Read more about how we’re thinking and where we’re investing here.

There has been an explosion in machine learning activity, and Shivon Zilis recently mapped out the current <a href="https://www.oreilly.com/ideas/the-current-state-of-machine-intelligence-2-0">machine intelligence ecosystem</a> as we enter 2016. This is one of the key areas that we’ll be following this year. <img class="aligncenter" src="https://d3ansictanv2wj.cloudfront.net/mi-landscape-2.0-r9_update-f2048bd678f522066e294ba8233b6615.jpg" alt="" width="3000" height="2250" /> While the opportunities here are tremendous, the exuberance surrounding machine learning distracts startups from a key hurdle: it’s data, not algorithms, that will dictate who wins in this space. Algorithms have largely been commoditized by now, so a machine learning company built around publicly accessible data isn’t defensible. Access to unique data isn’t a problem for incumbents like Google, Facebook, and Amazon. But, startups face a serious chicken and egg problem: they have to convince people to give them data, but the machine intelligence service won’t be useful until people (and a lot of people) are actually using the service and sharing their data. Matt Turck recently wrote about this “cold start” problem in a great post, <a href="http://mattturck.com/2016/01/04/the-power-of-data-network-effects/">“The Power of Data Network Effects</a>.” I’ve seen recommendation services struggle with this challenge. It takes too long to train the bots and users drop off the platform before the service gets valuable. Does this mean that startups can’t successfully play in the machine learning space? Not necessarily. First of all, there’s plenty of data out there. Likewise, there are plenty of opportunities to leverage machine intelligence to solve a business problem or improve day to day life. The key is to build a strong enough use case that compels people to give up their data before the benefits of machine learning and data network effects really kick in. Turck talks about the “data trap” strategy where startups build fun and free side apps to start gathering data. A great example is <a href="http://www.clarifai.com/">Clarifai</a>, a deep learning and image recognition company. Clarifai’s free consumer app, <a href="https://itunes.apple.com/app/apple-store/id1005812175?mt=8">Forevery</a>, offers instant value for its users, by making it easier to organize and find photos. With each new Forevery user, Clarifai has a bigger data pool to refine its image recognition technology. Mint.com is another example where users gave access to very valuable and unique data, even though Mint didn’t do much on the machine learning front in the beginning. The first step is to develop some kind of data acquisition strategy. The next step is to effectively communicate the value proposition. The focus has to be on tangible benefits for the end users, rather than the coolness of the technology platform. In Zilis’ summary of today’s machine learning ecosystem, she noted that “many machine intelligence companies have figured out that they need to speak the language of solving a business problem.” That’s a great sign. The bottom line is that if startups want to succeed in machine learning, their top priority should be building proprietary data sets. Creating a strong use case is the most effective strategy to get that data set. It’s not easy, but the good news is that it’s incredibly hard to unseat a service once data network effects kick in. ------- <span style="text-decoration: underline;">Follow-up note:</span> I have been getting a lot of feed-back on the post: while the thesis of data over algorithms is true for most machine learning applications, there are two scenarios in which algorithms are very important: a) situations with limited data in which you try to learn the most in minimum time with minimum data and b) very complex data (e.g. computer vision applications). In both scenarios, deep learning algorithms can make a significant difference.

<span style="font-weight: 400;">As 2015 comes to an end, it’s time to </span><span style="font-weight: 400;">reflect on what we’ve done and where we’re headed. It’s hard to believe that it has been just a little over a year since we </span><a href="https://www.versionone.vc/announcing-version-one-fund-ii/"><span style="font-weight: 400;">launched Fund II</span></a><span style="font-weight: 400;"> with $35M in commitments. Since then, we’ve accelerated our pace - backing exceptional entrepreneurs, feeding their growth, and seeing others out with successful exits. </span> <span style="font-weight: 400;">Here are some of the most memorable highlights for us in 2015: </span> <b>New portfolio companies</b> <span style="font-weight: 400;">This year, we welcomed five new companies to the Version One portfolio: </span> <ul> <li style="font-weight: 400;"><a href="https://www.versionone.vc/announcing-headout/"><span style="font-weight: 400;">Headout</span></a><span style="font-weight: 400;">:  an on-demand marketplace for tours, activities and experiences</span></li> <li style="font-weight: 400;"><a href="https://www.versionone.vc/shift-messenger-announcement/"><span style="font-weight: 400;">Shift Messenger</span></a><span style="font-weight: 400;">:  a communication platform for hourly workers</span></li> <li style="font-weight: 400;"><a href="https://boosterfuels.com"><span style="font-weight: 400;">Booster Fuels</span></a><span style="font-weight: 400;">:  still in private beta with the mission to eliminate the gas station errand permanently</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">and two more unnamed</span></li> </ul> We also publicly announced two companies, which we invested in at the end of 2014: <ul> <li style="font-weight: 400;"><a href="https://www.versionone.vc/introducing-volley/"><span style="font-weight: 400;">Volley</span></a><span style="font-weight: 400;">:  a community that connects people who need help with those who like to help</span></li> <li style="font-weight: 400;"><a href="https://www.versionone.vc/introducing-latest-investment-perceptiv-labs-future-smart-drone-technology/"><span style="font-weight: 400;">Vertical.ai</span></a><span style="font-weight: 400;">:  an autopliot app for drones that helps film dynamic aerial tracking shots</span></li> </ul> <b>Exits</b> <span style="font-weight: 400;">We celebrated the exits of three of our portfolio companies this year:  </span><a href="http://techcrunch.com/2015/02/03/fundable-acquires-clarity-fm/"><span style="font-weight: 400;">Clarity</span></a><span style="font-weight: 400;">, </span><a href="http://www.geekwire.com/2015/homeaway-buys-vacation-rental-startup-dwellable-plans-to-shutter-service-and-move-team-to-austin/"><span style="font-weight: 400;">Dwellable</span></a><span style="font-weight: 400;">, and </span><a href="http://techcrunch.com/2015/08/06/hackaday-acquires-diy-hardware-marketplace-tindie/"><span style="font-weight: 400;">Tindie</span></a><span style="font-weight: 400;">.</span> <b>Portfolio Summary and Investment Focus</b> <span style="font-weight: 400;">Our existing portfolio companies have continued to build their businesses with nine of them raising new rounds.  We participated in these follow-on deals to further support our founders.</span> <span style="font-weight: 400;">In total, our 23 active companies collectively raised $65M in 2015.</span> <span style="font-weight: 400;">While our highest concentration of companies is now in the SF/Bay Area (7), we remain geographically agnostic and are excited about the growing ecosystems outside the Valley, particularly Toronto/KW (6), Seattle (5), and NYC (2).</span> <span style="font-weight: 400;">In addition to investing, we focused our time on developing our investment theses, particularly around </span><a href="https://www.versionone.vc/marketplace-handbook/"><span style="font-weight: 400;">marketplaces</span></a><span style="font-weight: 400;"> (we published a handbook) and </span><a href="http://ondigitalhealthcare.com"><span style="font-weight: 400;">healthcare</span></a><span style="font-weight: 400;"> (we published a blog series with </span><a href="https://www.usv.com/blog"><span style="font-weight: 400;">Union Square Ventures</span></a><span style="font-weight: 400;">). Other themes that we are bullish about for the new year include: mobile marketplaces, B2B marketplaces, mobile SaaS, machine learning and artificial intelligence, Bitcoin and blockchain technology, Fintech 2.0, eSports, and virtual reality.  </span> <span style="font-weight: 400;">We’ve learned that we love all things with strong </span><a href="https://www.versionone.vc/network-effects/"><span style="font-weight: 400;">network effects</span></a><span style="font-weight: 400;">.  And of course, we stay inspired by our founders and the amazing new entrepreneurs we meet as they strive to make a positive impact.</span> <span style="font-weight: 400;">Finally, we are indebted to our LPs, colleagues, partners, portfolio companies, friends, and followers, for their continued support.  Thank you all for a great year!</span> <span style="font-weight: 400;">Our warmest wishes to you and yours for a happy holiday season and here’s to a 2016 that is rich in creativity and innovation.</span> <span style="font-weight: 400;">-Angela &amp; Boris :)</span>

