Someone on a web forum I read recently asked why it is that social networks seem to have great trouble monetizing through ads. I have a pretty good chunk of experience in that area, so I thought I’d offer up my thoughts based on what I’ve seen. For perspective, I’ve done affiliate marketing on the self-serve ad platforms for the big three search engines at various points in time over the last 6 years (in fact Google recently sent me a cookbook and an apron for that anniversary, thanks guys). More recently I spent a few weekends tinkering with and a few weeks running affiliate ads on the big two social networks. I also have a couple moderately successful Facebook apps (around a quarter million installs each) that have used both ads and other forms of monetization with varying degrees of success, and for which I’ve purchased ads from Facebook and third party apps on it, so I’ve seen a bit of both sides of the coin.
Everything I’ve seen so far has led me to the conclusion that an Adsense-style model of self-serve, targeted ads isn’t going to work for Facebook and Myspace anywhere near the way it does for Google, but it’s not going to be a total bust either.
The first thing that makes me think that is that the networks themselves seem to be convinced I’m correct. Their attempts at a self-serve ad platform are at best half-assed, not the sort of thing you’d build if you expected to bring in billions from it. I remember the early days of Adsense, and they were a little rough compared to what you see there now, but still significantly more advanced and less buggy than you’ll find on the social networks today.
The first thing I noticed in both platforms was that neither ad manager works in Firefox, which is by far the most standards-compliant browser that has any real market share. Facebook won’t let you get to the ad manager once you’re in the ads and pages section, and Myspace requires a separate ad account but unless you’re using IE it didn’t even let you log in until recently. They’ve added Firefox support finally, which is one more improvement to the platform than I’ve seen added to Facebook’s so far.
Given the surge in popularity of non-IE browsers over the last couple years, you would think that if Facebook felt their AdSense-type offering had any potential, they’d go the extra mile to make them cross-browser compatible. In fact, the way web development is done almost everywhere, it usually takes effort to make a standards-compliant website IE-compatible, not the other way around. It’s doubly ironic too considering that both Myspace and Facebook bug anyone using IE6 to upgrade on every page. The message they’re sending is “Please, use a better browser if you just want to upload some photos and cyberstalk that hot girl who went to your middle school, but not if you’re actually trying to pay us thousands of dollars. Then you have to get IE6 running in a virtual machine.” I’m all for launching early and iterating, but Firefox has the kind of market share and growth, especially among the sort of people who might be placing self-serve ads, that these days means you just have to support it out of the gate.
But moving past that serious annoyance, after dusting off the old Internet Explorer and getting my hands dirty what I found surprised me. I had heard that ads on Facebook, when purchased on a Cost-Per-Click (CPC) basis don’t get good clickthrough rates (CTR), and thus generate low effective CPMs (which stands for “Cost Per 1,000″ meaning the amount you’re paying per 1,000 ad impressions). That only seem to be part of the problem though. While they certainly were well below the CTR you’ll find on Google (which is not really a fair comparison, for reasons I’ll elaborate on in a second) using the advanced targeting that both systems offer I was able to get my message out to a rather highly-focused group of people and achieve a workable CTR. It wasn’t great by any means, but it wasn’t abysmal like I’d heard and given the enormous reach both networks have it was more than enough to justify the time spent.
The real problem I found was that users clicking through simply didn’t convert. It didn’t matter what I did with the ads, on which platform, or which affiliate program and product I was using, I just couldn’t get many users who clicked through to actually buy. At one point I even included in the ad texts the name of the site they were pointing to, the item (with picture) and even the price. The clicker knew exactly what they were going to see when they landed. For instance in one ad I targeted a book at people who had listed themselves as fans of the author. My ad said something like “Buy Book X, the latest by Author Y, at Amazon.com for $15.99″, and included the picture of the book taken directly from the Amazon page. There was no ambiguity at all.
And yet the vast majority of clickthroughs didn’t purchase. After many iterations of the ads, drilling down the demographics and targeted keywords, improving the text to optimize for conversions rather than clickthroughs, I was able to get to a point where I could break even at a certain CPC where, if I were able to purchase the same ads on Google (with vastly less demographic targeting) for the same price I’d be a billionaire. To compare I ended up running them on AdSense and paying almost 20 times the CPC, while making about the same $0 in profit doing so on a much higher volume (due to getting 3-5x the CTR).
Unfortunately, on the social networks, by the time I had gotten the conversion rate to a point where I was breaking even on a given CPC by adding more granular targeting, my CTR had dropped by somewhere between 20% and 50% depending on the product. That meant that the networks, which of course optimize for their own profit, effectively stopped showing my ad. They show the inventory with the highest effective CPM (eCPM) first, because that’s what they profit most from. eCPM is simply the product of CPC and CTR, times 1,000.
eCPM = CPC * CTR * 1000
So you can see that if your CTR falls by 50%, your CPC has to double for Facebook/Myspace to make the same eCPM. By the time I had gotten to where the ads were breaking even (still far from profitable enough to be worth running) they were barely being shown. They seemed to at best run late at night when presumably all of the higher-eCPM ads had already hit their daily budget.
Of course both Adsense and the social network platforms are markets. The fact that I wasn’t able to get ads shown at the CPM I needed to make a profit indicates that someone out there is making a profit selling something. Simply using the social nets and looking at the ads gives you a good idea of what and who, and using the ad managers gives you a pretty good idea of what they must be paying. As far as I can tell, companies that make novelty t-shirts, get-rich-quick schemes, and dating sites are buying users at somewhere around 10-20 cents eCPM on Facebook.
On Google, I don’t even think you could get an ad to run at 10x that price, which is why I just don’t think the social networks ar ever going to get high eCPMs from self-serve ads. The people we targeted when running ads for Draftmix on Google were pretty much identical to the ones we’ve targeted for our apps on Facebook, but they cost us over 50 times as much.
Why do users on social networks click fewer ads, and convert less frequently when they do? The answer to the first part is obvious. When you show an ad on Google, you’re showing it to someone who wanted to find what you’re telling them about. It’s like an ad in the Yellow Pages. The only people seeing it are more or less specifically looking for it. If you’re running an ad for the latest book by J.K. Rowling, you can just show your ad to people who searched for “new Harry Potter book” and you’re going to get ridiculous click throughs and conversion rates.
People on Facebook, on the other hand, are just hanging around, using stupid apps that tell them which state they should move to, and leaving wall posts on each others’ profiles. When you’re advertising there, you’re like a life insurance salesman trying to approach people at a bar on Friday night. All of the demographic targeting in the world won’t make up for the fact that they aren’t looking for what you’re selling, and in fact aren’t looking to be sold anything at all.
The second one I’m not sure about and I found much more surprising. What compels someone, who clearly has little or no interest in a product, to click a highly descriptive ad for it and then not purhcase? Your guess is as good as mine. Until the social networks figure that one out though, they’re going to have to find a much better revenue stream if they want to be the next Googles.