Twitter’s acquisition of Summize provides a good case-study on how-to, and how-not-to, commercialize new web services. To start at the end, USV’s Brad Burnham summarizes the lesson in 9 words: “Create something simple. Let the market pull you in.”
photo by jphilipson
Much like the lesson, Brad’s summary is a simplistic presentation of a complex idea. Appreciating the lesson requires appreciating where Summize was, where they are, and where they can go.
Where Summize Was.
In their own words Summize was a “novel and scalable opinion mining and search for blogs, user reviews, and other user-generated content on the web.” John Borthwich, a Summize investor, described it correctly as “complicated.”
The service crawled social content, applied sentiment based algorithms, and enabled users to search aggregated product reviews. The technology was solid, the results were rich, and you could, for example, easily see that a specific memory reader had “0.23076923 swell opinions.”
Not only was the service complex, it was geeky.
Geeky was once a requisite when following Christensen’s playbook for attracting early-adopters. However, in this era of the web, where the definition of an “early-adopter” has changed, geeky/complex now impedes early-stage adoption.
Summize, with great technology powering a feature-rich service, could not get traction and was unable to attract additional investment. Rather than continuing to iterate an unsuccessful commercialization strategy the team made a brave but necessary move - they “repointed the technology” and launched Twitter search in March of this year.
Where Summize Is.
I can only imagine the internal debate. The decision to transition from a service that was beautiful for computer scientists, phd’s, and those interested in “0.23076923 swell opinions” to a service that was, on the outside, simplistic, could not have been easy.
In essence Summize down-geeked, stripping away complexity to wrap a powerful technology in a simple service. You couldn’t not understand what it was, what it did, how to use it. Conversational search: “see” what people were saying.
Any debate about the strategy would have been silenced once the results came in. Any guesses when Summize down-geeked?
Only the features that were required to provide core value were surfaced, others were removed; either stripped away completely, automated behind the scenes, or buried within an advanced tab. Even the advanced features that remained within the advanced tab were simplified. The sentiment index was collapsed from a spectrum of five options (requiring a 1,200 word explanation) to a simple, binary choice: positive or negative.
As the service grew, and individuals became familiar with it, Summize gradually introduced new features to the mix, adding local search and translation features after people were comfortable with the basic offering.
The Summize team, and their various advisers, executed a beautiful plan, bravely repositioning their technology in what could have been perceived as a more juvenile package. AdaptiveBlue has been friendly with the Summize team for a while and we’re thrilled for their success and are eager to watch the positive impact that they’ll have on Twitter.
Where Summize Can Go.
As users become increasingly familiar with the new Twitter search the team will be able to introduce an increasing amount of features and functionality to the mix. In an ironic twist they may ultimately end up back where they started - with a complete search and sentiment service.
Not only that but the technology may prove to be a killer feature for Twitter and also their path towards a business model by bring an advertising model to the now-web of tracking people “ideas, concepts, names, thoughts, etc.” For more on this see an excellent post on konterkariert.
Lessons to be Learned.
There are important lessons to be learned from the story of Summize. It’s a rich case-study on commercializing web services in the post-Christensen world of technology adoption.
- Be a minimalist: trade off perceived completeness in the offering for increased interaction. Reducing the friction in the interaction options increases usage. Hierarchies of choice add clutter to the UI and increases the cognitive load. Collapse offerings to a single call-to-interact. (think: Hypem’s hearting)
- Be a magician: minimize choice and simplify functionality by automating elements behind the scenes. Users will appreciate the trade-off in control for simplicity and automation. (think: Feedly)
- Be a guide dog: start with just the core offering and slowly walk your growing user-base towards a more complete offering. If the feature isn’t necessary for delivery of the core value, remove it. It only introduces additional complexity at launch and can be introduced later. (think: FriendFeed)
- Be a mirror: differentiation should not come from the presentation of a feature if it already exists in a well-defined and understood way. Catalyze the time it takes to understand what the feature is and how to use it by mimicking the presentation of the existing feature. (think this vs this)
“Create something simple. Let the market pull you in.”
