A few decades from now, when someone writes up an HBS case study on Google, how will it read? My suspicion is that Google is going to treat us to a very painful object lesson: what happens to your business when you compromise the values of a beloved core product. Here's the thing: having the most complete, most accurate, most relevant search results means never having to say you're sorry. You can add ads, you can do…
This is my perception of how Google evolved:
In the early days, it used to be like running grep on the whole internet but getting your results sorted by a very relevant algorithm (page rank). Then over the years they added a lot of features that fell into three categories:
* operators to run more powerful queries
* adding more complexity to the sorting of the results (localization, penalizing websites trying to game the algorithm, ...)
* adding widgets given the query (ups tracking, weather, calculator, ...)
These were all good, and I loved the fact that you could run some very clever and complex queries to get exactly what you wanted (and the googledorks were born...).
But recent improvements to Google have all been about "guessing what the user wants" (the best illustration of that is the search as you type) and while it's hard to argue about the validity of this approach my problem with it is that it won't give me any fallback options when it guesses wrong.
To be honest, Google does a very good job at guessing what I want. But the frustration caused by the few times it doesn't get it just makes me miss the "old google". And with every new recent improvement I keep being reminded of the MentalPlex [1]