Just how skewed media planning has become was recently brought home to Jason Brownlee, the founder of UK audience research and consumer insights company Colourtext.
His daughter landed her first proper job in marketing, working for a global media agency. “Pretty much from day one,” he says, “she was set to planning Google pay per click campaigns.”
That’s very different from when Brownlee started out in the 1980s, in the Messenger newspaper group. While there, he began learning about broad reach TV and radio channels too which, by their nature, were brand rather performance focused.
“My daughter’s experience was completely the other way around,” he tells Dave Winterlich. “It was all about, ‘What’s the ROI? (return on investment) What are the conversions?’ Basically looking at the world through the other end of the telescope.”
READ MORE
As more people come into marketing this way, the impact on brand building is stark. For a start, they get used to the data that flows – almost for free – out of the performance channels.
“It’s easy to get hold of and easy to use for things that the performance channels want to be measured on, which is very often attribution,” he explains.
We’re now in a situation where businesses are increasingly born online too, often selling direct to consumers and typically starting their lives by reaching out through Facebook, Instagram or Google.
Because all they know are these platforms and the easy metrics they provide, they find it harder to figure TV, radio or outdoor, places where “they just don’t recognise the data landscape at all,” he says.
The result is that “those channels that we love and appreciate for their ability to build brands and market share just don’t get represented in the data picture, and that is a strategic issue,” he explains.
It’s all about attribution, something which really took off in the early days of direct mail campaigns, and moved online with ease, providing a direct link between offer and action, whether sales or conversions, which typically spiked in minutes.
But with broad reach media such as TV and radio, attribution is much more difficult to assess. He points to the “offline attribution trap” arising from research by Radiocentre, the UK industry trade body.
It’s surveys found advertisers that shied away from radio advertising often said they had tried but it didn’t work. But when Radiocentre dug deeper, it found that what advertisers actually meant was that their attribution platforms didn’t pick up results.
That compares to a situation where they can say, with ease, that “for every quid we put into Facebook, we get X back – and can see it,” says Brownlee. “So they took the money out of radio and just put it back into Facebook, and that was the trial.”
Even though uplift studies show time and again that radio is contributing, it doesn’t show up in attribution data. It doesn’t help that consumer responses to radio and TV advertising are rarely instantaneous. For a start, they’re most likely driving a car or making dinner, and unlikely to interrupt themselves.
Research shows that just 23 per cent of the response that an average radio ad generates occurs in the first hour after exposure. Only 8 per cent happens in the first 20 minutes.
Because attribution models look for an almost immediate response, that means that at least 92 per cent of the response radio is actually generating is missing.
Worse still, when consumers who heard the radio ad are ready to buy, it’s the performance channels they turn to, which then get the credit for what radio has actually done, simply because they are closest to the conversion.
“We thought that if this was happening on radio, we bet it’s happening on TV,” says Brownlee, who conducted the research with TAM Ireland, which oversees TV audience measurement and ratings. By studying TAM data along with other data sets, and examining the experiences of 12 brands, it obtained a clear picture of what he says is referred to in the industry as the “Google tax”. “It’s where television is doing the hard part for a brand, building familiarity, putting the brand into people’s head and nudging them close to some kind of action, or basically remembrance,” says Brownlee.
Pay-per-click advertising is ‘probably being credited for, and taking the benefit of, things that it hasn’t actually stimulated’
— Jason Brownlee, founder of Colourtext
At the same time, Professor John Dawes of the Ehrenberg Bass Institute’s famous 95/5 rule applies. It says that of all the people that might conceivably buy a product in a particular category, only five per cent of them are actually in the market, ready to buy, at the moment when they experience your ad. That’s because they may only buy once every six months or every three years, depending on the category.
But it’s during these in between periods that TV is building up mental availability and favourability towards your brand. Then, when somebody finally gets around to making that purchase, what do they do? “They type the brand name into Google, click a paid link and the search platform sends a bill to the advertiser for that click, but it’s the TV campaign over all those months and possibly years that has been doing all the work,” says Brownlee.
So the brand is paying twice: “First for the TV advertising, which created the interest in the first place, and then again for that paid click that captures it at the doorway.”
He likens it to nightclub goers having to first pay the bouncers and pay again at the door.
“That’s not to say that pay-per-click is bad. It works really well. But it’s probably being credited for, and taking the benefit of, things that it hasn’t actually stimulated, and putting a little bit of tax on every brand, because the brand has already done the work,” he says.
Depending on the category involved, anywhere from 3 to 12 per cent of what TV is doing is being misattributed to pay per click, he reckons.
That matters. If one medium is under credited and the other over credited, media buyers will withdraw from the first, a problem that compounds over planning cycles.
It also ignores what TV’s “persistence dividend”, the longevity with which a brand is remembered after a campaign. Yet, again, none of that is captured by current attribution models.
That’s no surprise really, given that those models were designed by the platforms, to serve their own ends, says Jill McGrath. TAM Ireland data, by contrast, is built on the basis of measurements agreed by all stakeholders. But is very clearly not an attribution model.
“Because, what is marketing about?” she says. “It’s about building familiarity and building that reach at scale over time, consistently talking to your audience to build that relationship with them.
“Those are the metrics that really matter and they are getting a bit swamped by the shiny metrics that come out of, and are created by, platforms like Google and Facebook, which of course they are created to self-serve.”
As Brownlee puts it: “They’re marking their own homework.”
Discover more Inside Marketing podcast episodes at irishtimes.com