Messaging platforms are gaining more and more mindshare. Michael Wolf’s deck on the <a href="http://www.businessinsider.com/michael-wolf-predicts-what-will-happen-in-the-tech-industry-in-2016-2015-10">Future of Tech and Media</a> shows that messaging will soon blow past social media as the dominant media activity and will add 1 billion new users by 2018. Given these numbers, the opportunities for messaging platforms are getting very interesting – particularly in terms of how to leverage messaging platforms to help people solve problems at work and in day to day life. Improvements in Artificial Intelligence (AI) and Natural Language Programming (NLP) now make it possible to: 1) use conversational language as command line (the bot understands what you want) and 2) automate the execution of the command (the bot does what you asked for). Extending this to a messaging platform…it’s possible to text your request and have it fulfilled by an automated and scalable backend. As the first stage, we’re already seeing messaging bots that act as gateways or aggregators to existing services. These are typically shortcuts to on-demand services, for example, to order food or an Uber. However, the potential for such aggregators is rather limited, since there’s not much value add and most on-demand services will likely have their own messaging bots in the near future. For example, 99designs has a Tasks Bot for Slack. <strong>Growth areas for messaging bots</strong> The real opportunity in messaging is to help people solve problems in areas where the underlying service doesn’t exist or the user doesn’t already know about the service (aka discovery). The challenge here is you need a lot of data about customer preferences and behaviors. At the beginning, there’s a high level of human involvement needed to train the AI bot on what the user is really asking for and how to best solve their need. <a href="http://www.buzzfeed.com/alexkantrowitz/time-to-meet-the-wizard-facebooks-messenger-head-pulls-back#.krlzoOq33">BuzzFeed</a> provides a look at the human training involved in developing M, the artificial intelligence that Facebook is building into its Messenger app. Currently, whenever someone sends M a message, the AI engine takes a first stab at the response, but this response is always reviewed, and then possibly edited, by a human trainer. All this data is fed back to M so that it will learn to get better and better. The idea is that in time, more and more tasks can be executed by AI and the service can scale. Startups <a href="http://techcrunch.com/2015/04/07/operator/#.3z5uqzd:g0Kj">Operator</a> and Magic are following a similar playbook. But it’s clear that any company needs a lot of funding to support all the human assistance while building the AI backend. On top of the funding challenge, there’s a real risk looming for startups who enter this space that Facebook or another platform will soon own the business. <strong>Automating tasks in the enterprise</strong> The other interesting opportunity for AI is to automate common chores for business users. We’ve already seen several virtual assistants go after the meeting scheduling market: e.g. <a href="https://x.ai/">x.ai</a> and <a href="https://claralabs.com/">Clara Labs</a>. The key here is removing any friction by being part of the regular workflow; for example, copy your virtual assistant on an email for an upcoming meeting and AI will take over from there. It’s easier for startups to gain traction in the business space, since the tasks are relatively cut and dry. For example, a bot can quickly grasp how to schedule a meeting with predefined rules versus having to field consumer requests on anything from ordering an anniversary present to drawing a picture of the sunset. For this reason, you probably don’t have to years of stealth mode to build a solution catering to enterprise users. In addition, the rise of Slack as a platform provides the opportunity to add further value: e.g. schedule internal meetings or collect regular updates from group members (see this <a href="https://medium.com/point-nine-news/slackbots-9144feee6f6#.4aj6kf96w">good overview</a> of the current Slack bots). However, given the fact that external communication in the enterprise is still centered around email, it will be interesting to see how AI players will bridge the different communications platforms or if they will just focus solely on email, Slack or another platform. The other big question is if these apps will stay focused on narrow use cases or try to position themselves as a complete personal assistant. The bottom line is we’ll be looking closely at the AI and messaging space in the coming year. The emergence of a new communication platform always sparks major innovation. Winning the consumer space will take heavy people-power and funding. There are easier opportunities to be had in the enterprise side, but you’ll need to figure out which use cases and which communication platform to pursue.

We know that companies are successful (or not successful) because of their people. Over the past few years, it has become relatively acceptable to work with a leadership coach and investors typically have a handful of good coaches in their rolodex ready to support and develop young founders. Yet even with this rise in popularity, leadership development is still primarily limited to CEOs and select executives. <a class="zem_slink" title="Shopify" href="http://shopify.com" rel="homepage">Shopify</a> is the first startup to really understand the value of coaching and leadership development across the organization. It’s about to launch “Managify 101,” an ambitious at-scale leadership program geared for all company managers. The evolution of coaching at Shopify from executive to at-scale offers important lessons for other startups. <strong>The evolution from one-on-one to full-scale production</strong> Coaching at Shopify began ordinarily enough. <a class="zem_slink" title="Cody Fauser" href="http://www.crunchbase.com/person/cody-fauser" rel="crunchbase">Cody Fauser</a>, one of Shopify’s first employees, was promoted to CTO. It’s a common situation at tech startups: Cody was one of the best and brightest, but didn’t have C-level leadership experience. Executive coach Cam Gregg began working with Cody one-on-one. Soon enough, Cam began coaching other Shopify executives as well. At the beginning, there was concern that having the same coach working with multiple executives would create conflict and tension. These concerns proved unfounded, primarily because Cam “provokes, but doesn’t force” (his words). If Cam thinks a conversation should take place among upper management, he’ll provide the catalyst for it, but won’t act as a go-between. As <a class="zem_slink" title="Harley Finkelstein" href="http://harleyf.com" rel="homepage">Harley Finkelstein</a>, Shopify’s chief platform officer, explained, “Coaches often become conduits between executives. But, that’s silly: we need to be talking amongst ourselves.” One of the remarkable things about Shopify is that they hired Cam as a part-time coach when the company had around 60 employees and full-time when they reached 160 employees. Cam now heads up Shopify’s Talent Acceleration group which offers two levels of coaching services throughout the company. One-on-one coaching for executives includes the full coaching toolkit: 6-10 hours of assessment, and regular biweekly meetings over the course of 6-9 months. Then, there’s ad hoc mentor coaching for 1^st and 2^nd level managers. Now, Cam’s group is about to launch “Managify 101,” a 90-day leadership development program that brings the benefits of coaching to the whole organization. Designed for newly promoted managers and new manager hires, Managify kicks off with a keynote from Shopify’s CEO, <a href="https://twitter.com/tobi">Tobi Lütke</a>, with other executives sharing real-world leadership stories. The one-day kickoff will be followed by a set of workshops and peer coaching over the next 90 days. This program is an impressive investment in talent development. As Cam explained, “if you’re prepared to play at your edge the whole time, what you’ll get from us is an incredible commitment to your ongoing development.” Management buy-in has been critical to the successful development of coaching programs at Shopify. Harley lets all of his direct reports know that he has been working with Cam for over four years and has seen a huge impact from the experience. From day one of any training session, Cam already has the social proof that Shopify execs believe in the results. “Where I’ve seen other companies get tripped up is they caravan in a coaching team for middle managers without any executive buy-in,” said Harley. “You can’t just drop coaching into the middle layer of the company and expect it to have an impact.” Talent development is often an afterthought or an obligatory check on the checklist once a company reaches a certain size. It’s refreshing to see Shopify develop such an innovative approach to making sure company leaders are nurtured and developed at all levels of the organization.

