Facebook, Twitter and the economics of attention


Ever since Facebook went public in one of the largest technology stock issues of the last decade, an offering that valued the company at more than $60 billion, one of the biggest questions hanging over the company’s head is whether it can generate enough revenue to justify this massive market value. There have been questions raised about both the lack of a strong (or at least convincing) mobile strategy and the relatively lackluster performance of Facebook’s advertising model. And some wonder whether the giant network will ever be able to make advertising work at all, because of the nature of the network and the purposes for which most people use it — and whether alternative networks like Twitter aren’t better equipped to do so.

One of the more recent discussions of that idea comes from startup advisor and entrepreneur Alistair Croll, who argues in a recent blog post that despite its enormous size and reach — something that should theoretically make it perfectly suited for advertising — the nature of Facebook’s network actually makes it less effective at doing this than Twitter. In a nutshell, Croll says, we are more likely to be open to commercial messages appearing in our Twitter stream because the network is asymmetric (meaning you don’t have to approve every follower, and users can send you messages by simply using your Twitter name).

Is Facebook’s structure even conducive to ads?

Facebook, by contrast, is based on a more symmetric model — one that is designed to allow users to control their social network at a very granular level, by approving each and every “friend,” and therefore implicitly accepting that messages from them will appear in their timeline. The network has added subscriptions and other features to try and duplicate the asymmetric model that Twitter uses, but those don’t seem to have really taken off for most users, many of whom complain about subscriber spam. And that kind of structure, Croll argues, makes it a particularly ineffective medium for advertising:

At its core, Facebook is also built around the idea of a network of friends, which means some people (and their messages) are foreign to that network. Our “immune system” reacts when someone promotes something to us. Our first reaction is that their computer has a virus and it’s spamming their social graph —- not that they might be genuinely recommending something.

This is a point that we have made in the past as well. And it isn’t just a threat to traditional advertising, the kind General Motors was doing when it decided to yank its $10-million budget just before Facebook’s IPO (the company is apparently reconsidering this move). It’s also a potential risk for the kind of social advertising Facebook is trying to sell to big customers, including features like “sponsored stories,” which embed messages from Facebook users inside ads. As I tried to point out in a recent GigaOM Pro report, the network’s future depends on it being able to solve that problem (subscription required).

If Facebook can’t make ads work, then it is AOL

Sir Martin Sorrell, the chairman of giant advertising and marketing conglomerate WPP Group, put this challenge well when he said recently that he wasn’t sure whether advertising would be effective on Facebook or not, because it is a social place where users expect to interact with their friends, not with commercial messages. As he put it: “The point is that Facebook is a social medium, not an advertising one… you interrupt social conversations with commercial messages at your peril.” As Croll notes in his blog post, this could be one of the reasons why Twitter might actually be better suited to advertising content:

Facebook is stuck in an awkward place compared to LinkedIn, Twitter, and even Google. It’s not a “vertical” social network like LinkedIn. It’s not an “open” social network like Twitter [and] it’s not a default part of using the Web like Google is.

Obviously, Facebook is far from doomed when it comes to either advertising or generating revenue. One of the few advantages the network has — apart from just its massive size — is the “open graph” platform it has created, which connects it to millions of websites through “like” buttons and comments and other features (something Google has been trying to replicate to some extent with its Google+ network and the +1 button). There is at least the potential that Facebook could use these connections to develop an advertising network that places ads on partner sites and targets them using the behavior of its users, avoiding the anti-social problem.

The big danger is that Facebook never makes this or any of its various mobile efforts fly properly, and winds up being just an AOL-style platform that “tries to keep people inside it, making money off the in-system credit taxes,” as Croll puts it in his post. It may have a few billion dollars that it can spend on figuring the problem out, and a billion or so users on whom to test various solutions, but there’s no guarantee that it will succeed — or that success is even possible without radically changing the nature of the network.


former facebook user

the recommendation model will be as poor as the advertising model.
facebook encourages homogeneity, therefore the like bottom just like the want bottom will be at most a decent customer retention tool, it does not bring new costumers.
There is (ironically) a good research about the lack of effectiveness of the like bottom, done by some Harvard Researchers.
If you consider the active user’s base we might find out the facebook is actually loosing a lot of users.


What do you guys think about having a premium subscription, in which no advertising is one of the features? Most people know you don’t get something for nothing. It would have to be cheap, though, like 4.95/month or something. Amplifying recommendations is good, with links to other websites for the advert.

Anuj Agarwal

Hey Mathew, do you think Facebook should charge companies using their open graph platform. So many companies have acquired millions of users using facebook platform for literally zero cost. Do you think this this could be additional business model for facebook?

Mark Rogers

This analysis is spot on. On Facebook the size of the user’s graph (the group from whom they see updates) is constrained by this need for mutual validation. It averages 130 people. These will be friends, colleagues and those with whom I share a common interest. I don’t see any updates about my area of common interest because:

a) Facebook doesn’t optimise search [epic understatement, the search is horrible] meaning it is hard to find and subscribe to themes;
b) Knowing that Facebook isn’t a great place to reach people with a particular interest, the best content tends to sit elsewhere (Twitter for example)

You are unlikely to reach those with whom you share a common interest via your timeline. For example, even if you have one friend who is a cyclist, they may well not be online when you update. The nature of viral distribution (if it is to work) demands that you reach not one person but many. In fact academic research on Facebook shows that ideas are passed on in multiple short chains, between two or three people. This normally happens as a result of external information. By external information I mean anything happening outside Facebook. What has in the past triggered “Facebook” virals is mainstream media coverage amplified by email or Twitter happening outside Facebook.

You could put this rather simply: Facebook is not the best place to reach people on Facebook, Twitter is.

There is more on this here:


and we summarised it just before the floatation


We are publishing a paper evaluating the effectiveness of Facebook for viral diffusion later in 2012 in the Institute of Direct Marketing’s Journal of Direct and Digital Marketing.

Three things Facebook does work for:

– Community support. CRM works well on Facebook. Brands can reach an engaged community and communicate efficiently with them.
– Entertainment Media distribution. Facebook has had real successes distributing games like Farmville and movies like The Dark Knight Returns. The ability to serve media to people whose social graph you can mine is extremely valuable (though still largely untapped).
– Local business outreach and advertising. Local business is one area where Facebook’s limited but trust-based graph works in its favour. If I have a great experience in my local Pizza shop, my recommendations are likely to find others in my graph who are my neighbours. They know me and they know the Pizza place. There is some anecdotal evidence that local Facebook marketing is more effective than big brand marketing.

We have evidence and references for the above if anyone needs to know more @sentinelproj


To me, the interesting question is why people go to Facebook..It’s not to look at ads, it’s to socialize with other people. Unlike magazines, where people were passively reading stories and many are highly interest specialized so targeted ads can work, or even TV where, in the past, attention could be interrupted by ad, this doesn’t happen on social networks. If Twitter sticks ads in people’s way, then, like TV, people will learn to ignore the ads over time. If Facebook does the same, people will ignore them as well. This, at least, is my humble prediction.

Troy Morris

The solution isn’t difficult. It comes from leveraging that “weakness” as their strength by removing advertising and amplifying recommendations from friends and finding a way to monetize that. It’s why Sponsored Stories do better than an ad and why the rumored “want” button is valid.

Erica Wissick

I agree with amplifying recommendations from friends, which would keep the inherent social network structure intact. The “Like”buttons could play a key role in the recommendations, and having them displayed as a second “wall” which can be separated by a “tab” but displayed similar to a Pinterest model, I think would be very effective. The solution is giving the user’s friends the control of advertisers.

Comments are closed.