Sponsored by

Spending less to achieve nothing is not a saving

By

4 minutes

We have become extraordinarily good at optimisation, but we have not always been clear about what we are optimising for.


Car dashboard


The industry has never been better at optimising media. So why is advertising effectiveness going backwards?

Here is a puzzle worth sitting with. UK adspend keeps climbing, and our optimisation tools have never been more sophisticated.

Yet, as Omnicom Media UK chief strategy officer Charlie Ebdy told The Future of Media London recently, average advertising effectiveness is heading the other way.

The usual explanation is audiences: scattered across more screens, giving each ad less of their time. True. But some of the decline may be self-inflicted. We have become extraordinarily good at optimisation, but we have not always been clear about what we are optimising for.

The fuel economy problem

Imagine a car with the best fuel economy on the road. Every journey, the dashboard glows green: fewer litres per mile than ever. Brilliant. Except nobody checked the satnav, and you are heading very efficiently in the wrong direction.

That is a fair description of a lot of media optimisation. We optimise towards the numbers we can see most easily: cost per thousand, viewability, click-through, completion rates and, increasingly, attention. They arrive daily, they are tidy and they make for a reassuring dashboard.

None of these numbers are useless. But they tell you what a campaign cost and how it was delivered, not what effect it had on people: whether anyone became more aware of the brand, more likely to consider it or more likely to choose it.

Economists have a name for what happens next. Goodhart’s law says that when a measure becomes a target, it stops being a good measure. Ask a buying platform for cheap impressions and it will find them. It just won’t ask whether they changed anyone’s mind.

The platform that clocked off at 8pm

Here is a real example from our data, with the numbers anonymised. A UK business-to-business advertiser ran a connected TV campaign across six platforms. Same creative, same week, same audience. On the usual efficiency measures, nothing looked wrong.

Then the team looked at the outcomes: how many people each platform moved towards considering the brand, and what each of those people cost. On the best-value platform, it cost £1.51. On the most expensive platform, it cost £4.02. That is 2.7 times as much for the same result.

The reasons were hiding in plain sight. The best-value platform had quietly stopped serving just as primetime began, with only 5% of its ads running in the evening. Another was showing the ad nearly three times each to a small group of viewers, rather than reaching new people. Neither problem set off an alarm. Delivery was on track and costs were in line. By efficiency standards, the campaign was fine.

The fix was almost embarrassingly simple. No new budget, creative or suppliers. The best-value platform kept serving into the evening, lifting its primetime share from 5% to 19%, and we eased frequency back so each ad found someone new. A week later, the campaign moved 17% more people up the funnel for every pound spent, and all six platforms became cheaper per person moved.

Why we keep optimising the wrong thing

If the answer is that simple, why isn’t everyone doing it? Mostly, timing.

Outcome data has traditionally been slow, costly and rationed, with brand lift studies reserved for the big campaigns and reported weeks after they finish. By then, the money has gone. So day-to-day optimisation falls back on whatever signals are available while the campaign is live.

Across the more than 60,000 campaigns we have measured at Brand Metrics, the same lesson keeps surfacing: the cheapest impression and the most effective one are often not the same thing. But you only find that out if you look.

This matters more as buying becomes automated. AI-driven systems will optimise relentlessly towards whatever signal we feed them. Feed them cost and delivery, and they will deliver cheap reach with impressive precision. The machine will do exactly what we asked it to. The question is whether we asked for the right thing.

Three questions worth asking

Efficiency is not the enemy. But it is a means, not an end. Spending less money to achieve nothing is not a saving.

So, before the next optimisation report lands, ask three questions of any campaign:

  1. What is it actually optimising towards: a cheaper cost, or a change in what people think and feel about the brand?
  2. When will we find out whether it worked: while we can still act on the insight, or months later in a wrap report?
  3. If we could see brand outcomes on every campaign while it was live, which decisions would we make differently?

For most of us, the honest answer to that last one is “quite a lot”.

Our industry has spent two decades perfecting the art of efficiency. The next step is to point that skill at what actually matters.



Want to get involved in the discussion?

Your email address will not be published. Required fields are marked *