It has been twenty years since Ebay arrived on the scene and showed what happens when you bring people together over the Internet to do commerce. Since then, there has been tremendous innovation within online marketplaces. Startups have honed in on every vertical imaginable, branched out from product to services, and tested out new paths to profitability. I launched my first marketplace a few years after Ebay. Since then, I have been investing in and closely watching marketplace activity. Over the past few months, Version One has been diving into every detail of marketplace strategy while putting together a handbook on “<a href="https://www.versionone.vc/marketplace-handbook/">Building Marketplace Startups</a>.”  Through the process, I was struck by just how much marketplaces have evolved since eBay sparked the first wave of product-focused marketplaces in 1995. Here are the top six changes, and there are many more to come. <strong>1. Standard commerce tames the wild west</strong> In many ways, Ebay is the wild west of marketplaces. It’s a virtual free-for-all where everything is in the hands of the sellers, from pricing to packaging and shipping times. The marketplace itself is very hands-off, as any buyer who tries to contact Ebay’s customer service soon realizes. Amazon introduced many traditional commerce elements to its marketplace to make buying from a marketplace seller almost as easy as buying from Amazon itself. On the service side, we have seen consumer preference for standardization (Uber) over the wild west (SideCar). Every Uber driver will charge you the same rate to take you from Point A to Point B. Yes, you have a better chance of finding a cheaper ride with Sidecar, but you’re going to need to spend time browsing through a lot of choices (lowest fees, larger cars, etc.) in order to find what you want. <strong>2. Managed marketplaces unlock new supply</strong> Over the past few years, we’ve seen marketplaces take on additional parts of the value chain to deliver a better overall experience. For example, <a href="https://www.beepi.com/">Beepi</a>, a peer-to-peer marketplace for buying and selling cars, inspects every car listed on the platform, and gives sellers a guarantee: Beepi buys the car if you don’t sell it within 30 days. Other examples are Threadflip and Real Real, which handle the logistics/cataloging for their sellers in women’s clothes and luxury products respectively. By adding concierge/logistics support, these marketplaces unlock new supply. Most sellers of luxury items aren’t inclined to go through the time and trouble associated with cataloging and listing their clothes on a marketplace. But a managed marketplace takes care of all the hassle. <a href="https://www.versionone.vc/the-emergence-of-managed-marketplaces-greater-opportunities-but-a-lot-more-risk/">Managed marketplaces</a> also create new demand by helping buyers overcome the trust issues associated with peer-to-peer marketplaces. There’s a big difference in trust levels when buying a car via Beepi vs. Craigslist. <strong>3. Mobile unlocks new use cases</strong> Since the launch of Ebay, nothing has had a bigger impact on our buying and social habits than the rise of the mobile phone. Today mobile-first marketplaces, like <a href="https://www.varagesale.com/">VarageSale</a> and <a href="https://offerupnow.com/">OfferUp</a>, are creating new categories of Craigslist competitors. In the age of mobile where long-tail discovery (whether paid or organic) is less important, becoming a “home screen” app is critical to building a moat against the competition. Uber and Lyft do a great job locking in customer mindshare by becoming a home screen app on their customer’s mobile phones. <strong>4. New models emerge for monetization</strong> Traditionally, marketplaces made their money by charging sellers listing or transaction fees. The problem is that any kind of fee adds friction to the marketplace and encourages sellers to take their transactions off-platform whenever possible. We’re now seeing <a href="https://www.versionone.vc/freemium-model-future-marketplaces/">SaaS’ Freemium model</a> come to marketplaces where the base listing is free or very low and sellers pay for enhanced services like better placement. The low entry cost encourages more suppliers to list (increasing liquidity), but revenue can be hard to come by this way since only a small percentage of suppliers will choose to pay for the better placement. <strong>5. The rise of transactional service marketplaces</strong> Uber inspired a whole new generation of service-based marketplaces. But compared with their product-focused marketplaces, service platforms face a serious monetization hurdle. It’s hard for a service platform to stay in the loop and know when the service was performed and for how much. That makes charging by transaction fees tricky. The first service marketplaces were primarily lead generators. They made money by charging providers to contact customers. But the new generation of service platforms is trying to be as transactional as possible. In order to do so, you need to commoditize the service, with boxed offerings that include pre-defined scope, duration, pricing, deliverables, etc. This simplifies the end user experience, enabling customers to complete a transaction in just a few clicks. The less back and forth between provider and customer, the more likely they are to complete the transaction on platform. <strong>6. Decentralized marketplaces emerge</strong> Over the past year, we have seen the evolution of virtually fee-free decentralized marketplaces – such as <a href="https://openbazaar.org/">OpenBazaar</a>, <a href="http://lazooz.org/">Lazooz</a>, and <a href="http://www.openname.com/">Openname</a>. These marketplaces flatten hierarchies and centralize control: anyone sells, anyone buys. Everything – from trust, rules, identity, and payment – are at the peer level. Today these deep web marketplaces are known as places for selling/buying drugs, weapons, and other illicit goods. But there are some who are betting that the decentralized model will revolutionize all online commerce – and not just be limited to the stuff that can’t be sold on eBay. It’s still early, but with the emergence of blockchain, there is a huge potential to disrupt global commerce. <strong>The bottom line</strong> Considering that buying and selling goods and services are such an integral part of daily life, there’s still great opportunity for innovation in new verticals, models, and monetization strategies. I believe there’s a lot more room for new entrants and disruption. Some of the world’s next billion dollar companies will be marketplace startups.

<span style="font-weight: 400;">Nearly four years ago, I joined the <a href="http://www.insightdatascience.com/">Insight Data Science</a> team and we launched</span><span style="font-weight: 400;"> an intensive 7 week post-doctoral training fellowship bridging the gap between academia and data science. Since then, over 400 Insight alumni have been hired as data scientists or data engineers at top tier companies like Facebook, LinkedIn, Twitter, Airbnb, and Google.</span> <span style="font-weight: 400;">Although I formally left the company in March 2013 (but continue to have an advisory role there), I still field countless questions from entrepreneurs who are looking to hire a data scientist. Because, let’s face it: data scientists aren’t necessarily easy to find. When we launched Insight, the term “data science” was still new. Most founders don’t know where to begin the search, who to look for – not to mention how to bring a data scientist on board so he or she can make valuable contributions to the company.</span> <span style="font-weight: 400;">To that end, I’ve decided to summarize some of what I learned at Insight about recruiting and training data scientists. Keep in mind, I won’t share all of Insight’s secret sauce, but will hopefully provide enough high level lessons and principles that you can apply to your own hiring process.</span> <b>Finding a data scientist is hard. Let’s look in a different haystack!</b> <span style="font-weight: 400;">When looking for a data scientist, the default approach for many companies is to go after software engineers with a strong interest in statistics, data mining and machine learning. Understandably, there’s a preference for people with strong programming skills.  However, drawing math-loving engineers from the tech talent pool isn’t a sustainable strategy.  This simply reallocates scarce resources.</span> <span style="font-weight: 400;">So, what’s the alternative? Many PhDs in engineering and the sciences work with “big data” and code their own algorithms. Did you know that particle physicists analyze 5TB of data every day? Or that mechanical engineers design and code models for computational fluid dynamics? It’s a huge opportunity: PhDs can be the new untapped supply for data scientists!</span> <span style="font-weight: 400;">The advantages of hiring PhDs are that you don’t have to teach them the fundamentals of math and they have a working knowledge of how to code that make them a functional data scientist. In addition, you earn their loyalty for taking a chance on them straight out of school and the cost of hiring is lower.</span> <b>Can everyone make the leap from academia to industry?</b> <span style="font-weight: 400;">Academia and business are two different worlds and there are several kinks to work out when transitioning a grad student</span> <span style="font-weight: 400;">their academic environment to industry. Here are three pointers learned from my experience with Insight.</span> <ol> <li><em><span style="font-weight: 400;">Teach the language of industry</span></em><span style="font-weight: 400;">.  Just because most academics work in Matlab or R, this doesn’t mean that they aren’t capable of translating their skills to Python, SQL or MongoDB. Learning a new programing language only takes a week or two, so don’t let specific programming languages deter you from interviewing someone who knows their math.  In the meantime, you can communicate and evaluate a candidate’s abilities with pseudocode.</span></li> <li><em><span style="font-weight: 400;">Guide workflow from rigor to iteration</span></em><span style="font-weight: 400;">.  The culture of academia is such that feedback from professors is infrequent and we only defend our thesis when it is absolutely perfect. As a result, PhDs are trained to work towards rigor as opposed to embracing iteration.  This is obviously contrary to a startup’s culture. As you engage with a data scientist candidate or new hire, find ways to test their ability to ask for, receive and respond to critical feedback.  Perhaps send them home with a 3-day data project and have them check-in every so often.</span></li> <li><em><span style="font-weight: 400;">Find actionable insights and not just interesting facts</span></em><span style="font-weight: 400;">.  In the spirit of rigor, academics tend to report anything and everything as opposed to thinking about KPIs.  This is arguably the hardest mindset to shift because PhDs have spent years in an environment where research is not pushed for commercialization or productization. They aren’t necessarily going to be thinking about the business implications of the data right out of the gate.  </span>During the interview process, provide a data set and ask candidates an open-ended question like “How can we use this data on product usage to increase our revenue / engagement?” Then post hire, you need to provide the right context to help a new data scientist find the knobs you can turn and the levers you can pull to grow your company.</li> </ol> <b>Three ways to hire a data scientist</b> <span style="font-weight: 400;">If you want to hire a data scientist, there are three approaches to take:</span> <ol> <li><span style="font-weight: 400;">Tap your network for existing data scientists and other people who deal with lots of data (quants at banks, etc.).  Remember, there are many people who do the same work as a data scientist but their roles might not be labelled as such.</span></li> <li><span style="font-weight: 400;">Go directly to universities and begin a recruiting and training process, following the principles above.  This is a time-consuming process, especially when you need to create the framework from scratch. If you do choose this route, I’d suggest building a small team of data science advisors to help you. </span></li> <li><span style="font-weight: 400;">Partner with companies like </span><a href="http://www.insightdatascience.com"><span style="font-weight: 400;">Insight</span></a><span style="font-weight: 400;">.  They are currently in the SF Bay Area, NYC and Boston, and have started an online program for those outside these geographies. If you’re a founder or company who would like to work with them on your search for a data scientist, feel free to ping me for a connection or email them directly at </span><span style="font-weight: 400;"><a href="mailto:info@insightdatascience.com">info@insightdatascience.com</a></span><span style="font-weight: 400;">.</span></li> </ol>

