2 Things We Need For Better Display Advertising Relevance

a B 52 bomber dropping its payload

Carpet bombing is still the default behaviour of corporate advertisers, since it is easier to increase volume than accuracy

This morning, I've been targeted with display ads for a conference in Glasgow, enterprise fintech software, a romantic novel, accountancy training, and about 30 more things I don't even know what they are because I closed or flicked past them so fast.

What these ads have in common is their irrelevance to me.

Irrelevance is a major headache for display advertising, and always has been.

What has changed in digital media is that better targeting is becoming available. A large part of problem is that lazy advertisers are not taking advantage of the better targeting, and instead just spending big bucks on reach.

But the bigger part of the problem is that advertising networks are not taking relevance seriously enough to steer default-lazy advertisers towards better targeting.

As long as advertising networks make it the default and easy approach to advertise with mass undifferentiated reach, instead of proactively defined targeting, then they are effectively encouraging ad irrelevance.

In an obvious way, it is in advertising networks' interests to encourage large untargeted ad campaigns, because large reach is associated with large spend.

However, it is not in their interests, or advertisers' interests, or publishers' interests, over the longer term if this encourages ad irrelevance.

The rise of ad blocking is a direct response by rational people to the callous irrelevance of advertising.

Irrelevance encourages ad blindness and ad blocking, which erodes then destroys the ecosystem of advertisers and publishers. Ultimately this is not good for anyone because advertising, if done well, is a virtuous or at least innocuous way of spreading the word about things that make people's lives better. (Making people's lives better is presumably the only reason you would be in business to sell stuff, and the only reason people would buy your stuff.)

So to save display advertising I think there are two things we need:

  1. Incentives to target better;
  2. Disincentives to increase reach blindly without targeting.

Currently none of the major ad networks provide any automatic incentives or even informational prompts in either direction. e.g. Facebook "Boost this, boost that" -- and you can definitely go ahead and boost stuff without having the slightest clue to which people you are boosting it.

To some extent the market helps here, because increased competition drives up costs and increases the pressure to perform. However, as long as advertising networks like Google AdWords not just permit but encourage lazy broad-reach campaigns, the method of improving performance is going to be via increased spend, instead of investing effort to target better.

The advertising networks need to find ways to nudge advertisers to understand the value of targeting and thus invest the resources to work on it.

In this respect Google is making small steps in a geeky kind of way: just adding more targeting options over time. This doesn't actually help to promote the use of the targeting though. Anecdatally, Google's display inventory must still be really undifferentiated. This can be seen in the lack of specific enough (and enough specific) topic/interest targeting audiences, the lack of accurate location and demographics, the lack of transparent managed placement options, and the frequency with which you get the "too narrow" warning when only mixing a couple of targeting dimensions.

Advertising networks should not warn you about your targeting being "too narrow". The opposite: they should warn you about being too broad and spammy. Right now, no system does that. Spam all the people.

I also think Google's custom affinity audience building is completely obscure. Overall, again anecdatally, the way the Display network performs craply compared to Search just shows they are not facilitating strong targeting. The rush to remarketing is evidence of the need for better audience definition.

Facebook is winning in targeting accuracy because of its profiling of users, but compared to what we all need, is still crap at classifying people's interests well enough beyond mainstream topics like sports teams.

Twitter ad targeting is nowhere which is why the ads are universally crap and the users hate them. An intelligent 7 year old could suggest better targeting options for Twitter than they currently offer advertisers.

Until the networks improve, what do we do?

What can businesses and advertising agencies do? Firstly just keep our fingers crossed that these giant highly funded corporations full of PhDs finally get round to developing ad targeting based on the insane amounts of data they are hoovering up... which after all is their raison d'etre.

In the meantime we need to use the available tools as accurately as their current targeting allows. This means combining targeting dimensions including demographics, location, context, and audience profiles.