A recent post on the popular weblog Mashable outlines one person’s opinion on why Google and search, as a whole, are going to lose ground over the next ten years to suggestion technology. I respectfully disagree.
At the highest level – we will work through the author’s points individually in just a moment – Google is a mixture of the smartest people alive today, unlimited money (check their tens of billions in the bank), and loose creative license. In other words, you have the smartest coders with time and money on their hands to invent what they see to be the future.
We don't shill.
Check out TNW's Hard Fork.
That’s hard to compete against, but not impossible. Historically, Google has indeed acquired small, nimble competitors in a number of areas. However, one thing has remained core to Google during the company’s history, and has served it well: search. Search, for the last 12 years, has been Google’s bread and butter – the thread that holds Google products together. Given the momentum of the larger search market and Google (be it its market share or profit), to say forthwith that Google is soon to jump the shark seems a bit daring.
In her article Yuli Ziv states that “the search process is inefficient,” and that it would be better if we could “skip [the search process] and let technology offer you the [result]… without asking you to take any action.”
I have to point out that for finding information there is literally nothing more efficient than search. With three search engines, say Google, Wikipedia (its built in search), and WolframAlpha in one action I can have nearly any piece of information that I need. I can find anything online with Google. I can find detailed reference on any topic with Wikipedia. And with WolframAlpha I can work out nearly any calculation. All for free, right away.
The article points out that if search could be replaced with “suggestion” we perhaps might be able to save time in our daily lives. Again, I disagree. Perhaps I am exceptionally flighty, but my tastes tend to change. Even if a service was created to have excellent recommendation features it would suffer out of the gate by my new opinions on both new and old things.
The computer that it could recommend to me for this coming summer might hardly be the one that I want, then. Of course, I could feed it more data over time, but that would require input. I could train the recommendation software diligently in hopes that it would help me when I need it. Doesn’t that sound like more work than just running a search when I need one?
Will Foursquare Kill Location Search?
If Foursquare is taken up by the common folk, something that I still find improbable, the article states that it could auto-recommend to me certain nearby locations. Am I downtown at “the corner of 44th street and 6th avenue? There is a nice coffee place on the block, and according to your last 10 check-ins, you must love coffee.”
Interesting, and probably feasible functionality that will be created and perfected in the next year. It in no way kills search. Sure, I might love coffee, but generally I have a specific action or location in mind. Even if I feel like having coffee, chances are that I want to go to my favorite location. What I want is its hours and a map there, something that no recommendation service is ever going to be able to intuit from my current location.
In a few words, suggestions are tools, features if you will. Search is a product, and a big one at being that. The rabbit shall not eat the wolf.
Social Matching Will Be Useful, But I Still Want Search
Continuing, the article points out that as LinkedIn becomes more intelligent, it will be able to “connect directly with people who match your business goals.” Indeed, and I look forward to it. But just what are my business goals? I have a suspicion that again they change frequently enough to be hard to suggest against in any real context. What do I mean? If I log onto LinkedIn looking for someone, I am not looking for a generic person.
No. I want that person. Again, suggestions can be trained and can be a useful feature addition to any product but are hardly a replacement for my ability to say “find this now, forget what I have done in the past.” I always try to do new things, and you cannot predict what action I want to undertake next. At least I can’t, and I’m me.
Content Recommendations Already Exist, And Have Yet To Challenge Search
Scroll to the bottom of this post and you will find content recommendations, suggestions to other things that you might want to read. Most blogs have this feature, the proffering of more content that is in a similar vein at the conclusion of a post.
All this requires is the input of a data point: the article that I read. And it can feed me similar content. Useful, and something that I like to see on blogs that I read. How could this sort of recommendation replace search? Theoretically, if it had a weighted list of my priorities for a day, my exact mood and what genre of content that I wanted consume do to it (do I want to take ten minutes off and just enjoy Metallica videos?), and any number of other things it might be able to give me useful, accurate recommendations.
Instead, however, it does not have anything more than my past history. So, what it can do, is provide rather general guesses as to what I read. As I said, useful. But I still need to search to find the exact thing that I want.
What is my ratio of recommendations to searches? In fact, its sliding away from recommendations. Now that Google is nearing nirvana in terms of its ability to give me exactly what I need or want in five seconds I have no need to let a computer flounder around with quasi-help when I can just go to the solution. Don’t get me wrong, Google does recommend search results to me based on my previous behavior. But that still requires search!
Suggestions Are Only The Core Of Your Shopping Experience If You Are Confused
If I am shopping for a computer, and know nothing about computers, suggestions could be useful. I might employ their advice to help me pick out a computer to fit my tastes.
Alternatively I could use Google to learn about the product category that I am shopping in and then use that data to make an informed and therefore accurate selection for purchase. Given that I am plunking down a wad of cash, I don’t mind a five-minute trip to Google to learn all about what I am shopping for.
I trust that quite a bit more than a recommendation feature that is probably tied to the store that I am using that is most likely programmed to steer me towards products of a higher dollar value that are higher powered than what I require.
What All This Means
Suggestion technologies, recommendation engines and the like are tools that should be built into search to make it better. They will not overtake it because they are always working with datasets that are so limited as to be simplistic. Are my Facebook Likes going to paint a good picture of me? They would not now and I doubt they every will.
Don’t sell your Google stock just yet.