<em><strong class="it cb">Note: We’ve updated this handbook — read the latest and greatest 3rd edition <a href="https://www.versionone.vc/marketplaces-guide-ed3/">here</a>!</strong></em> <span style="font-weight: 400;">At Version One, we love marketplaces. Just look at our portfolio, and you’ll understand how important we think marketplaces are to the future of commerce.</span> <span style="font-weight: 400;">Over the past year, we have focused much of our blog content on helping marketplace founders build their companies. That’s because we realized that while there is a lot of great information for tech startups, not much of it deals specifically with marketplaces. And anyone building a marketplace company knows that marketplaces face unique challenges - including how to solve the chicken and egg problem with supply and demand and how to monetize when services are delivered offline.</span> <span style="font-weight: 400;">To that end, we put together a handbook, </span><a href="https://gregburnison.ca/code/version1v/images/Marketplace-Handbook-11-08-2015.pdf" target="_blank" rel="noopener noreferrer"><b>A Guide to Marketplaces</b></a><span style="font-weight: 400;">. It compiles many of the insights we’ve learned from working with great marketplace companies and analyzing the industry.</span> We hope the handbook helps you in your own journey to break down walls in how goods and services are bought and sold. There’s no single way to build and scale a marketplace, but the book can help you figure out your own path to build supply and spark the virtuous circle of supply and demand. <p style="text-align: center;"><a href="https://gregburnison.ca/code/version1v/images/3px.png"><img class="alignnone wp-image-2505 " src="https://gregburnison.ca/code/version1v/images/3px-700x990.png" alt="3px" width="351" height="496" /></a></p> <p style="text-align: center;"><strong>Download the <a href="https://gregburnison.ca/code/version1v/images/Marketplace-Handbook-11-08-2015.pdf">PDF</a> or <a href="https://www.dropbox.com/s/ze84q6pvquroi8n/Marketplace%20Handbook-Nove%208%202015.epub?dl=0&amp;preview=Marketplace+Handbook-Nove+8+2015.epub" target="_blank" rel="noopener noreferrer">ePub</a> version of the handbook and refer to it as you need it.</strong></p> <span style="font-weight: 400;">Lastly, we’d love to hear if you find the content useful. What should we expand on? Did we leave anything out? </span><span style="font-weight: 400;">Please leave a comment below or reach out to us directly.</span> <span style="font-weight: 400;">Happy reading!</span> <span style="font-weight: 400;">Angela and Boris</span> <em>May 2016 Update:  we summarized this handbook in slide deck form for quick reference which you can access <a href="https://www.versionone.vc/marketplace-deck/">here</a>.</em> <em>March 2018 Update: we just published the second edition of this handbook <a href="https://www.versionone.vc/marketplaces-guide-ed2/">here</a>.</em>

Given the frequency of media stories focused on a handful of startups with billion dollar valuations and high profile rounds, it’s understandable that founders get impatient with their own company’s natural pace of growth. This sense of urgency is intensified by the fact that growth is an elemental part of being a startup. <a href="http://www.paulgraham.com/growth.html">Paul Graham</a> summed it up: “a startup is a company designed to grow fast.” Growth is what separates a startup from a plumbing or barbershop business. However, every company needs to grow at its own pace. Some business models simply don’t scale as quickly and need time to mature – and this isn’t necessarily a sign of failure or a bad business model. For example, trying to grow an e-commerce site like you’re a tech company will not end well. Gross margin levels are typically not a big deal for a tech company, but can be the make-or-break area for an e-commerce company. In addition, e-commerce and tech companies deal with very different operational complexities: touching/shipping physical products vs. quickly fixing a bug for a SaaS product. And, these company types have very different acquisition strategies: an e-commerce site typically relies more on paid advertising where you can easily lose your shirt if you’re not careful as you scale. There is a limit to how quickly an organization can scale. Be aware of the impact that high growth can have on your company’s culture, team, and execution ability. I have seen many startups get screwed up in all three dimensions because they grew too fast. Yet having said this, there are times when fast growth is necessary. If you’re in a <a href="https://www.versionone.vc/type-market-operating/">‘winner takes all’ market</a>, going slow and steady and then ending up second or third won’t help your cause. The bottom line? Get inspired by your peers with hyper-growth and large fundraises, but don’t forget to think about what’s the right speed for your market, startup, and team.

<span style="font-weight: 400;">Over the past two weeks, we have introduced our </span><a href="http://bit.ly/1iBW8mg"><span style="font-weight: 400;">social platform KPI dashboard</span></a><span style="font-weight: 400;"> to track the key metrics for social platform startups. We first covered the </span><a href="https://www.versionone.vc/social-platforms-metrics-1/"><span style="font-weight: 400;">high-level metrics</span></a><span style="font-weight: 400;"> (such as DAU, MAU and stickiness) and then took a deeper dive into </span><a href="https://www.versionone.vc/social-platforms-metrics-2/"><span style="font-weight: 400;">content engagement</span></a><span style="font-weight: 400;">. </span> When we talked about engagement, we related it to a specific action performed by a user (i.e. post, upvote, share, etc).  As a result, last week’s discussion was very much content-centric, but as you may have noticed, we also suggested tracking events related to connections made between users. This is because we feel that stickiness is not only a function of common interests, but also the depth of relationships on the platform. This is probably one of the reasons why anonymous social networks have not scaled in the past (i.e. Secret). When building your platform, consider how much users care about the people that they are interacting with and whether the communication is bidirectional. Also, are the relationships pre-existing from an offline social graph or new ones generated online? <span style="font-weight: 400;">Refer to our </span><a href="http://bit.ly/1iBW8mg"><span style="font-weight: 400;">KPI dashboard</span></a><span style="font-weight: 400;"> - which has now been updated to included </span><b>Relationship Engagement Metrics. </b><span style="font-weight: 400;">Make a copy of the spreadsheet and use it however you see fit.</span> <a href="https://gregburnison.ca/code/version1v/images/SS3.png"><img class="aligncenter wp-image-2493 size-large" src="https://gregburnison.ca/code/version1v/images/SS3-700x452.png" alt="SS3" width="640" height="413" /></a> <p style="text-align: center;"><i>Our KPI dashboard for Social Platforms: <i>access the updated Google spreadsheet with relationship engagement metrics </i><a href="http://bit.ly/1iBW8mg"><i>here</i></a></i></p> <p style="text-align: left;">Some examples of key relationship metrics to track are:</p> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">Total number of connections made (</span><i><span style="font-weight: 400;">note: that in the case of public social networks, like Twitter, that use the “follow” model, you may want to count both the number of one-way followers and double-followers)</span></i></li> <li style="font-weight: 400;"><span style="font-weight: 400;">New number of connections made</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Number of connections per user (plot a distribution)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Number of new connection made per user (this gives you a sense of virality)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Total number of groups (public or private sub social networks with 2+ people)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">New number of groups</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Number of groups subscribed to per user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Number of messages received and/or sent per user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Number of interactions with unique individuals per user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Average % of connections that any single user interacts with</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Averages and % growth rates for all these metrics</span></li> </ul> <span style="font-weight: 400;">Based on these metrics, you can identify which users have a strong influence on others. Then follow these users’ journeys to learn from their behaviour and actions.</span> <span style="text-decoration: underline;"><strong>Main Takeaways from this Series of Posts</strong></span> <span style="font-weight: 400;">Our thesis is that social platforms are sticky due to two things: </span> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">Unique/user-generated content</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Strong underlying relationships between users</span></li> </ul> <span style="font-weight: 400;">As you build your social platform and track your key metrics, ask yourself the following: How can you drive daily engagement to your platform?  Can you do it with content only (i.e. in the case of a community) or do you also need to focus on building relationships between users?  </span> <i><span style="font-weight: 400;">If you haven’t already done so, you can download the </span></i><a href="http://bit.ly/1iBW8mg"><i><span style="font-weight: 400;">KPI template here</span></i></a><i><span style="font-weight: 400;">. I’d love you to share any questions, comments, or ideas on these metrics (and any of the social platform metrics) in the Comments section below. </span></i>

