Spinnakr calls for death of the data scientist in push to democratize analytics
Spinnaker founders Michael Mayernick and Adam Bonnifield
Adam Bonnifield and Michael Mayernick think Google Analytics is too complicated, the data scientist is superfluous, and the A/B test is too slow at delivering results. Analytics and optimization, they argue, need to be radically democratized and made accessible to even the most technologically novice business and content owners. And what’s more, those people should have one system that takes care of everything that analytics are useful for, from insight to recommendation to action.
These are the theories that underlie Bonnifield and Mayernick’s startup, Spinnakr, which is launching in beta today and is backed by about $1 million in seed funding from Andreessen Horowitz, 500 Startups, and Berlin-based Point Nine Capital. Spinnakr monitors activity on websites and sends notifications to the owners when it notices significant activity. It then recommends a particular action so that the website owner can capitalize on the incoming traffic, even to the point of personalizing the messaging or website tweaks so the actions apply only to the specific set of people coming from a particular channel, or at a particular time. The goal is to make real-time optimization very easy and highly personal.
Bonnifield and Mayernick, who split their time between Washington DC and Mountain View, CA, met while rivals on the college debate circuit. They later went on to become close friends, and both worked on Senate political campaigns in 2006 and 2008, where they specialized in using Web optimization tactics to help increase online donations.
“When we were doing online fundraising in 2006 and 2008, there was this conventional wisdom that there wasn’t any money online,” Bonnifield says. As they, and a guy called Barack Obama, would prove, however, that wasn’t true. Says Bonnifield: “If you can unlock the small dollars and do it at scale, you can make so much more money.” The founders are now applying that experience and against-the-grain logic to the Web optimization problems outside the realm of politics.
Bonnifield says existing analytics options are too complicated and require too much manpower to be useful. “They’re forcing you to learn the rules of the game of doing data analysis,” he says of options such as Google Analytics and Optimizely. While that approach is effective for companies that can afford to dissect data, carry out extensive A/B testing over a number of months, and apply their learnings accordingly, Bonnifield and Mayernick see a much bigger potential market. They believe there is a massive opportunity for a product that makes sense of analytics without requiring human analysis and allows optimization on an up-to-the-minute basis. “There needs to be a completely different model for thinking about this stuff.”
Instead of just collecting and reporting data, Spinnakr runs the stream of visitor data through several real-time insight engines as visitors arrive on a website. Those engines then detect trends as they develop, such as the arrival of a certain visitor type, a spike in a set of search terms, or a surge in traffic originating from a particular Web page. The system then prioritizes those events and immediately notifies the site owners of the most important ones. It then lets site owners respond to the trends by customizing their messaging via an email, using a template, or making the change directly in Spinnakr’s application. The founders claim that its early tests have shown Spinnakr users frequently see a 100 percent conversion lift compared to the standard 10 to 30 percent lift achieved by traditional Web optimization methods, such as A/B testing.
For example, a restaurant might see an enormous traffic spike to its website because of a review published on the New York Times website. Spinnakr might notice that the review praised the restaurant’s soup. Given that insight, the app would send a notification to the restaurant owner that says something like, “Hey, you’re getting all this traffic from NewYorkTimes.com and it's because your soup got a great review. How about putting a picture of soup up on your website for those visitors?”. The theory is that the soup photo will help convert those Web visitors into actual customers.
A potential Spinnakr rival thinks it’s an interesting problem to try to solve but almost impossible to be done effectively. Suhail Doshi, the founder and CEO of Mixpanel, an analytics platform also backed by Andreessen Horowitz, believes that it’s hard to make personalized analytics successful because you have to know what specific users care about. In that sense, Doshi says, the Spinnakr approach is analogous to that taken by news aggregation services, none of which have been particularly successful. Everyone has different needs and so it is hard to pull out an insight that is actionable. "The problem is every business is pretty specific,” he says.
Doshi says the “radically simplified” approach compared to more complex analytics is like having a choice between Microsoft Paint and PhotoShop. While the latter requires special expertise and a significant time investment, it produces more sophisticated and powerful results. Doshi warns against the “dumbing down” of analytics and optimization for businesses. “What they want is really powerful analytics.”
Where Doshi sees a barrier, however, Bonnifield is convinced of the opportunity. “We see this market growing 100 times in size in the same way that digital and online fundraising exploded when we were working on it,” he says. The app will benefit sites that get a lot of organic traffic, and businesses that get different value from different kinds of customers, he says. While Spinnakr isn’t revealing exact numbers, Bonnifield says there are “several hundred” active websites on the platform, and it is optimizing tens of millions of pageviews of month.But will the Spinnakr approach ultimately pay off in the long run?
No doubt it's nothing that a quick A/B test couldn't sort out.
[Disclosure: Marc Andreessen of Andreessen Horowitz is an investor in PandoDaily in a personal capacity.]