May 5, 2013

Examples of Missing the Point

Missing the PointI remember reading James' article when he left Google, but it has been a year since he left and when his blog post started marking the rounds again last month, it was picked up elsewhere in the blogosphere, it started me thinking again.

Lets look at one of James' quotes and then dig into it a bit more.

Google could still put ads in front of more people than Facebook, but Facebook knows so much more about those people. Advertisers and publishers cherish this kind of personal information...
Lets think about what we use Google for versus what we use Facebook for. To paraphrase the Jobs-To-Be-Done theory, exactly what do we hire each of these companies to do?

Google Me This...

Most often, we go to Google when we want to find out something we don't know. This is search. Google was, and for most things still is, the place we go to help us find out information we don't know. Google's excellent search algorithms seemed to work like magic. What had previously been hit or miss when searching online, so much so that sites like dogpile.com came along to search multiple engines for us at once, suddenly was far more likely to be a hit when using Google.

From there, Google started to branch out, both through innovation and acquisition. They bought companies that had innovative products which were related to their core search function or they created new products, like Gmail and Calendar, to help people make more effective use of the data they already have in their possession. We, the users (not the customers, those are advertisers who pay Google) could then search our own data.

As Google learned more about us over time, they started using our own data to help (or as some would say, creep us out) us find things we didn't even know we wanted. This sounds great, but their first attempts at this were fairly crude and often produced results that seemed strange. This mostly showed up in 'related' advertisements that seemed completely unrelated to what we actually were looking for. Over time, their accuracy increased as the algorithms matured and we users put in more of our personal data.

The problem with all of this is that while Google knows a lot about what we're trying to discover, they had a much harder time learning about what we actually do. They could use inference, looking at the behavior of a small, hopefully representative sample of people and inferring that other people with similar search patters would (hopefully) find the same results helpful. This could work, but wasn't nearly as good as first-hand knowledge of everyone.

Viagra LogoWhen it failed, it failed badly. Lets say you're a college student studying health science. You've got a report due on erectile disfunction so you spend a little time learning about it via Google. Your report finished, you think nothing else about it, until you have a first date come over to your place, ask to use your computer for a moment and sees nothing but Viagra ads around their search results.

Oops.

This is an (admittedly made-up but potentially very embarrassing) extreme example of the problem, but it does get to the heart of it... Google historically hasn't know the difference between information and intent. For a company who makes the majority of its revenue helping advertisers, such as Viagra creator Pfizer, connect with customers who want their product, high false positives resulting from a failure of intent means revenue that is not realized.

So, if Google has an intent problem, why do they think that social is the way to fix that?

Facing the Book

Mobile TargetingFacebook is, at its core, a massive store of data of intent. You check in to places you actually visit, not just ones you want to know more about. You 'Like' your favorite bands, TV shows and comments of your friends. You tell everyone what you had for dinner last night, possibly even posting a picture of it. In general, its your 'life stream', a beautiful store of actual behaviors. Its a marketer's fondest dream, mostly because they don't have to ask for it; you give it up for free.

Everything you put into the system becomes a potential vector for Facebook to use in selling ads to be displayed in front of you. These ads can be more effective than those Google serves because they relate to your actual behavior. If you checked in at a Mexican restaurant, other Mexican restaurants in the same area can assume that you are a good target for their advertisements. If you stated that you were frustrated with a particular business' service in a status update, that business could choose to not advertise to you in the future as they likely already lost you as a customer. It saves them money by not serving an ad to you.

Yet Facebook's data has failure points as well. While they may win on intent, they don't have the breadth of all knowledge that Google has amassed. Facebook's knowledge is limited to what users enter (or what they can acquire from partners), but Google's mandate is to seek out, collect and collate all the world's data. Having such a wide range of knowledge, especially when it comes the length of human history, gives them a much better perspective than what Facebook has collected in the last 10 years.

But My Point Is...