<span style="font-weight: 400;">Last week, we introduced our </span><a href="http://bit.ly/1iBW8mg"><span style="font-weight: 400;">social platform KPI dashboard</span></a><span style="font-weight: 400;"> to help provide a framework for founders to think about their businesses and preempt the inevitable questions that arise during the fundraising process. In the first post, we talked about the </span><a href="https://www.versionone.vc/social-platforms-metrics-1/"><span style="font-weight: 400;">high-level metrics to track</span></a><span style="font-weight: 400;"> - such as DAU, MAU, and your stickiness ratio. </span> In the context of these high-level metrics, the only requirement for a user to be considered “active” is that they somehow engaged with the product. This could be as simple as logging in to the platform once or as engaged as creating content every single day. This means you need to dive deeper into your users to understand not just the total volume, but also the quality of these users and what it means to be engaged. <b>Defining engagement on your platform</b> <span style="font-weight: 400;">The first step is to create a list of actions that a user can perform on your platform.  Here are some examples of events to record. Note that this is not a comprehensive list nor are all events applicable to every product.  </span> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">Account created:  when a user creates an account</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Account verified:  when a user’s account is successfully verified</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Profile created:  when a user creates a profile</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Profile completed:  when a user completes his/her profile</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">User sign-in:  when a user signs onto the platform (note:  most platforms now have auto sign-in so users are always logged in)   </span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">User sign-out:  when a user signs out of the platform</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Content created:  when a user posts text, a url or an image (track this by content type)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Content viewed:  when a user views content</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Content shared:  when a user forwards content to others (track this by internal sharing and external sharing)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Content upvoted:  when a user “likes” or “favourites” content created by another user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Comment added:  when a user responds to content created by another user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Comment upvoted:  when a user “likes” or “favourites” another user’s comment to content</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Following:  when a user “follows” another user (or content thread)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Followed:  when a user is “followed” by another user</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Connection requested:  when a user requests to “friend” another user (in the case where relationships require a double opt-in)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Connection made:  when a user accepts a friend request, or when a user follows another user who follows him/her back</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Group created:  when a user creates a group (public or private)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Membership requested:  when a user asks to join a group</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Membership granted / group joined:  when a user successful joins a group</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Message sent:  when a user sends a message to another user (you can also separate public and private messages)</span></li> </ul> <span style="font-weight: 400;">Once you have this list, you can create a framework of engagement.  I previously described the concept of an </span><a href="https://www.versionone.vc/engagement-pyramid/"><span style="font-weight: 400;">engagement pyramid</span></a><span style="font-weight: 400;"> which organizes a user’s behaviour in a hierarchy.  Recall:</span> <ol> <li style="font-weight: 400;"><span style="font-weight: 400;">At the top, place the activity that you determine to be a sign of someone who is most engaged with your product. Typically, this is the factor driving your most important metrics.</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Next, consider all the other activities that someone can do, and place them in decreasing order from what is the hardest / has the most friction for someone to do / requires the most energy, to what is easiest.</span></li> </ol> <span style="font-weight: 400;">By doing this, you can identify the highest level of user engagement and subsequently, allocate resources to achieve that outcome.  </span> <span style="font-weight: 400;">While not obvious, the network effects on social platforms are, in fact, </span><a href="https://www.versionone.vc/network-effects/"><span style="font-weight: 400;">two-sided</span></a><span style="font-weight: 400;">. </span><span style="font-weight: 400;">Participants generally follow the</span><a href="http://avc.com/2011/06/dont-forget-your-logged-out-users/"> <span style="font-weight: 400;">100-10-1 rule</span></a><span style="font-weight: 400;">, with 1% creating content, 10% engaging with it, and 100% consuming it.  Morgan Brown wrote a </span><a href="https://segment.com/blog/growthhackers-community-metrics/"><span style="font-weight: 400;">great piece on community metrics</span></a><span style="font-weight: 400;"> that draws a deeper analogy to marketplaces, referring to participants as “supply” and lurkers as “demand”.</span> <a href="https://gregburnison.ca/code/version1v/images/pyramid.jpg"><img class="aligncenter wp-image-2336 size-full" src="https://gregburnison.ca/code/version1v/images/pyramid.jpg" alt="pyramid" width="455" height="492" /></a> <p style="text-align: center;"><em>A sample <span style="font-weight: 400;"><a href="https://www.versionone.vc/engagement-pyramid/">engagement pyramid</a> </span>for social platforms </em></p> <span style="font-weight: 400;">Knowing that “user-generated content is king” for social platforms, the most active users (creators) are therefore those who post status updates, photos, or links that are drive others to come to the platform time and time again.  These creators are inspired when they are recognized for their contributions via shares, comments, favourites, etc., for their posts. </span> <span style="font-weight: 400;">Given this positive feedback loop, it becomes clear that the goal of a social network is to encourage the logged-in user to be a contributor.  Also, we note that the defensibility of social platforms increases as users participate in the same way – when they create and consume – because, after all, content is what ultimately drives engagement.</span> <b>Build your engagement dashboard</b> <span style="font-weight: 400;">Now that we have defined engagement, the next step is to understand how “healthy” that engagement is.  From the list above, you can build a content engagement </span><a href="http://bit.ly/1iBW8mg"><span style="font-weight: 400;">dashboard</span></a><span style="font-weight: 400;"> similar to what we did for the high-level metrics.  </span> <p style="text-align: center;"><a href="https://gregburnison.ca/code/version1v/images/SS-engagement1.png"><img class="alignnone wp-image-2486 size-large" src="https://gregburnison.ca/code/version1v/images/SS-engagement1-700x447.png" alt="SS-engagement" width="640" height="409" /></a> <i><span style="font-weight: 400;">Our KPI dashboard for Social Platforms: access the updated Google spreadsheet with content engagement metrics </span></i><a href="http://bit.ly/1iBW8mg"><i><span style="font-weight: 400;">here</span></i></a></p> <span style="font-weight: 400;">For instance, on a monthly basis, you can tally all the “positive actions”.  Let’s use images as our example.  In this case:</span> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">Total images uploaded to date</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">New images uploaded this month</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">% image growth (WoW, MoM, etc.)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Average images uploaded per user (take the mean and median)</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Average comments per image</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Average upvotes per image</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Average shares per image</span></li> </ul> <span style="font-weight: 400;">However, don’t stop there.  We recommend understanding the distribution of these actions (i.e. plot a histogram).  You can calculate useful metrics like:</span> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">% of users who have uploaded an image at least once, 2x, 5x, 10x in the past week, month, or year</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">% of users responsible for uploading 80% of all images</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">% of users who have engaged with images at least once, 2x, 5x, 10x in the past week, month, or year</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">and so on...</span></li> </ul> <span style="font-weight: 400;">We also recommend that you track your user journey while using the platform in order to understand if there are some actions that tend to lead to others and where there can be improvements made to push people to be content creators.  I wrote a post on </span><a href="https://www.versionone.vc/correlation-probability/"><span style="font-weight: 400;">correlation vs. conditional probability</span></a><span style="font-weight: 400;"> that outlines an easy way to understand the relationship between two actions or events.</span> <span style="font-weight: 400;">If you haven’t already done so, you can download the </span><a href="http://bit.ly/1iBW8mg"><span style="font-weight: 400;">KPI template here</span></a><span style="font-weight: 400;">. If you have any questions or comments on these metrics, share your thoughts below. And next week, we’ll discuss the last piece of the puzzle - measuring engagement by relationships.</span> <span style="font-weight: 400;"><em>Now you can read Part 3 <em><a href="https://www.versionone.vc/social-platforms-metrics-3/">here</a>!  Or re-visit Part 1 <a href="https://www.versionone.vc/social-platforms-metrics-1/">here</a>.</em></em></span>

The Blockchain is such an innovative and fundamentally transformative technology, but Bitcoin and Blockchain apps have not yet scaled as much as I would have liked. This slow traction is likely due to two things. First, Bitcoin startups that focus on commerce/payment solutions in North America are competing with existing products with strong network effects. Credit cards actually work very well for most consumers to buy things online and in stores. Secondly, there’s still too much friction to bring on new users. The UI of most Bitcoin apps is cumbersome and uninviting. More importantly, apps require users to already own Bitcoin in order to engage. That approach won’t encourage the average consumer to give it a try. This is why I believe that the most successful Bitcoin/Blockchain company won’t look anything like what we’ve seen to date. In order to scale, a startup will need to address these two factors. <em><u>Enable a service that is not yet possible today</u></em> The first wave of Bitcoin apps went after payment-related use cases in North-America and hence against strong existing network effects. The second wave has started to go for more green field opportunities by focusing on those areas where current fees are really high (i.e. remittances), where consumers are underserved (the unbanked), or where the existing payment infrastructure is weak (developing countries). But the most promising opportunities may be in completely new use cases of the Blockchain and we have seen a few early attempts at this: managing licensing of digital products through the Blockchain or enabling distributed data storage - all use cases that enable services that - at scale - are only possible through the Blockchain technology. <em><u>Lower sign-up friction </u></em> Start-ups need to focus more on the user experience for first-time users. More specifically, a successful Blockchain app will allow users to engage and participate even if they don’t yet own Bitcoin and get them hooked on a service before having to get Bitcoin. The bottom line is I’d love to see Bitcoin/the Blockchain succeed, and am rooting for those entrepreneurs who can scale the technology up – perhaps by following these two principles…

