Ship Web 2.0 Features Early, and Don’t Fear User Hatred, Investor Dave McClure Tells Dogpatch Labs Audience
“Ship early, ship buggy,” was the advice San Francisco-based Web 2.0 investor Dave McClure gave to his audience at Cambridge, MA’s Dogpatch Labs last week.
“The world is filled with products that shipped too late,” he continued. One example? The online budget-tracking service, Mint.com, a company McClure invests in along with other sites such as online jobs portal Simply Hired and TeachStreet, a Seattle-based website that connects students and teachers. McClure said that Mint delayed in delivering its products because of concerns with security and its features, rather than focusing on the fact that no one else out there was providing what it did. (In the end this didn’t seem to hurt the company, which was acquired last fall by Intuit.)
But what if the product is messy and users hate it? That’s OK, said McClure, who runs the Founders Fund seed-stage investment program FF Angel and publishes the blog “Master of 500 Hats.” The worst–case scenario would be customers who don’t care at all. “Hate is a good discovery signal,” he says. “If people hate something you’ve done, guess what? You’ve found something that matters. Haters are good.”
Too many startups get trapped in the linear cycle of developing a feature, shipping it out, and starting the process all over again without ever stopping to measure or reflect on the customer experience and conversion metrics around the first feature. Companies should continually be examining whether an existing feature improves user experience, rather than focusing on getting the next one out, says McClure, who launched Simply Hired and ran marketing at PayPal from 2001 to 2004. Dogpatch, created by Polaris Venture Partners, and David Cancel, CEO of web marketing startup Performable, hosted McClure’s lunchtime presentation. Polaris principal Jon Lim said after the event that the firm is working to further cross-fertilization of ideas between its bi-coastal communities, which also include San Francisco and New York.
McClure, who spoke later the same night at an MIT Enterprise Forum event on early-stage funding, counseled his audience of more than 100 to go home and immediately kill a feature (most Web startups have too many anyway, he says). Once users start complaining about the removal of a feature, that’s the one companies should really focus on improving and bringing back better than ever. The advice came as an inspiration after one of McClure’s earlier points, said in jest to get attendees to understand why they should kill features: “The reality is you suck, your product sucks, and you will fail…hard.”
That perspective was woven into McClure’s slide show presentation on what he calls pirate startup metrics, so dubbed for the fact that the acronym for their five tenets spell out AARRR (a sound McClure playfully made after roughly each point he made in the presentation). Those five tenets would be acquisition, activation, retention, referral, and revenue. These elements are a better guide to customer satisfaction than traditional measures like unique visitors per month, he said.
Acquisition metrics signify how users are directed to a Web page, and activation is defined by what keeps them there. For activation metrics, McClure said, entrepreneurs should look at how long a user spends on a site, how many pages they click on, and whether that converts to a sign-up.
Retention metrics should track how often a user returns to a page, an action companies should encourage with regular e-mail communications, McClure said. For him, referral means how existing users get their friends to a page, which can be through non-traditional methods of social media or widgets. And revenue—well, that’s how a Web startup makes its money, which McClure suggests shouldn’t be solely through Google ads.
For McClure, it seems less is more. In addition to advising attendees to pare down their list of features, he also cautioned them against cluttering their products’ homepages. A company’s main feature or call to action should take about half of a Web page’s real estate, he said. Next should come roughly three smaller but recognizable storylines, and a list of all the other stuff available on a website should take up only a small slice of space, he said.