Google employs a lot of really smart people. I hesitate to even consider offering advice to them as I know that the data of their is far more comprehensive than anything to which I have access. If they feel that having social data with intent is necessary, I'm inclined to believe them. What occurs to me though is that they might be going about it in the wrong way.

Two years ago, when Google first introduced their +1 button, I wrote that it seemed to be a solution missing a use case. Later that year when Google+ the social network was released, things started to make more sense as Google was pre-populating data prior to release their platform for accessing it. When Google+ failed to gain much traction beyond the geek section that was frustrated with Facebooks repeated and egregious privacy violations, I wasn't really surprised. Having a viable social network requires a large network effect. For it to be genuinely useful, there have to be a lot of people there to use it. Since Google+ hasn't gained a large audience, at least not outside of particular social groups, it is likely that the data Google collects from their social network is only of limited use.

But they haven't given up. The last few years, Google has been fairly ruthless in shutting down products that they feel are not in keeping with their social focus. Either Google+ is doing better for them than everyone believes it to be or the company is in serious denial about its network. Its possible they feel that the network hasn't simply been around long enough to achieve the network effect it needs. It is quite possible that they are playing a really long game, believing that their success is inevitable (or at least probable enough to keep going down the road).

Searching for an alternative

Did it have to be this way? No, I don't see that partnering with Facebook, with Google reliant on Facebook for all social data to help with relevance, would be in Google's best interest. Nor would it make sense for Facebook to be totally reliant on Google as a search interface for its customers.

If collaboration is out of the question and there isn't enough room in the world for two general-purpose social networks with large enough network effects to provide relevant information in all areas, then where does that leave the two companies?

Both companies have enough smart people working there that they realize status quo is simply not an option. Facebook is content to continue to find new ways to collect and catalog their users personal data. Google is doing this as well but seems to only now be waking up to the wealth of data they have been sitting on for years, namely Android.

With Android, they have deep hooks into the OS of a device that is with you at all times. They know your location, your route, what you're looking at in each of those places, what you're taking photos of and how long it takes you to get from place to place. Facebook's new Home platform allows them to collect some of this same data, but not to the level that Google can get as the platform owner.

While Google collecting all that data takes Facebook's creepy factor and 1-ups it, the reality is that if you are willing to purchase an Android device, which is created by an advertising company, you should expect them to be capturing and using every bit of your data they can get their hands on. This is the deal you make when you purchase a Google-branded handset. If you don't like that, there are alternatives.

Google has other long-term projects, such as its self-driving cars and Google Glass that will also allow them to collect more data than they do today. These projects show different degrees of promise, but none of them will provide the mature data that Google craves in anything remotely close to a short-term timeframe.

Another option, and one Google is somewhat pursuing, is to purchase or partner with smaller firms that operate local or specialized social networks and collect their data. This can definitely help them with data focused on specific subjects, but being able to aggregate that data across disparate platforms and then correlate it all back to a single user profile is very difficult, if not impossible, given so many numerous and incompatible data formats.

There may be no good general social data source available for Google beyond their attempt to create it. Yet, it just seems destined to fail, not because it is inferior to Facebook but because it isn't enough better to get people to switch. I trust a great deal of data to both companies, likely far more than I should trust to either of them. Yet for all that data, I never click any of their ads. They make no direct revenue from my behavior, only indirect by mining my behavior and my searches for use in spotting trends.

Maybe this is where Google should have invested all that money they spent to create Google+, on improving their search and relevance scores, doubling down on search. Its not impossible, with a lot of work and a lot of partners, to infer all that social data that Facebook has locked away in its platform. The truth is, they don't have to be the best at this, just good enough at it to keep their users from moving to a competitor. But then again, maybe Google+ got them just good enough data with a lot less work. Time will tell if it was enough.

Pointing to the future

James's reasons for leaving Google are sound, given that both parties are working to achieve different things. Leaving, especially doing so in a professional manner, is the right choice when such a philosophical divergence occurs. Neither party would have been happy with the results had James remained. Here's to hoping that James has found a better home at Microsoft and that Google finds the data they're looking for, wherever it might be.