<span style="font-weight: 400;">Back in July, we wrote about</span><a href="https://www.versionone.vc/network-effects/"> <span style="font-weight: 400;">network effects in marketplaces and social platforms</span></a><span style="font-weight: 400;"> and how they are critical to defensibility. In the post, we talked about the difference between direct and indirect network effects, and how they may or may not relate to virality. Since then, we have been further developing our investment thesis on social platforms and want to add more nuance to the discussion.</span> <span style="font-weight: 400;">In our post, we defined network effects this way:</span> <p style="margin-left: 40px;"><i><span style="font-weight: 400;">“When a new user/member is added to the network, it increases the value of the product or service to all other users.” </span></i></p> <span style="font-weight: 400;">This definition (like many others we’ve read) mainly refers to the </span><b><i>quantity </i></b><span style="font-weight: 400;">of users: i.e. more is better. Most likely, this is because the concept of network effects originated in the commerce or software world where every new user brings the same amount of added value to the product or experience. </span> However, network effects work a little differently on social platforms – and the value of a network shouldn’t be measured just by size, but also by the types of users themselves. More specifically, you need to look at the underlying relationship between the new user and existing users, and/or the influence of the new user. <span style="font-weight: 400;">The qualities that matter vary by the dynamics of the social network. For example:</span> <ul> <li><span style="font-weight: 400;">When a close friend joins a messaging platform like WhatsApp or a private social network like Facebook, you get more value than when that same close friend joins a community like Reddit.</span></li> <li><span style="font-weight: 400;">On a public social network like Twitter, an active user with lots of influence (i.e. a celebrity) will be of more interest and drive more value for the entire user base than a lurker.  </span></li> </ul> In addition, all strong social platforms experience a time of exponential value creation:  when critical mass is reached, tremendous network value is unlocked.  As we discuss the quality of connections within a network, we also need to review the concept of critical mass beyond sheer volume of users, and how its onset differs depending on the type of social platform. Think about the following examples: <b>Communities</b><span style="font-weight: 400;"> like Reddit are fun (aka valuable) after reaching a critical mass. Here, the actual size of the network is most important and it matters less if there’s little to no relationship between participants. The challenge in reaching critical mass is that communities typically have much less mass appeal or applicability (the strongest communities are pretty niche).</span> <b>For social networks</b><span style="font-weight: 400;">, the number of users needed to reach critical mass is likely less than for communities.  This is because the value of the network is tied to the number of meaningful connections made from underlying social graphs (i.e. your network of friends or colleagues) </span><i><span style="font-weight: 400;">and/or</span></i><span style="font-weight: 400;"> the number of influencers on the platform.  Note:  we’re using "influencers" as an umbrella term for celebrities </span><i><span style="font-weight: 400;">and</span></i><span style="font-weight: 400;"> great content creators without an existing brand.</span> <p style="margin-left: 40px;"><i style="line-height: 1.5;">For “private” platforms that use a friending model like Facebook</i><span style="font-weight: 400;">, the “</span><a style="line-height: 1.5;" href="http://ryangum.com/chamath-palihapitiya-how-we-put-facebook-on-the-path-to-1-billion-users/">7 friends in 10 days</a><span style="font-weight: 400;">”</span><a style="line-height: 1.5;" href="http://ryangum.com/chamath-palihapitiya-how-we-put-facebook-on-the-path-to-1-billion-users/"> growth strategy</a><span style="font-weight: 400;"> illustrates how the value of the network grows first from a core group of users you care about. Then inevitably through the bigger network of networks, personal networks become less fragmented and social circles (school, work, family) overlap.</span></p> <p style="margin-left: 40px;"><i><span style="font-weight: 400;">For “public” platforms that use a follow model like Twitter</span></i><span style="font-weight: 400;">, influencers typically lead the exponential value creation mentioned above as they’re responsible for creating most of the content that is consumed.</span></p> <b>For messaging</b><span style="font-weight: 400;">, you can arguably reach critical mass with just two people who text a lot between each other.  It’s probably true that the more utilitarian a platform is, the more important the quality of relationships between participants is than sheer numbers.</span> <b><span style="text-decoration: underline;">Our main takeaway</span>:  </b><span style="font-weight: 400;">When building your social platform, don’t just think about increasing your user base in terms of raw numbers, but also consider the value and effect that particular users bring to the platform.  Whether you target a core group of friends for private social networking, or a majority of a group if the underlying social graph is group-based, focus on user quality in order to leverage the inherent virality of social platforms to ultimately reach critical mass.</span>

<a href="https://twitter.com/SeanSilcoff">Sean Silcoff</a> had <a href="http://www.theglobeandmail.com/report-on-business/careers/careers-leadership/for-version-one-ventures-boris-wertz-failure-is-in-the-job-description/article26340269/">a thoughtful article in Saturday's Globe &amp; Mail</a> on my path from entrepreneur <a href="https://www.versionone.vc/how-to-bootstrap-your-way-into-becoming-a-vc/">to investor</a>. It is a well-written piece with many great quotes and anecdotes and included the most important lesson about early-stage investing that I learned over the past 8 years: <blockquote>“If the fail rate of our companies is not high enough, it’s not because we are exceptional investors, it’s probably because we’re not taking enough risks. You need to invest in things that nobody has seen, where nobody recognizes the opportunity, and where nobody else would invest – and be right.”</blockquote> Taking huge risks can be scary for entrepreneurs and investors alike, but risk lays the foundation for what can ultimately be special companies and outsized returns. Every day I remind myself that there is no such thing as a stupid idea, as a market that is too early, or a company that has too many missing pieces. Many of our bets might fail. But when they succeed, it’s pure magic to get to be a part of something special from the beginning.

When you’re in the midst of building a company, it’s hard to fully assess the significance of various activities and decisions- particularly for first-time entrepreneurs who don’t yet have the wisdom learned from prior mistakes. As an investor (who was also once a first-time entrepreneur), I have noticed several key areas where founders either overestimate or underestimate the value to their company. Here they are… <u>5 things founders often underestimate</u> <strong>The value of a great hiring process</strong> Your company won’t survive if you’ve got a great idea surrounded by a mediocre team. Founders should stay involved in every hiring decision for as long as possible. When you’ve reached the scale when it’s no longer possible to be hands-on in each decision, you need to make sure you have strong hiring methodologies in place – such as <a href="http://www.businessinsider.com/amazon-bar-raisers-2014-1#ixzz3kUMrbTDM">Jeff Bezos’ bar raisers</a> that were implemented in Amazon’s early days to help weed out “cultural misfits.” <strong>The importance of reference checks</strong> I’m always surprised at how often people still get hired without extensive reference checks. When you’ve got a good feeling about someone, it’s tempting to just move ahead without any kind of due diligence. But informal reference checks can be very revealing, particularly if you can talk to people that weren’t provided by the candidate, but know him or her very well. <strong>The need to build culture from the beginning</strong> Founders often mistake culture with things like morning yoga, a company climbing wall or food truck lunches. But culture is deeper than just free perks and it’s also something you can’t just “switch on” down the road when you have time for it. From the start, you should be thinking about the type of company you want to build and then make sure your systems and decisions support that vision. <strong>The importance of the right set of investors and advisors</strong> This point might seem self-serving, but too often I’ve seen founders look at investors as checkbooks. Money is important but over and over again I’ve also seen just how important great investors and advisors can be in shaping an early stage company. <strong>The value of focus </strong> Your company and your product will never be all things to all people. And not every customer opportunity will be worth pursuing in the long run. You need to narrow your focus to a few key priorities, and then make sure this focus is communicated clearly and repeatedly throughout the company. &nbsp; <u>4 things founders overestimate</u> <strong>The importance of money</strong> No amount of money in the world is going to get you to product-market fit. And raising too much money before you find product-market fit will usually kill your start-up. <strong>The importance of launch and fundraising announcements</strong> Start-up founders typically want to make a big PR splash when they’re launching their product or closing a funding round. A major launch announcement can make a lot of noise in moment, but it’s more important to focus on generating continuous demand than your 15 minutes of fame. Likewise, fundraising announcements can drive investor interest for future financing rounds, but I’ve found it’s usually better to just stay under the radar and build the company. <strong>The value of senior employees with industry experience</strong> Young startups often feel compelled to bring on senior employees with impressive resumes and a deep knowledge of their vertical. While specific expertise can help in some cases, you should hire for smarts, passion and hard work, rather than experience. <strong>The magic of a silver bullet</strong> It would be great to find that silver bullet to take your business to the next level. The problem is that you rarely ever find them. When founders waste their time chasing after that one non-existent thing that will solve all their problems, they ignore all the small incremental improvements that will make a real difference. <em>This list is hardly exhaustive. What did I miss? I’d love to hear your thoughts on what else founders overestimate or underestimate… </em>

