Graph Search and the Like.

The question the new Graph Search at Facebook continually elicits in me as I've discussed it on various lists, as well as read a couple dozen articles on it, is:

Do I really need my whole graph to find what I need?

First.. how much and what do i need?

Advertisers, marketers, graph search makers, all operate on the assumption that we all need much more than we do.. and if the last 5 years had taught us anything, it's that a lot of people borrowed a lot of home equity to buy crap they later dumped at Goodwill..

In other words.. yes.. we do need some things, a plane ticket, rental car.. a new laptop.. etc. But I do think many know how to get those things.. without necessarily getting all that much input from others.

And that leads to my other point: how many others do you need, and how much of their input?

This weekend I had a guest here.. who rented a car from Avis.. and it's the third time she's signed up for the lowest level car and then been given a 3-series BMW or a Mini.. for $25 a day.

That's a nice to know factoid.. but if everyone coming to SFO knew it.. she would never get a BMW for a tin can on wheels price.. we talked about whether she would share this anywhere.. and she said no.. she would not share it. Though she's very active on many social communities.

Another angle: about 7 years ago, I was in a book club with Jerry Michalski and about 5 others.. and we would read books on ants and viruses and ecosystems.. trying to apply those understandings to what was going on online.. we did it for a couple of years and it was very helpful.

But one of our conclusions after talking through two dozen books and working through the logic of different takes on systems and people and flows of information was that in the end, you only need the right 5 people to help you find the things you need, get the right ideas, advice, etc to make good choices.. and these were verbal conversations because most often, even if these people were highly active online, they wouldn't necessarily share certain information online, for various reasons (it took too much time, there were consequences for having those opinions, they didn't want to be bugged, etc). In fact, much of the time the good intel didn't make it to the searchable web for months or a year or two later.. and I still find that true today, even with Twitter, FB, quora, tumbler, etc. People who really know stuff don't want or need to show it off.. and there is downside for sharing the data.

So these questions linger for me.. as I think about Graph Search.. which may have some value.. but I am highly skeptical of what, how much, etc.

There will be some value.. but I think maybe it will be comparable to the kind of "lift" that an Ad gets, when some new technology is added to the Ad selection or whatever.. often that lift is just a couple of percent better than before but to Ad people.. that's great.. because they are doing something at scale.

For us.. for individuals.. if Graph Search got us 10% better intel over what we could otherwise find using existing search systems.. would that be worth the increased personal exposure and loss of control over our data we give away in a system like this...

And lastly, I'm skeptical because I do believe Facebook's biggest issue is trust -- people withhold information intentionally. It's not a safe place and most people know it.

Graph Search makes Facebook a lot less safe. Which leads also to the question: do I need to know who in my graph likes something salacious? Really, does this help us develop better relationships or just make our current relationships a bit more unsavory?

So if people search, see what's exposed, and cut down their sharing even more, then the effectiveness of Graph Search goes way down. That 10% bump in quality information you got with Graph Search could turn out to drop 20 points.. you might find that you have -10% quality over your search results compared to before Graph Search.

I think Graph Search will only work when we have Personal Data Stores, and can set terms for use of our data, and then our friends can search our non-public, but friend-shared information, without fear that a company like Facebook will sell us out.

Until then, I'm very skeptical of Graph Search at Facebook, other than as a model for the sea change to come where we will drive our own data and interactions, and treat Facebook as the bar or restaurant it is, where I would most definitely want the in-person protection of clothing. As it stands now, we just got more naked in Facebook, which doesn't deserve to also hold our personal information the way it does now (leading to our naked state there). It's just a Cheesecake Factory online, but most people don't see that yet.