Yesterday I read a very insightful and honest article by Eric Schmidt on the Google Europe Blog titled “The New Gründergeist“. His article inspired me to write this blogpost to reflect on the problems he mentioned.
The biggest problem with the search industry is that everybody copies the existing model of Google instead of approaching it in an entire new way. Even though the existing model has been doing a perfect job for years, new problems, trends and innovation require new solutions. To solve the problems Eric mentioned, there needs to be a shift in the approach to search.
“You often hear people talk about search as a solved problem. But we are nowhere near close.” — Eric Schmidt
I completely agree with this statement. I’ve experienced the same problems that Eric mentioned in his article with search over the past years and this gave me the motivation to start a company that approaches search in a different way. Adrian Aoun wrote an article on TechCrunch called “Why The Future Of Search May Look More Like Yahoo Than Google” and I completely agree with him.
Google is great when you know exactly what you are looking for, but when searching for ‘New York’, I get the best links (according to Google’s algorithms) about the topic, without any categorization or alternatives to individual links. Maybe I just booked a flight and thus a link to Booking.com is not the most relevant for me. I rather want an overview of things I can do in New York when I arrive there, preferably categorized. Things that are fun to do are very subjective and differ for each person, so I would prefer several subjective human overviews of curated listings and pick the results of a person I can identify with the most.
The social trend of the internet has significantly changed the way we find relevant information online. Google’s famous PageRank was very effective at interpreting websites linking to each other as a recommendation of A to B, rather than individual persons sharing websites through their social channels. Websites mostly used to be a group of people working together without an actual human face and acted more like media companies rather than a distribution channel, responsible for the style and quality of content available on their platform. Identifying which information is most valuable became much more complex now than it was in the early days. Nowadays, information online is much more personal and individual websites contain lots of different styles and quality of information, because they act more like a distribution channel for individual persons than a big content producer with a consistent style and quality they safeguard. To summarize: value of content has to be identified on the level of an individual piece of content and which person created it, rather than on the reputation of the platform/website itself that published the content.
The social graph of a person is very important these days. People are much more aware of what happens on the internet these days, privacy and trust are a big thing.
“Google isn’t useful because it’s popular; we’re popular because we’re useful.” — Eric Schmidt
The ‘filter bubble’ is not the right solution to these problems. This approach creates a false sense of (social) relevancy, because individual links are ranked based on privacy-invading methods and are not ‘social’ at the core. You don’t want a stream of individually relevant links, you want more context, subjective overviews about a topic, each ranked and categorized by an actual human independent on your search history. People most close to you in your social graph should be shown first.
I believe search needs to become more human, transparent and contextualized to remain effective now and in the future. When I find a bar in New York, I also want to see alternatives, preferably ranked based on different perspectives, but a ‘related:websiteofbar.com’ query on Google doesn’t provide me with any of this. We need a crowdsourced human filter on top of Google to supply what we need in this information-overloaded social web age and get out of the ‘black box’ of filter bubbles and SEM.