The Platforms Are Writing Your Marketing Reports. That Is the Problem.

IAB Australia just published the strongest Q1 digital ad spend number on record. $4.9 billion in the March quarter, up 15.3 percent year on year. Search and directories at $2.16 billion. Social video up 29.4 percent. Every headline framed it as the market going from strength to strength.

Here is what I keep thinking about: the money going into those channels is being measured by the channels themselves.

That should bother you more than it does.

I have been watching this pattern across Australian businesses for a few years now. The ones that are most confident about their marketing performance are often the ones with the weakest actual position when you look underneath the numbers. They have dashboards full of green metrics. They have attribution models that credit every channel generously. They have agencies presenting slides that show return on ad spend climbing quarter on quarter.

What they do not have is a clear answer to a simple question: if we turned this off, what would actually change?

The cult of performance, as AdExchanger named it this week, has been building for years. The short version is this: AI-driven ad tools have been trained to optimise for measurable performance, and they are extraordinarily good at it. The problem is that measurable performance and actual business impact are not the same thing, and the gap between them is getting wider.

Here is a concrete example. Google's AI ad products, when given control over search keyword decisions, will immediately start bidding aggressively on your own brand name. Your business comes up first in organic search for free. The AI still wants you to pay for the spot above it. The logic is internally consistent: someone searching for your brand name by intent converts well, so the AI targets them heavily. The attribution model credits Google Ads with the conversion. The ROAS number looks great. Nothing incremental happened.

Meta did something quieter in March. It expanded the surface area of what counts as ad engagement, so more accidental swipes and unintentional clicks now register as attribution touchpoints. More actions get credited to ads. Performance metrics improve. Spend follows.

The platforms are not doing this maliciously. They are doing it because they are optimising for the same thing they told you to optimise for: performance. The measurement and the product are built by the same organisation with the same commercial interest. That is the structural problem.

We have scored a lot of Australian businesses across six dimensions of marketing capability. The pattern that comes up most consistently is this: businesses that score highest on acquisition performance — the dimension that captures paid media, organic reach and channel diversification — do not necessarily score highest overall. And the ones that score highest overall are rarely the ones with the most aggressive performance marketing posture.

The businesses performing best when you look at the complete picture tend to maintain investment across multiple dimensions simultaneously. Strong acquisition scores, yes. But also strong brand scores, strong retention scores and strong data scores. The businesses running pure performance playbooks often show a specific signature in the data: acquisition up, brand flat or declining, retention weak, data infrastructure thin. The machine is working. The business underneath it is fragile.

This matters because performance marketing concentrates spend on people who are already close to converting. It captures intent that often exists because of brand investment made months or years earlier. If you drain the brand to fund the performance, you are borrowing from a balance sheet you cannot see.

One category is bucking the overall ad market softness in Australia right now: automotive. Specifically, the IAB data shows automotive recorded the largest year-on-year increase in share of general display advertising, up 1.7 percentage points. The notable contributor is Chinese EV brands entering the Australian market.

Think about why. A Chinese EV brand entering Australia has no existing brand recognition, no dealer network history, no word of mouth. It cannot run a performance campaign targeting people who are already considering its product, because those people do not exist yet. It has to build the audience before it can convert the audience. It is investing in display and brand channels because it understands that demand capture requires demand creation first.

This is not a lesson that applies only to new market entrants. It applies to any business operating in a category where their competitors have stopped investing in brand. The floor is lower. The opportunity to build share of mind cheaply is real. Most businesses are too focused on the performance dashboard to notice.

I want to be clear about something: the platforms are not lying to you. Google Ads does drive conversions. Meta does reach your customers. The attribution models they offer are not invented from nothing. The problem is that they are measuring what is measurable within their own ecosystem, which is not the same as measuring what is working for your business.

The most dangerous version of this is when the AI takes over creative decisions. AdExchanger covered a Chewy ad this week that featured a coin covered in maggots — pulled from the product catalogue by an AI that identified it as a strong thumb-stopper. The AI was right. People stopped scrolling. They stopped because the image was disturbing, not because they wanted to buy anything. The platform counted it as performance. Chewy's brand absorbed the cost.

When you hand creative decisions to an algorithm optimising for engagement, you get engagement. Whether it builds anything is a separate question the algorithm is not set up to ask.

The fix is not to stop using performance channels. They work. The fix is to change how you measure them.

First, test incrementality instead of attribution. Pick a geography or a customer cohort and turn a channel off for four weeks. Compare outcomes. That is the actual signal. Attribution models tell you who was in the room. Incrementality tells you whether the room mattered.

Second, audit what your AI ad tools are actually doing. Log into Google Ads and look at where your Performance Max spend is going. Check how much is landing on brand keywords and competitor searches. Set exclusions and watch what happens to the ROAS number. If it falls sharply, that is the gap between claimed performance and actual increment.

Third, maintain brand investment even when the performance numbers look good — especially when they look good. The time to invest in brand is when you can afford to, not when you have already depleted the asset and the performance metrics start declining.

Fourth, read the platform changes as they happen. When Meta quietly changes what counts as engagement, that changes what your attribution model is measuring. The number in your dashboard is not a fixed unit. It is whatever the platform decided to count this quarter.

Australian digital ad spend will probably hit another record in Q2. The platforms will report strong performance. Agency dashboards will show improving metrics. Most of it will be real.

Some of it will be the platforms getting better at measuring themselves.

Your job is to know the difference. The only way to do that is to build measurement infrastructure that sits outside the platforms you are measuring. It is harder than reading a dashboard. It is also the only way to actually know what is working.

Numbers on the board matter. Make sure they are your numbers.