One of the most common questions I get is “How can I become a Venture Capitalist?” I suppose this is because to many people, being a VC is a coveted position: you control large amounts of money and can pick and choose where that capital goes. [But let me tell you that the VC business is also much tougher than anybody imagines from the outside] The traditional career path in venture capital has been around for decades. You start working at a firm as an associate, with the assumption that your hard work will pay off and you’ll become a principal, and then make partner. But today, there’s an equally viable route where one can bootstrap his or her way to becoming an investor. This is the path I took after selling my startup to Amazon. The rise of the bootstrapping VC is made possible by micro-funds. An investor no longer needs to raise $100 million; a $10 million fund puts you right in business. In addition, the concept of a single general partner (GP) has become pretty widely accepted among investors now. Go back just 10 years ago, and most investors avoided solo GP funds. If you’re an entrepreneur considering the move to VC via a non-traditional route, here’s my advice: <em>Do your homework</em> The beauty of the web is just how much information is now at your fingertips. Many in the VC community are incredibly generous when it comes to sharing their insights and experiences via blogging and social media. In this respect, my mentor was <a href="http://www.avc.com">Fred Wilson</a>, one of the first investors who blogged on a daily basis. I still remember Fred’s series on <a href="http://avc.com/2008/08/venture-fund-ec/">Venture Fund Economics</a> from back in 2008. There have also been posts about the <a href="http://avc.com/2013/11/loss-ratios-in-early-stage-vc">appropriate loss ratio</a> for an early-stage fund and how the <a href="http://avc.com/2012/11/what-has-changed/">start-up landscape</a> is changing. Keep in mind that Fred was blogging this nearly eight years ago – a time when blogging and sharing knowledge was definitely not the norm for VCs. Today, you’ve got countless options; my advice is to find a couple of different investors and follow their blogs and tweets regularly. You’ll learn about the industry as well as the in’s and out’s of investing strategy. <em>Start with small angel investments</em> Reading and research is the first step, but at some point you’ll need to dive in and start investing. Between 2008 and 2011, I made around <a href="https://www.versionone.vc/our-portfolio/">35 investments</a> with my own money. Through the experience, I learned about what and who to invest in, along with what and who to avoid. At the same time, I was also building a track record. At this point the actual amount you’re investing is less important – it can be just $5K per deal. The key is to invest in a wide variety of deals so you start learning from your wins and losses. <em>Build your brand</em> The third step is to build your brand and network in order to attract dealflow and co-investors. If you went to Stanford or worked for a company like Google or Facebook, then you inherently have a great network. But if you don’t already come with a natural network, you’ve got to put the effort in. As an example, my friend <a href="http://blog.semilshah.com/about/">Semil Shah</a> has done an amazing job building his brand through a combination of smart writing, social media, and organizing get-togethers and targeted conferences. In terms of networking, your location is important to a certain degree. For me, Vancouver is home. While it’s got an active startup scene anchored by companies like <a href="http://www.hootsuite.com">Hootsuite</a>, <a class="zem_slink" title="BuildDirect" href="http://www.builddirect.com/" rel="homepage">BuildDirect</a>, and <a href="http://www.clio.com">Clio</a>, it’s not a huge tech hub – and certainly wasn’t eight years ago when I got into investing. This meant I needed to build my network by frequent trips to the Valley, Seattle, NYC, Toronto, and Los Angeles. Try to identify the key people in each of these start-up hubs and be bold (yet polite) in offering to buy them coffee. In the smaller ecosystems, you will probably be connected to 80% of the important people within your first two to three visits. Given its size and activity level, the Valley will always require an ongoing effort. I also spent one month working out of the <a class="zem_slink" title="Union Square Ventures" href="http://unionsquareventures.com/" rel="homepage">Union Square Ventures</a> office in New York, and one month in the <a class="zem_slink" title="First Round Capital" href="http://www.firstround.com/" rel="homepage">First Round Capital</a> office in San Francisco in order to get a deeper understanding of the start-up world in these two worlds. The bottom line is if you want to become a VC, there are plenty of paths in front of you now, some traditional and others less so. Bootstrapping isn’t necessarily a short cut to success. There’s a lot of hard work and risk involved, but you do have the chance to experiment and be autonomous along the way.

<span style="font-weight: 400;">It has been a while since we’ve written about healthcare.  Our last update was in June 2014 when we shared </span><a href="https://www.versionone.vc/thesis-update-healthcare/"><span style="font-weight: 400;">what we were excited about</span></a><span style="font-weight: 400;">:  patient-doctor networks and patient health data.  This hasn’t changed but we’ve certainly learned more from the many founders and physicians who we have spoken to over the past year.  We have a deeper understanding of the current system and consequently, see more opportunities to innovate.  But with great possibilities come even more questions on the future of healthcare. </span> <span style="font-weight: 400;">Over the past two years, as co-investors in </span><a href="https://figure1.com/"><span style="font-weight: 400;">Figure 1</span></a><span style="font-weight: 400;">, we have worked closely with </span><a href="https://www.usv.com/"><span style="font-weight: 400;">Union Square Ventures</span></a><span style="font-weight: 400;"> on developing our investment thesis.  Today, </span><a href="https://twitter.com/libovness"><span style="font-weight: 400;">Jonathan Libov</span></a><span style="font-weight: 400;"> of USV and I are excited to share our research:  5000 words on what we’ve seen and what we’re curious about in digital healthcare.  We’ll be publishing it in 4 parts over the course of 4 weeks.  Each week, we will cover a series of questions and answers on the topics of mobile, data, apps, networks and marketplaces.  </span> <span style="font-weight: 400;">You can read our first instalment here: </span> <strong><a href="http://ondigitalhealthcare.com/">http://ondigitalhealthcare.com/</a></strong> <span style="font-weight: 400;">... and sign up at the top of that page to be notified about the next post.   </span> We hope that you’ll find this piece informative and thoughtful, and that it inspires you to join us in our discussion on digital healthcare.

This week everyone has been talking about <a class="zem_slink" title="Zirtual" href="http://zirtual.com/" rel="homepage">Zirtual</a> - the on-demand virtual assistant company that collapsed on Sunday night, suddenly laying off more than 400 employees (the company now is going to be acquired by Startups.co). Zirtual CEO Maren Kate Donovan blamed the collapse partly on an outsourced CFO firm and numbers that failed to account for two additional pay periods. You can read <a href="http://fortune.com/2015/08/13/zirtual-maren-kate-donovan-2/">Donovan’s story</a>, as well as <a href="http://fortune.com/2015/08/13/zirtuals-outsourced-cfo-gives-his-side-of-the-shutdown-story/">Ryan Keating’s account</a> (the outsourced CFO) in Fortune. In the wake of Zirtual’s story, many are now saying that a startup should never outsource its financials. But I don’t necessarily agree. I don’t know the specifics of this story beyond what’s in the media, so I can’t comment on Zirtual’s experience. But generally speaking, my advice to startups is that you should <a href="https://www.versionone.vc/never-ever-outsource-your-core/#ixzz3impoAb71">never outsource activities that are core</a> to your business. For most startups, core functions are product design, product development, marketing, sales, PR, and customer service. But it can be okay to outsource the things that don’t create a competitive advantage, such as: legal, accounting, hosting, and analytics. However, whenever you do outsource a function, you need to be on top of your service provider. For finance, your CFO firm should deliver monthly numbers within 2-3 weeks of month’s end…and you need to go through these numbers and understand them. For legal, you need to read and understand any and all documents that your lawyer creates. It’s the same as with any service provider or vendor. If you don’t stay on top of the service they’re delivering, you can end up either getting poor quality results or paying a lot of money for the level of work you’re getting. Of course, as you scale, you will start to take your outsourced functions in-house. For example, you’ll probably need a controller once you reach 30-40 people and really large companies have their own corporate counsel. But the bottom line is if you have been outsourcing your financials – and have been happy with the results thus far – you don’t need to suddenly change your strategy just because of this high profile example.

These days, we’re hyper-focused on marketplaces, communities and social platforms. We love these platforms for many reasons, but the primary one is the power of their network effects and how they create a high degree of defensibility for startups. <b>What are network effects? </b> “Network effects” is a frequently used term and it is sometimes confused with having a large customer base.  Network effects is more than just a large number of users - they kick in when the value of a product depends on how many other users there are. When a new user/member is added to the network, it increases the value of the product or service to all other users. This increased value can be in the form of cost reduction (in user acquisition as an example), higher liquidity (in a marketplace), stronger community or deeper relationships (in social networks), etc. Generally speaking, network effects are categorized into two categories:  direct and indirect. <b> </b><b>Direct Network Effects</b> The simplest network effects are direct: an increase in usage leads to a direct increase in value for all other users. Social sites like Facebook, Twitter and <a href="https://www.wattpad.com/">Wattpad</a> have a direct network effect. If I’m a user, I gain some benefit if you join. When we think about direct network effects, we consider the roles of the participants. In a marketplace, users are buyers, sellers, or sometimes both. Likewise, in a community or social platform, users can be content creators, content consumers, or both. In a marketplace, the roles of a buyer and seller are typically distinct - effectively creating two-sided network effects. When there’s overlap between buyers and sellers, it’s easier to get your network off the ground at the beginning as you’ll achieve liquidity more quickly. For example, in Etsy, Craigslist and Airbnb, we see some overlap between buyers and sellers (or hosts and guests in Airbnb’s case). While overlap helps a marketplace at the beginning, two-sided network effects add to the defensibility of your marketplace over time (since there’s more fragmentation). For example, with Uber, there’s little overlap between drivers and riders. Social platforms and communities can also exhibit two-sided network effects, although they are less obvious. Participants generally follow the <a href="http://avc.com/2011/06/dont-forget-your-logged-out-users/">100-10-1 rule</a>, with 1% creating content, 10% engaging with it, and 100% consuming it. The defensibility of social platforms increases as users participate in the same way - when they create <i>and</i> consume - because, after all, content is what ultimately drives engagement. We also need to consider how network effects fall on the local-global spectrum: <ul> <li><em>Local network effects</em>: The value-add of a product may not increase with the overall size of its user base, but is dependent on the size of a smaller subset. In this case, a user’s experience might be influenced directly by the number of people and actions of a small subset of users (i.e. connections through the user’s personal/professional network or neighborhood). We see local network effects at play with instant messaging and private social networks as well as in the on-demand economy.</li> <li><em>Global network effects</em>: These take place with a larger-sized network with few barriers in payment or language.  We see global network effects on global marketplace like eBay and public social platforms like <a href="http://figure1.com">Figure 1</a> and Twitter. Keep in mind that some marketplaces, like Etsy, will fall in between local and global network effects, as there can be shipping and language limitations.</li> </ul> <b>Indirect network effects</b> Network effects can also be indirect. In these cases, when more people use a product or network, it sparks the production of complementary products and goods - thus increasing the value of the original product. A common example is with hardware and software; the more people that use a hardware product, the more likely it is that developers will build software and apps for that hardware. But indirect network effects can also apply to marketplaces, communities, and social platforms. For example, developers build applications and products on top of APIs of existing marketplaces and platforms – for example, Shopify has a very active developer/value-add app ecosystem. In addition, countless services sprout up to support popular platforms – like Airbnb management services (Guestly and Pillow) and lease options for Uber/Lyft (Breeze). <b>How are network effects different from virality?</b> When information can be shared rapidly and widely from one user to another, or when the rate of adoption increases with adoption, we have virality. In other words, the product grows faster as more users adopt it (to a certain limit). Network effects and virality often go hand-in-hand, but not all network effects are viral and not all viral products have network effects. For instance: <ul> <li>Marketplace network effects often have low virality.  While you can achieve some virality through incentives, the two-sided network effect does not lend itself to adoption of users on one side to come on-board with invitations/interactions from the other.</li> <li>News outlets/channels, gaming, and communication products can be highly viral products without strong network effects.</li> </ul> Network effects are critical to scaling and sustaining a marketplace, community or social platform.  As you build yours, think about how users benefit each time someone else participates, and in what way:  directly or indirectly.  Understanding how network effects impact your startup brings you one step closer to building a more defensible product.

