Now the dust is settling from Yahoo’s recent purchase of the news aggregator app Summly, the real picture is becoming clear: Yahoo CEO Marissa Mayer is making a big play on the future of search, and it looks very different to Google. In fact, it seems to be based on the next wave of natural language processing from SRI, the creators of Apple’s SIRI.
Let’s be clear about a few things: Marissa Mayer knows a lot about search, having previously been Google’s VP of Search Products and User Experience before becoming VP of Local , Maps and Location Services. In short, these are probably the parts of Google the average user touches most, and in particular the areas where brand loyalty has been strongest.
On the surface, the purchase of Summly doesn’t make much sense. Summly runs on Natural Language Processing, which means that it scans articles to understand their content in order to do its little trick of summarising them; but it didn’t develop much of this technology itself. Also, the app doesn’t seem to have many happy users (or indeed, users at all: usage figures suggest about one news story was read per download per day between November and March) given its one million downloads, which implies either that it was downloaded by robots, or that the app isn’t spectacularly useful. It also didn’t have much revenue and the team didn’t even build the app; they outsourced it to Somo, a marketing agency. In fact, it has been suggested that the purchase was simply a way to get on the good side of a powerful investor by giving a disappointing startup a soft landing, or a way to get Yahoo a lot of earned media coverage.
The picture becomes clearer when we consider where Summly got that piece of natural language processing: it was licensed from SRI International, an organisation born out of the Cold War need for new technologies and which got into natural language processing to help the CIA pick out the most important parts of Soviet documents. More recently, SRI International created a personal assistant app: SIRI. Even more recently, it appears Yahoo struck a deal with SRI International to license some of its technology, and the acquisition of Summly formed a part of that deal.
An interesting twist in this story is that the Summly app appears to have been a publicity stunt with long-term celebrity backers and publicity for an app that wasn’t actually that widely used. As consumerisation becomes increasingly important in IT, especially mobile, it is hardly surprising that the professional backers of the Summly experiment would focus on creating a personality-based brand to help promote the eventual product. It certainly doesn’t harm Yahoo’s attempts to appear as a more youthful, dynamic company either.
So, why did I say that this purchase is a play at the future of search? Yahoo is unable to compete with Google on its own terms: with our short attention spans and ever-increasing information overload we will use the tool we know works best, and it’s no accident that when we want to know something, we Google it. Marissa Meyer will know all too well how effective Google’s search tool is, so instead of trying to overhaul Google (especially considering Yahoo’s search currently runs on Microsoft’s Bing), it makes sense for Yahoo to sidestep it and take a completely different technology path. She is transforming Yahoo to become a mobile company, and intends to corner the market on mobile search.
Google runs on BackRub, a very clever technology which allows it to judge a page’s usefulness through examining its popularity and how frequently it has been referred to. This technology is at the core of Google’s search and has been a key factor in its success. Now, it seems Yahoo is going down a very different route of providing search by summarising webpage content.
The ultimate challenge facing Yahoo in this field is to make search more useful than Google can; and in order to do that it needs to be able to take natural language search to the next level. Content is easy; true value for the user in a hurry on a small screen lies in context.
That’s why I say this is the future of search, and in particular search on mobile. On desktop, we have the screen space to browse results, the bandwidth to use features like Google’s page preview, and usually we are sitting in comfort when doing so. On mobile, the picture is different: we use mobile browsing because we need information now, where we are, and usually we either worry about our data tariff or our data speed. We also have a small screen, which not only makes browsing results a painful experience, but also makes it harder to validate that we have the correct result. So being able to understand what the search engine is looking at, what you’re looking for and presenting you with that really would be a mobile killer app.
I wrote recently about Google’s business model being to get us all using Google tools, and again about Samsung trying to reduce its reliance on Google by developing its Tizen mobile OS. Since then, rumours have surfaced of closer integration between Yahoo and Apple (another company trying and struggling to reduce its use of Google services, such as Maps). Of course, information from Yahoo is already fed into some iOS apps including Siri itself. The Yahoo’s current use of Bing to underpin its search engine makes that seem unlikely, but what if Yahoo is able to create a context-aware search engine that understands what you are looking for? Then we really will be living in a science-fiction future.