The Future of Search
I have made a few statements in the past that haven’t been far wrong from the truth, so I have something of a track record. I’m no Nostradamus by any means, but I can certainly see how some things are going to turn out based on my experiences.
For example in 2008 I noted that mobile versions of websites would quickly fall out of favour and later in the same week I proposed that people would eventually have music-as-a-service, rather than purchasing and storing music themselves.
Around the same time, I left a company who couldn’t understand that their future should be web-based – but they were putting all of their money on Windows Forms and Windows Presentation Foundation (sigh). I upgraded to HTML5 at this time too, despite everyone telling me that XHTML 2.0 was the future (chortle).
So that is my track record. Here are my thoughts on the future of search.
Search is big business. While you don’t make money directly by pointing people to websites that may answer their questions, there are clearly a lot of ways to indirectly make billions of dollars from search. The World Wide Web didn’t start with search – the original idea was to create directories for websites. The directories quickly became overloaded and hard to manage – even the largest human-edited ones. Search was something of a revolution – robot-powered indexes of websites attempting to supply results for searches.
The big search revolution in search so far has been the polishing of a series of algorithms that give people the most relevant possible results. The original search engines sucked. They relied on people being honest with their meta-tags and a lot of people were far, far away from being honest. So at the current time, search engines are competing based on relevance and it is hard to see what a new search engine could do to compete with the big guys with their massive algorithms.
I think the next disruptive technology for search will be personal eco-systems. You will be able to create priority sites that you prefer results from. For example, I would rather read content on Mozilla Developer Network (MDN) than on W3Schools, so I would add MDN to my eco-system. I prefer to read Stack Overflow’s question and answer format than old bulletin-board forums, so I would add Stack Overflow to my eco-system. I would rather shop at Amazon than Currys, so I would add Amazon to my eco-system. If I really didn’t like a website, I would black-list it from my searches.
All of this would be done with a flick of a mouse in my toolbar… or maybe just by learning from what results I select.
When I perform a search, sites in my eco-system would appear higher than those outside and black-listed sites wouldn’t appear at all. The results are relevant just like they are today, but my preferred websites get priority over other websites – it’s a system of personal relevance, rather than general relevance. This doesn’t mean a search for “replacement guttering” would only search my eco-system. The search could well determine that sites in my eco-system don’t have a relevant result.
Many people are already performing this filtering manually. If I search for an answer to a programming-related issue, I skip past MSDN, W3Schools and other forums if I can see a Stack Overflow result – when a search engine does this for me based on my preferences it will get incredible loyalty from me. That’s the future of search.