I often get asked about how to determine the valuation for a marketplace startup that is starting to scale. So I’ll run through some basic math for how we value marketplaces at version one. The main multiple we like to use for marketplace businesses is GMV (Gross Merchandise Volume…the total volume of goods sold on the marketplace). Our rule of thumb is that marketplaces at scale are valued at roughly 1x annualized GMV (typically about 6-8x annual revenue). These are for marketplaces that are growing fast and are category leaders. Our assumptions for this valuation: <ul> <li>Scale: &gt; $1b GMV</li> <li>YoY growth of at least 30%</li> <li>Take rate of about 10-15%</li> </ul> Taking <a class="zem_slink" title="Etsy" href="http://www.etsy.com/" rel="homepage">Etsy</a> as an example… <ul> <li>Consensus estimate of approximately $270M for 2015 revenue</li> <li>$2.4B for 2015 GMV</li> <li>Take rate of about 11%</li> <li>YoY growth for most recent quarter (ending Q1 2015): 45%</li> </ul> Etsy had a market cap of about $2.22B as of July 20, with a revenue multiple of 8.2 and a GMV multiple of 0.92 (but note that its GMV multiple was as high as 1.7 immediately following its IPO). <a href="http://www.bvp.com">Bessemer</a> offers a <a href="http://www.slideshare.net/taliagold/valuation-presentation-blog-version">good overview of current valuations</a> for different business models: SaaS, marketplaces, consumer, and ecommerce. You can see in their chart below that marketplaces get some of the highest revenue multiples because of their operational leverage and high defensibility at scale. <a href="https://www.versionone.vc/how-we-determine-valuations-for-marketplaces/screen-shot-2015-07-13-at-11-08-40-am/" rel="attachment wp-att-2430"><img class="aligncenter wp-image-2430 size-full" src="https://gregburnison.ca/code/version1v/images/Screen-Shot-2015-07-13-at-11.08.40-AM.png" alt="Screen Shot 2015-07-13 at 11.08.40 AM" width="631" height="529" /></a> &nbsp; <strong>So what about early-stage marketplaces?</strong> These calculations apply to marketplaces that have already solved the chicken-and-egg problem, reached liquidity, and have become a category leader. But you might be wondering how to value your early-stage marketplace startup. Early-stage companies are valued very differently. In this case, metrics don’t count - we’re evaluating the team, idea, and vision. Then as the marketplace starts to scale fast, the multiples are often very high because growth is high as well. But you need to understand that your marketplace will ultimately be valued at 1x GMV. If you’re a founder of an early-stage marketplace, you should put your focus on two things: <ol> <li>Growing GMV</li> <li>Proving out take-rate</li> </ol> Sometimes we see entrepreneurs who pitch impressive GMV numbers, but haven’t proven out that they can ultimately get to a significant take rate. For example, a marketplace that generates leads instead of being part of the transaction might have a take rate as low as 2-3%. In this case, the GMV needs to be 5x bigger than a comparable marketplace with a 10-15% take rate. The bottom line? Don’t wait too long to prove out your monetization. <em>(Thanks to <a href="http://www.startupcfo.ca/">Mark MacLeod</a> for giving me the idea for this blog post.)</em>

A few months ago, I wrote about the questions that we typically ask when <a href="https://www.versionone.vc/marketplace-traction-data">evaluating a marketplace opportunity</a>. Now, inspired by our friend <a href="https://twitter.com/chrija">Christoph Janz</a> at <a href="http://www.pointninecap.com/">Point Nine Capital</a> who created a <a href="http://christophjanz.blogspot.com/2013/04/a-kpi-dashboard-for-early-stage-saas.html">KPI dashboard for early-stage SaaS</a>, we have put together a KPI template for marketplaces. We recognize that every marketplace is different, but each one is also similar at its core: each marketplace has a seller (supply) and buyer (demand) side, and acts as an intermediary to bring them together. We hope that our version of a Marketplace KPI dashboard helps founders manage their business and pre-empt those due diligence questions in a fundraise.  Below are some screenshots of the template which you can also access via the Google doc <a href="http://bit.ly/1IWdtkc">here</a> (make a copy of it and then you can edit away). <a href="https://gregburnison.ca/code/version1v/images/Marketplace-KPI-Screen-Shot.png"><img class="aligncenter wp-image-2428 size-large" src="https://gregburnison.ca/code/version1v/images/Marketplace-KPI-Screen-Shot-700x691.png" alt="Marketplace KPI Screen Shot" width="640" height="632" /></a> <p style="text-align: center;"><em>Screen shot of <a href="http://bit.ly/1IWdtkc">Marketplace KPI Google Spreadsheet</a></em></p> There are plenty of notes within the dashboard itself but here’s a high level overview. The dashboard is separated into three dimensions to measure the efficacy of your business:  1)  overall marketplace metrics; 2) seller/supplier metrics; and 3) buyer metrics. <span style="text-decoration: underline;"><strong>Overall Marketplace Metrics</strong></span> <b>Gross merchandize volume </b>(GMV) is the total sales dollar value for goods sold or services purchased through the marketplace over a certain time.  Given that GMV is one of the most important marketplace KPIs, founders should track its growth rate on a monthly and yearly basis, and understand its makeup by customer acquisition channel.  With GMV and the total number of transactions, we can compute the <b>average order value </b>(AOV) since: <i>          GMV = # of Transactions * AOV </i> <b>Revenue</b> is the income that the company receives from facilitating connections in the marketplace. It comes in the form of transaction fees, listing fees, and/or the offering of premium seller/supplier services.  With revenue and GMV, we can calculate <b>take rate</b> via: <i>          Revenue = GMV * Take Rate</i> In addition to take rate, we can evaluate business efficacy by calculating the total <b>customer acquisition cost</b> (CAC) of buyers and sellers/suppliers as a percentage of revenue. <span style="text-decoration: underline;"><strong>Seller/Supplier Metrics</strong></span> Start with general seller/supplier and growth KPIs, for example: <ul> <li>Number of sellers/suppliers</li> <li>Sellers/suppliers growth rate</li> <li>Number of listings (n.b. in a services marketplace, this is not applicable so perhaps consider the number of active suppliers at a specific time instead).</li> <li>Listings growth rate</li> <li>Average listing price</li> <li>CAC</li> </ul> Engagement KPIs are most important to the seller/supplier side so make sure you track these at the very least - examples: <ul> <li>Cohort analysis: percentage of sellers/suppliers still active 1 month and/or 1 year after signing up</li> <li>GMV retention: average percentage of Month 1 GMV generated by sellers/suppliers in Month 12</li> <li>Concentration: percentage of revenue generated by the top 20% sellers/suppliers</li> <li>Net promoter score (NPS)</li> </ul> <span style="text-decoration: underline;"><strong>Buyer Metrics</strong></span> Similar to seller/supplier metrics, we can categorize these into general and engagement. General KPIs - examples: <ul> <li>Number of buyers</li> <li>Buyer growth rate</li> <li>Average dollar amount purchased per buyer</li> <li>Average number of orders per buyer</li> <li>Average order growth per buyer</li> <li>CAC</li> </ul> Engagement KPIs on the buyer side - examples: <ul> <li>Repeat buyer contribution: percentage of buyers who have purchased more than once; and percentage of GMV generated from buyers in previous months</li> <li>GMV retention: average percentage of Month 1 GMV generated by buyers in Month 12</li> <li>Concentration: percentage of revenue generated by top 20% buyers</li> <li>Cross pollination (if applicable): percentage of buyers whose second purchase is in a different category</li> <li>NPS</li> </ul> <span style="text-decoration: underline;"><b>Closing Notes</b></span> Any thoughts or questions on the dashboard? Please comment below, as we’re sure there are many others that will benefit from the discussion. We’d also love any feedback on other important KPIs to add, or ones to remove. We’re looking forward to iterating!