Confessions of a broadcast exec: ‘If you pay peanuts, you’re going to get monkeys’

Here are the excerpts, lightly edited for clarity:

What’s the state of blind buying?
Blind buying has gone full circle, and it’s multiplied. People have removed the context from programmatic buying because they’re after the audiences, but they don’t know where the ads are appearing. What gets to me is the sheer two-facedness of agencies saying they want to support quality publishers and then just pumping all their money into YouTube, Facebook and blind networks. It’s shocking — they literally don’t know where their ads are going.

Has that gotten worse lately?
The money is still being spent; it’s just going to everyone within the agency networks rather than publishers — and marked up at incredible rates.

Like what?
Recently, I’ve seen some news audiences being sold for £18 ($23) CPMs and costing £5 ($6.40) — that’s how big the markup is. That may not happen all the time; it’s a question of where you can get away with it. “I’m using bespoke tech to create an audience just for you, using proprietary data, so adding true value — that’s why it’s that price.” It’s what I call “bollocksology.” That’s my bugbear — why are the agencies blowing all their money into this shit? Maybe they’re just fighting for their own crumbs out of digital, but still.

Where is blind buying most prevalent?
A lot of it is so-called entertainment networks — sites that describe themselves as entertainment, like mediatakeout.com, but in my view half of the content on there is salacious and the other half is pornographic. But buyers aren’t getting that kind of information — they’re just looking for people interested in entertainment.

Are people incentivized enough to change?
The problem is that once you start asking for audiences to be verified and for more filters, you’re whittling down the actual level of liquidity in the market, and that doesn’t suit anyone, least of all Google. The attitude to brand safety is a bit like putting pedestrian lights on a road after someone gets knocked down by a car. It’s not in Google’s interest to put filters in because it reduces the inventory, likewise with Facebook.

What about data quality?
We have tried using third-party data suppliers, and [our data matching] has gone from being 30 percent accurate to being wholly inaccurate — and that’s with something as simple as being male or female, never mind interests. That boils down to the recency and the controls of it. If you go and look at the big DMPs, I am single and married, male and female, I am both renting and owning accommodation — I am everyone. Because the controls haven’t been put in.

How is that?
There’s an element of people chancing their arm, particularly at the vendors, and an element of people not knowing what they’re doing, and there’s certainly an element of people turning a blind eye at agencies. They’re [agencies] also answering to brands that want as wide a reach as possible for as cheap as possible, and the clients are increasingly saying they don’t want any margin to be made — so everyone is getting squeezed, but if you’re going to pay peanuts, you’re going to get monkeys. Some brands see it as collateral damage, others won’t want to know, and others have their managers trying to justify their existence. Everyone is under pressure, but why is it OK to throw your money into a dark pit without knowing what it actually means?

P&G has cut spend because of brand safety and ad ineffectiveness.
Yes, and JPMorgan has done the same, but they’re just two brands. This problem of major brands appearing on disreputable websites has been happening since 2003.

So, what will it take to change?
Competition [to the duopoly] will help. AppNexus clearly thinks it has a chance, but Amazon is the biggest sleeping giant. Amazon could just flick a switch — it’s got your credit card data, your viewing data, it knows your interests and what you want to buy and then it has voice control and Alexa. For publishers, it would be a good alternative just as an ad server to Google — it’ll probably bring in much richer data, and it would certainly be more accurate [than Google].