Building an Excellent Web or Mobile Startup Adam Jackson Meltwater Entrepreneurial School of Technology (MEST) June 2012
What we’ll cover today: • The founding team • The goal: Product / Market Fit • Identify the metrics that matter • User acquisition • Test, test and test • Barriers to Entry • How & When to raise money • Growing the team
The Founding Team • If you are not a tech person, partner with one. • Best odds for success : work with who you know and trust • Second: “try before you buy” – FounderDating, MeetUp, etc • Values matter most – Work ethic, culture, commitment, integrity – Ideas don’t matter – teams do. Don’t compromise on this step.
The Goal: Product / Market Fit • (my) Definition of a Startup : a collection of people that deeply believe the same set of theories who come together to build a product and company to make those theories a reality. • Build something people want to use
Timing Matters • … “Timing is everything. But it’s also the hardest thing to control. It’s hard; entrepreneurs are congenital y wired to be too early. And being too early is a bigger problem for entrepreneurs than not being correct. “ … – Marc Andreessen, 2012 @ Stanford • But don’t let that discourage you – just go build something!
The Minimum Viable Product • Build something compelling and get it out to users • Doesn’t have to be pretty, but aim for useful, addictive, fun, essential • Your initial product may be way off … but you may accidental y find something people want! – Great example of this : InstaGram • Could be a … – Service, App, Website, Game, Physical product, Content (podcast, premium educational site, community, etc)
• “If you are not embarrassed by your first release, you’ve launched too late!” –Reid Hoffman (LinkedIn / Greylock)
Identify the Metrics that Matter • Web & Mobile apps – Installs / unique visitors – Daily / Monthly active users – Goals completed / engagement – Sharing / referral – Response rate to notifications (PUSH or email) – Revenue • Understand traffic sources & cohorts – Track where your traffic / users comes from to identify the best channels – organic or SEM – Create cohorts by segmenting users by relevent attributes: • Spending habits, visit frequency, week they joined, device they use to access the site, demographic data, etc
Acquiring Users • Buying users – Facebook / google / mobile ads / affiliate networks / email / direct response – Test every channel that makes sense and continue to test the marketing message & the creative until you find a sustainable way to buy users for less than what you monetize them for – If you can buy users for less than your ARPU, you have a nice and scalable SEM model that should be easy to raise money against.
Organic User Acquisition • Search: always practice good SEO when possible. – Lots of good long-tail search oppty still out there. – Google favors good content over farmed content – you can de-thrown an incumbent fairly easily • Social: make it easy for users to share your product with others. – Gamification works : Leaderboards, points, badges, levels – Goal = Network Effect the more users that are on it, the more useful/fun it is for the whole group.
The Viral Coefficient • K = i * conversion % – i = # of invites sent out by each new user (email/fb/twitter/sms) – Conversion % = the % of invites that convert to new users • The goal is to get a K >= 1.2~1.3 – The higher K is, the faster the resulting viral growth is • K = 2 means each new users bring in 2 more users, which wil bring in 4 more, which wil bring in 8 more, etc…
Maximize Virality • Leverage the social graph – the key to viral distribution • Make sharing implicit instead of explicit • Spotify with music sharing on social graph • Path with implicit geo checkins • Allow your service to connect to the social sharing platforms that matter to your users.
Understand the Desire Engine • Trigger – Internal : app asks the user a highly targeted question that they must enter the app to answer – External: user sees the app posted on a friend’s timeline and user must install that app to see what the friend is doing • Action: what your app is trying to get people to do – Take a picture, answer a question, follow me, etc • Variable Reward: – This is the emotional payoff of the product - we design this to maximize dopamine release in order to make the product habit-forming. – Stick to the 7 sins for success here : Lust, Gluttony, Greed, Sloth, Wrath, Envy, Pride • Commitment – After delighting the user, get some info from her that will make her next run through the loop more fun • Reminders – Stale users are useless. We must keep them coming back for more. – Use email / SMS / iOS reminders – Keep it relevant!
Your Instincts Are Wrong : Test Everything • Once you find a MVP you wil have a flood of ideas for features to build next – Your instincts on which features will work will be wrong more often than not. Why is this? • So, build as many of them as you can and conduct lots of A/B (multivariant) testing • Build tracking into everything: – Every click, swipe, tap, install, uninstall, interaction. – Google Analytics, MixPanel, Omniture, etc
The Testing Cycle Should Never End • Test 2, 3, 4+ versions of the same feature and see which performs best. • Performance = the metrics that are important to you (Goals) – Time on site, clicks per session, revenue, friends invited, etc
Example: Website testing w/ Google Analytics
The Results Goal 1: Report Purchase Version A sent 75% more traffic to the purchase page, but actual sales conversion rates between A and B were the same. This means A wastes traffic! Goal 2: Value lookup Version B wins by smal margin Goal 3: Servicing Link Version B converted 253% better. A lot of these clicks were probably ones that would have been wasted on Goal 1 in Version A.
Build Barriers to Entry • Every successful model wil be copied immediately by companies around the world. Build barriers to protect against this: – Excel ent service (Zappos) – Quality perception (Google) – Network effect (LinkedIn) – Out-execute (GroupOn) – Price & Efficiency (Amazon) – Customer Development : Pay close attention to what your users want, constantly watch the data and talk to them. Stay ahead of the competition.
How and When to Raise Money • Incubators are a great place to start – Great exposure to network of people that can help you get started and grow • Build a great advisory team – Find people who have successfully done what you’re trying to do. This could be in the same space you’re in or just someone who is very good at fund raising, tech scaling, organic growth, etc – They will lead you to great investors (or may become great investors) • Angel List (www.angel.co) – Get referred in by another startup or an angel – Build an excellent profile (screenshots, team, press hits, advisory board) – Keep it private until you’re ready to raise – Ideal for seed1 ($50-150k) & seed2 ($250-1mm+) – Don’t use AL too early or too late in your life cycle • Crowd-funding – Indie-gogo, Kickstarter
Growing the Team • Hire slowly, fire quickly • “ You wil never say ‘Boy I wish I would have waited longer to fire that guy’” – Me, ~2006 • You will churn 20-40% of your staff in the first 18 months. Do not be discouraged – this is normal. • Always hire the best quality people that best fit the values and culture the founders have set from the beginning – Smartest is NOT always best for every role.
Growing the Engineering Team • Scaling up technically : experience matters most • It takes an engineer to manage an engineer • SCRUM • Let devs set their own estimates, then buffer them to the business side. • Do post-mortem after every release so team keeps improving • Bin predictability into your release schedules (once you find the business model that scales)