One of the things I was struck by (apart from the fact of how woefully inadequate the New York Hilton is as a major exhibition venue) last week at ad:tech was the staggering number of “Ad networks” touting their wares on the exhibition floor. Of the roughly 300 exhibitors in New York, at least a third included the word “network” in their description. From AdBrite to Zanox, everybody seems to be in the business of “connecting advertisers to publishers”. It’s clearly the business to be in; but for how much longer?
Buy low, sell high
Ad networks make their money on the margin – by buying inventory cheaply from publishers, and selling that media on at a higher price to advertisers. How successful a network is depends principally on the difference between these two prices. As with any intermediary business, however, there’s always the danger that your customers will go straight to your suppliers, putting you out of business, or at least forcing you to trim your margins until the pips squeak (any travel agents reading this, you know what I’m talking about).
So will ad networks get squeezed out? Well, it depends on the network, and how they make their money. At present, one major way that ad networks make money is through simple arbitrage – they exploit the fact that an advertiser cannot easily reach all the publishers where they’d like to place ads, and simply aggregate and repackage inventory before selling it on at a marked-up price. The ad networks are adding a little value here (mainly in the aggregation of inventory), but it’s pretty thin.
The trouble with this model is that it’s very sensitive to downwards price pressure: because the networks add little value, if someone can offer a similar solution at a better price, the only way a “reach” network (as these networks are known) can compete is on price, cutting into their profits. Plus, the margins that many ad networks charge (up to 70% in some cases) provide a strong incentive for publishers to try to find other ways of selling their inventory.
Fancy a game of risk?
As a result, most ad networks play a smarter arbitrage game: they buy inventory from publishers on a CPM (cost-per-thousand impressions) basis, which is nice and reliable for the publisher (they know for a given volume of traffic what revenues they’ll get), and then sell that inventory on on a CPC (cost-per-click) or CPA (cost-per-action) basis to advertisers. The essential feature of this is that the ad network assumes the risk of converting CPC/CPA pricing on the buy-side to CPM pricing on the sell-side – get it wrong, and you’ll end up selling inventory for less than you paid for it. But because the network is adding value by assuming this risk, it can justifiably insert a mark-up into the pricing and make a profit from the deal.
This model is more robust from a pricing perspective, but is still vulnerable, because more and more publishers are willing to sell on a CPC basis (e.g. via Adsense) and assume some of the pricing risk themselves, drawing value away from the network, and pushing down the margins. So the networks with the best prospects are those which can add extra value to the inventory as they pass it through to the advertiser.
This “extra value” comes in a number of guises, including:
- Providing access to specialized inventory (e.g. in-stream video or mobile)
- Focusing on specific verticals (e.g. travel)
- Optimizing ad delivery to maximize inventory value
- Adding targeting data to the inventory (demographic or behavioral) to maximize value
At ad:tech, many of the networks (and especially the start-ups) were presenting some spin on one or more of the above themes. All of them, however, are dependent to a greater or lesser extent on scale – the very thing that a budding ad network startup doesn’t have.
Size does matter
Ad network scale is important for a couple of reasons. First, an advertiser or agency only wants to deal with a small number of ad media suppliers (networks and publishers). So lots of little ad networks are going to fall off the bottom of the agency/advertiser’s list. If you work for an advertiser or agency, I’m sure you really enjoy the dozens of calls you get every week from network sales reps. At the moment, there’s a niche to be exploited in offering particular inventory or verticals, which can justify (for an agency or advertiser) the hassle of dealing with a small, specialist player. But ironically, the more such players who enter the market, the more of pain it is for the media buyer. And once the bigger networks start to offer these services, they’ll pick up a big chunk of business through sheer inertia.
The second reason scale is important is related to the third and fourth points above: scale enables optimization of ad delivery and the ability to enrich inventory with targeting data. The reasons are simple: firstly, if you’re looking for the best place to serve an ad, the more places you have to choose from, the better your chances of getting the best price for that ad. And secondly, if you’re using cross-network behavior of users to add value to inventory (for example, noting that people who visit baby sites also visit pizza delivery sites), the more sites in the network, the better chance you have of spotting these correlations and using them to add value to the inventory (i.e. selling pizza delivery companies ads on baby sites).
All of which means that many of the newer/smaller companies I saw at ad:tech last week will likely fail to achieve the critical mass needed to make an ad network business model work. This may be ok for them, since I’m sure every single one is hoping to build up a bit of momentum and then sell up to a bigger network, but there will certainly be losers as the game of musical chairs comes to an end – you don’t want to leave selling up too late, or your competitors will simply be able to steal your customers rather than paying you for them.
So you can expect to see continual consolidation in this industry, and a few large-scale players emerge, offering a range of inventory types and verticals, and focusing their offerings on adding value in delivery optimization and behavioral targeting. But who might these players be? This post is long enough already, so tune in for another installment where I’ll give my (not so) considered opinion on how the market’ll shake out.