Buying AI Won't Fix Marketing You Can't Measure

The Take: Buying an AI marketing tool does not fix marketing you cannot measure. It just lets you guess faster and bill it monthly. Fix the measurement first, then let the tools do their job.

Here is the position, and you are welcome to disagree with it. Spending more on AI tools will not fix your marketing. If you cannot measure what you already do, a smarter tool just helps you guess faster.

I have run marketing budgets on both sides of the table, as the client signing off the spend and as the agency spending it. The pattern does not change. The businesses that struggle are rarely short of tools. They are short of the thinking and the measurement needed to know which tools are worth the money. Layering AI on top of that does not close the gap. It widens it.

An AI tool automates whatever decision a business was already making. Across 731 Australian businesses we scored, measurement was the weakest of six marketing dimensions, so most tools scale a guess rather than fix it.

If a business cannot cleanly measure what a channel returns, an AI tool scales that same blind decision instead of fixing the underlying gap.

Ruben Schreurs, the global chief executive of Ebiquity, told the industry this year to stop wasting money on unnecessary new platforms and AI tools. His firm advises most of the hundred largest advertisers in the world, so he is not shouting from the cheap seats. He also put a number on the prize. The global marketing market is worth about US$1.15 trillion in 2026, and most of it is unoptimised. That is roughly the size of Switzerland's economy, sitting there, not working as hard as it should.

Read that again. The waste is not because the tools are bad. The waste is because no one is holding the spend to account. Schreurs made a point I make to clients every week. Measurement decisions across the board are not sophisticated. Most businesses cannot demand proof against an investment because they have no clean way to measure the return in the first place.

Marketing was a doing thing for most of its history. Make the ad. Book the media. Hope it works. AI did not invent that habit. It bolted a faster engine onto it. If anything it made the habit more expensive, because now a business can do the wrong thing at a scale no human team could ever manage by hand.

We scored 731 Australian businesses across six marketing dimensions and Data & Tracking came out lowest of all six, averaging 58.5 out of 100 against a next-best of 62.5. Just over one in five of those businesses scored below 50 on that dimension alone.

That is not a fringe problem. It is the largest and most consistent gap in the whole dataset. It sits directly underneath every AI purchase decision a business makes.

Methodology: figures drawn from New Rebellion's benchmark dataset of 731 scored Australian businesses across six marketing dimensions, scored as at July 2026. This is a benchmark sample, not a random sample of the Australian market, so treat it as directional rather than a national census. Full scoring method at How We Score.

Most businesses cannot cleanly tell you what a customer costs to acquire. They cannot tell you what each channel actually returns. They have analytics installed. It is half configured. Conversions are not firing properly and the mobile experience is leaking revenue that no one is watching. The tools are sitting there. The thinking to use them is not.

So when a business in that state buys an AI campaign tool, what actually happens? It automates the spend. It does not automate the judgement. The business ends up scaling decisions it could never measure in the first place. That is not progress. That is the same guess, run at machine speed and billed monthly.

Yes. Marketing Week's 2026 Career & Salary Survey of 2,350 marketers found 66.5% had identified an AI skills gap inside their own team over the past year, and marketing effectiveness ranked as the number one skills gap for the second year running.

Gartner found that 65% of chief marketing officers expect AI to dramatically change their role, but only 32% think their own skill set needs to change to match. They can see the wave coming. Most are not learning to swim.

This is the catch-22. A business will not invest in measurement until it can see the return, but it cannot see the return until it invests in measurement. So it waits, and while it waits, it spends the AI budget anyway, because the platforms make spending easy and measuring hard. Google's Performance Max and Meta's Advantage+ will happily take the money and report that everything is working beautifully. They are grading their own results. Of course the report looks good. The report is the product they are selling, and it will always flatter the channel that paid for it.

The Australian ad market is forecast to grow about 6.5% in 2026 to roughly $30.7 billion, and retail media is the fastest-growing slice of it, tipped to expand more than 24%. That growth is flowing straight into the most automated channels available, the ones that optimise themselves and report their own results. More money, moving faster, through systems that are very good at looking effective.

If measurement is weak, that is not a tailwind. It is a faster way to lose track of where the money went. The businesses that get hurt will not be the ones who ignored AI. They will be the ones who bought it to avoid the harder work, and never built the discipline to check whether it paid off.

If a business feels the pull to buy the next AI tool, stop for a week and do this first.

Start with two questions. Does the business know how much money it makes? Does it know how much money it spends? If the answer to either is unclear, no tool fixes that. Sort the basics before automating anything on top of them.

Get tracking right before scaling. Clean analytics, conversions firing, a mobile experience that does not leak. This is unglamorous and it is the highest-return work of the year. Fix the site and revenue can jump 20% overnight, before a single new dollar of media goes out the door.

Audit the tools already owned. Most businesses pay for capability they never switched on. Before buying the next thing, find out what is sitting idle in the thing bought last year.

Make the platforms prove it. Run the holdout. Turn a channel off for a month and watch what actually moves. If nothing changes, that is the answer. It is far cheaper to learn it now than after another year of spend.

Pick where to focus. A business cannot be good at every channel and every tool at once. Learn one channel properly, then add the next. Buying ten AI tools at once decorates a business that has not fixed its fundamentals yet.

Does AI marketing software fix poor tracking or analytics?

No. AI tools automate execution and reporting, not the underlying measurement setup. A business with broken conversion tracking or half-configured analytics gets the same bad data, delivered faster and at greater scale.

What is the weakest area of marketing maturity for Australian businesses?

Across the 731 businesses we have scored, Data & Tracking is consistently the lowest-performing of six dimensions, averaging 58.5 out of 100. More than one in five businesses scored below 50 on this dimension.

Should a small business invest in AI tools or fix tracking first?

Fix tracking first. Clean, accurate measurement lets a business tell whether any tool, AI or otherwise, is actually paying off. Without it, every purchase is a guess dressed up as a decision.

How can a business tell if its marketing platforms are reporting honestly?

Run a holdout test. Turn a channel off for a period and measure what actually changes in revenue or leads. Platforms that report on their own performance have an incentive to look effective, so an independent check is the only reliable read.

The AI tools are real and some of them are genuinely very good. This is not an anti-tool position. It is a position against buying automation to skip the thinking. The danger this year is not that businesses miss out on AI. It is that they buy it to avoid the hard part, which is the measuring. They end up spending more to understand less. The market that wins in 2026 is not the one with the most tools. It is the one that knows its numbers and can prove what works.

Want to see where your own measurement actually stands? Get your Hub scorecard and find out before you spend another dollar on AI.