How to solve the cold-start problem (and how not to)

One of the inherent problems of starting a company that depends on a two-way marketplace, whether it’s user-generated reviews, listings sites, or anything else that requires two parties - is the “cold start” or “chicken-and-egg” problem. If you launch with no content (or users, or reviews), you look like a chump. But you have to launch in order to start getting users, reviews, and listings.

I’ve been thinking about this a lot, and recently saw a shit-awful “solution” to it, so I thought I’d share some thoughts about some of the strategies I’ve seen employed.

Piggyback on an event

This is how Airbnb did it - twice. There’s a great interview with Nathan Blecharczyk where he talks about how they took advantage of the 2008 Democratic and Republican conventions to promote their site. Despite the fact that they got on CNN, because of the diffuse nature of their business, the travelers who’d stayed at Airbnb properties scattered home, and they still hadn’t reached critical mass. Which brings up…

Geographical targeting

Limiting yourself to one region allows you to reach critical mass sooner - when there’s a smaller available area, fewer people look like more. It’s not a panacea - you still need to go out there and pound pavement to front-load the user base or prime the content pump - but it’s a hell of a lot easier to do so when you’re covering one city instead of the world.

Bonus - If you’re launching in the city you live in, you not only save travel bucks by not having to fly around and create relationships, you probably also have existing relationships in your own community you can leverage. If you’re starting a comic book review site and you’re a regular at the local comic book store, you’ll be able to approach the owner, employees and other patrons for help, being a friendly face and all.

Seed the room

To follow on some of the geographic targeting advice above, don’t overlook seeding the room with service providers. These are going to be the pros (or semi-pros) who will profit most from your site. In the example of the comic book review site, reach out to the comic store owner to see if she’ll write some reviews for your site before it launches - in exchange for linking to her store’s web site, naturally.

Caveat - try to not be a dick about it. Or at least, try not to break any laws.

And if all else fails, use coach marks with calls to action

One of the worst things that can happen when you launch without critical mass is that the user sees… nothing. From the user’s perspective, she just visited your site hoping to find something great, and you’ve just wasted her time. This is not making users awesome. That said, coach marks can help you message around a cold start. This is something I’ve been seeing a lot on some of the mobile UI sites I use for inspiration - basically, you use fairly emphatic messaging to tell users how not to fail:

Airbnb tells you exactly what’s wrong and how to fix it: via

Labelbox gives you a pictorial tutorial upon opening the app: via

And I decided to do something similar with a mobile app I’m working on:

I really appreciate how the small form factor and short user attention span inherent in mobile apps really forces designers to remove friction and emphasize user delight :^) Google’s empty Hangout sad robot does a good job of this, too, although I wish there were more visual directionality…

How not to…

So, full disclosure: I’m working on a startup that competes with this company. I weighed whether or not I should reference them this way, but I figure 1) other designers and developers should see this and not replicate it and 2) I’m just petty that way. I’ve obfuscated the company’s name to protect the better funded.

After finding out about these guys, I went to their site and was presented with one option - sign up for an account. Okay, I’ll bite. But then after spending five minutes giving them my personal information, I click a link that reads “Take me straight to the search page” only to end up at this:

Dude, really? You’re going to harvest my information, not give me anything for it, and then insult me by linking to a Wikipedia page about what “beta” means? If you’re not sure your users know what beta means, use another word. Otherwise you’re just making them do your work for you. Also, it’s “what that means”, not “what the means”. *eyeroll*

One of the integral aspects of being a competent UX designer is empathy for the user. I know how hard it is to be so far up your own tube that you’re no longer objective about how to make sure the user feels she’s getting value from your product. But it’s totally worth the time to run through ways to make sure you make a good first impression on your first users. Hell, bring in a disinterested third party to run through something simple like this on a test server. Listen to his feedback.

So that’s what I’m thinking about regarding cold start right now. If you have any additional advice (we totally need it!) email me at alexis dot peterka at gmail dot com.

Thanky.