Facebook Beacon: Not what I signed up for

Finally got a taste of beacon yesterday. I bought a ticket at Fandango and it was posted in my Facebook newsfeed — fb_evil.jpg you can see the image to the left here. No opt-in, no notification on Fandango, no notification from Facebook that beacon was enabled — it wasnt what I expected — if you want to see a play by play of what is meant to happen, its here). I felt there was a violation of commercial trust unlike I had experienced in a while. Yes I get spam, targeted behavioral ad's but this is both more explicit and more of a violation of the compact that I thought I had with Facebook. I find it astounding to think that a site believes it has my permission to gather and use this data on a non aggregated basis. Age, sex, location — thats one thing — but a purchase that is then explicitly broadcasted out to a group of people, some of whom i know, some of whom I wish i knew and some of whom wish they knew me? If I hadn't found beacon blocker to be as easy to setup I would have closed my Facebook account.

What's the context here? The Facebook needs to rapidly figure out how it can earn an operating margin that reflects the way it is scaling its audience growth. They need to do this to justify the valuation that MSFT pegged on the company. And if private equity investors are going to invest they are going to need to see the path to torrid margin growth, and cash flow — even if this investment is purely financial (eg: at a valuation that is discounted to the strategic MSFT round). Facebook is supposedly making money today but there are two important caveats that add up to supposidely. First, most of the money Facebook is making is from an advertising agreement that blends rationale economics with MSFT willingness to pay a premium to Facebook to keep them out of Google's arms — in English — the advertising deal they have isn't market priced, the investment demonstrated that fact. Second, even if the advertising deal was made up of rationale / extensible economics — advertising using a third party network, doesn't have the kind of operating margin that Facebook needs to achieve to support a market cap of $15BN+.

Facebook is in experimentation mode — and beacon needs to be placed in that context. Nine months ago Facebook launched its digital asset business — that was meant to be a high operating margin, revenue stream. Digital assets sales didn't take off. Then came F8. F8 and the platform initiative has been a huge boon to web 2.0 companies looking for distribution — but as yet, it is not a business. Now we have social ad's and beacon. Both are meant to lead Facebook to the promised land of Adsense like operating margins. My guess is that they will fail and Facebook will have to try again. Not to say there is anything wrong with testing radical new approaches to monetization. Its very necessary — and its brave and unusual to see a company of this scale figuring out its business in the public eye. I think Facebook's market position is fascinating — not since Netscape have we seen a company with this scale of audience without a scaled business model to match. But unlike Netscape — Facebook has at least $250M – $500 of capital to invest to figure its business model out. They have taken the VC model and turned it upside down. They have sold strategic alignment for cash and a stub of equity — buying them time to figure out where the business is.

My bet is the beacon backlash will be severe. The service is presumptuous at best, it lacks transparency and its poorly executed (ie: Facebook's web privacy setting states "…you will still be notified on affiliate websites when they send stories to Facebook. You will be able to decline individual stories at that time." This is not what I experienced). And even if they navigate the backlash they will likely be forced to make it an opt'in feature – rather than opt'out. As a feature that you need to opt into beacon will fade — users wont be navigating Facebook's frustratingly translucent privacy settings to turn beacons on. I wonder if Facebook's platform effort might have given them to the idea that they have permission to launch beacon. This would be facinating — they seem to have bought into their own marketing pitch that they are a web wide platform, missing the fact that they are a site and the web is the platform. I dont believe they have users implicit consent to collect and explicitly serve up this data — I certainly haven't. The backlash that is brewing online is, I believe, very different to the backlash they experienced when they launched the newsfeed. Along with users, MoveOn is at the helm — they can marshall resources and attention in Washington and the EU — this is very different to the newsfeed.

But to run out a counter point — what if beacon is accepted — everything else aside is beacon actually gathering useful information? Like Facebook itself, beacon, is gathering chad's of people lives in a manner that is machine readable but of unclear relevancy. Facebook needs to broker partnerships with all the retailers who are make up my commercial relationships — my commercial graph – and like Facebook's social graph aspires to, it needs to map how relevant these relationships are to me. This is where I think the Facebook model is most flawed. When I friend someone on Facebook I have to put them into one of 14 categories representing how we connected. This might represent 14 categories that can then be mapped across the social network but its only a few steps more interesting than Friendsters rating of whether you were a friend or not. Its not how human relationships are mapped. The dynamic nature of human relationships is reduced on Facebook to a narrow, stagnant handful of categories in order for them to be fully accessible and managed by the code base. Dumbing human relationships down so that they can be more effectively managed by machines is problematic. Its almost back to the the cathedral and the bazaar — computer centric experiences vs. human one's. And what about the data — these commercial chads of data — two counter perspectives. I was reading an interview with Craig Venter this weekend. He talked about how the six billion letters of his genome are now mapped but that the only useful conclusion thus far is that he should be taking statin drugs (fat lowering drugs) a conclusion that could be reached with far less data — by simply looking at family history. Are these beacon data points actually representative enough to form a data set that will enable behavioral targeting? In discussing behavioral targeting with the founder of Tacoda last week he explained how second order data was often more predictive of behavior — not first order. I wonder how predictive and useful these data points would be anyway. Over to Twitter — compare a Twitter stream to the Facebook news feed. Facebook's news feed feels like media, its processed, organized and polite. Twitter is messy, its all these tiny little specs of people's lives in a stream of data that offers a surprisingly human feel — Facebook has processed, digested and organized — Twitter is a steady stream of chatter, some of it useful, much of it messy, but all of it very human. Similar story with Fotolog — it has none of the computing cetric categorization that dominates Facebook — the site is messy and hard to navigate if you are a machine, but for humans its seemingly a pleasure — average time on the site is now 25 min per user, per day. These are disctinctly human experiences online — I believe we need to foster these experiences and get machines and code to be smarter about us, not the inverse. There will be more — not less — to come on the subject of beacon. Oh, and btw Beowulf was good — worth seeing in 3D.

Trackbacks

  1. [...] Here’s one report about none of the above opt-out options working. It’s not clear why this person has had a [...]