[Editor's note: Seattle entrepreneurs Barry Chu and Dave Cotter share some of their startup lessons in starting the new mobile app SquareHub. The three-part series is running this week on GeekWire, starting today with their thoughts about minimum viable products. In part two, they'll talk about how they determined product features. Part three, set to run Thursday, will include their ideas on customer acquisition].
Reid Hoffman is wrong. As with most memorable quotes, there is nuance that may be missed by those eager to launch and learn as fast as possible. It’s this: the point of launching early is to learn from the market. But in some markets, if your product sucks, you learn nothing.
Enter the “minimum viable product” — or MVP. This is the lightest, most scaled-down version needed to learn what the market wants.
A good MVP lets you:
- Prove whether you are actually solving a problem and verify that demand is there. It’s OK to target a small number of early adopters and hard-core users for this.
- Spend less time planning, more time responding. You’ll learn more from user feedback than from trying to think through every product feature before launch.
To figure out if Reid’s quote applies to your product, ask a simple question: Are you entering an unknown or a known market?
- In an unknown market, users have no expectations because they’ve never seen anything like your product. A typical reaction is “Why would I ever do that?” (Remember when you first heard about Twitter?) You launch rough, fail fast, and iterate. Ideally, you find polarized users who either love or hate your product. By understanding them, you begin to understand and rationalize the size and scope of your target market.
- In a known market, there are plenty of competitors, comparable products and substitutes. Consumers have expectations for usability, design and features. Example: Google Docs vs. Microsoft Word or Excel. When launching Google Docs, Google needed to meet some basic user expectations for how a word processor or spreadsheet would work. A lousy product in a known market can convince you that your core idea is wrong, when in fact, the idea is valid but the execution has missed its mark.
To determine if we were entering a known or unknown market, we used a Porter’s Five Forces model of user-level product dynamics.
- Suppliers: Are their factors increasing potential users? Do you believe that demand for your product is increasing because of certain conditions of the market? Ideally, every startup should be able to answer yes, as this determines whether you believe your market will grow.
- New Entrants: Do users see lots of new entrants? If you see a new startup trying to solve the same problem, it’s clear many folks feel it and have ideas for how to solve it. In our case, there are many group messaging and task management products flooding the market.
- Rivals: Are there competitive services available? Are there services that folks currently use in lieu of a “perfect” solution, or current services that promise similar value?
- Substitutes: Are there easy substitutes users can fall back on? What’s your user’s next best alternative to your product?
- Buyers: Are users actively searching for a solution? Is there “pull” demand from users? Are folks actively and publicly asking for a solution? If so, you better listen to what they say.
These questions help estimate the size and growth of the product opportunity. The new entrants, rivals, and substitutes questions gauge user expectations.
When LinkedIn launched in 2003, it was entering an unknown market. Social networks were in their infancy. Friendster was at its peak. MySpace had yet to launch. And Facebook didn’t exist yet. In this environment, Hoffman could afford to launch a first version he was “embarrassed” about because:
- No true comparable existed.
- The closest substitutes were offline and cumbersome.
- The overall exposure for that “embarrassing” first version was small.
The minimum bar for LinkedIn was that their solution had to be better than emailing friends of friends or building a Rolodex person-by-person in order to be valuable (and learn from). If a user entered LinkedIn circa 2003 and saw a list of people they knew who weren’t in their Rolodex, they saw immediate value.
With a low minimum bar, in a place where networking really matters, LinkedIn’s founders could learn quickly without sacrificing reputation. They could address a need immediately with just a list of “folks in your space.” (By the way, LinkedIn ended 2003 with “only” 81,000 members – just a reminder that success doesn’t always look like Instagram).
At our startup, on the other hand, we are entering a known market. Users have well-developed expectations for what an app should behave, how messaging works, what social networks look like and what a “private family network” might mean for our target demographic.
Thus, we must meet or exceed basic user expectations, both as an app and as a network. Our MVP can’t be sloppy or “embarrassing,” or we don’t learn anything at all. It needs to be minimal enough to build with limited resources, but viable enough that we can learn from the known market we are entering.
So maybe it’s not so much that Reid was wrong. For many startups, he’s right, and frankly, it’s kind of hard to argue with his track record.
But, for some startups, especially those entering a known market, be sure to ask yourself what your MVP should be. It may not make sense to be embarrassed.
Next, in tomorrow’s post, we walk you through how we turned “minimum” and “viable” into a discrete set of product goals and features.