This article is part of a wider body of work I have pulled together in a longer research report on why marketing performance has become harder to prove, harder to defend, and in many organisations, harder to make matter. The report connects a set of “signals” that many teams are feeling in practice: weaker measurement signal, rising internal scrutiny, and the slow drift from marketing as a growth engine to marketing as a reporting function. This piece zooms in on one of the most corrosive patterns that emerges under those conditions: attribution theatre.
I will keep this practical, human, and a little uncomfortable, because it is not an academic problem. It is a day to day organisational behaviour problem. And in my view it is one of the root causes behind the decay of the marketing function in the 2020s.
On Monday morning, a finance email lands like it always does.
“Quick one. Which channel drove last week’s revenue lift?”
You stare at the question for a beat too long, not because you do not have an answer, but because you have twelve. And none of them are clean.
So you do what modern marketing trains you to do.
You open the dashboard. You select the attribution model. You build a story with numbers that look precise enough to soothe the room. And as you paste the chart into a deck, you feel it again, that quiet split between what you know and what you are about to say.
That split is where attribution theatre begins.
It is not a fraud. It is not a conspiracy. It is a coping mechanism that forms when signal weakens but the organisation’s demand for certainty stays high, or gets higher.
And once it takes root, it does not just mismeasure marketing. It reshapes the behaviour of the entire company.
Why attribution felt like progress
In the performance era, attribution felt like a revolution. For the first time, marketing could point to something that looked like causality.
Clicks. Sessions. Conversion rate. ROAS. A path that appeared to move from ad to site to sale. A control panel. You could pull levers and watch the needle move.
This changed marketing’s status inside organisations. It gave the function an answer to the oldest question in business:
“Is this working?”
And it came with a promise that was deeply seductive to leaders under pressure.
Not just “we think marketing is helping.”
But “we know which part is helping, and by how much.”
The problem is that the promise got culturally encoded. It became the standard for legitimacy.
Marketing did not just become measurable. It became expected to be measurable at a microscopic level, all the time, with clean answers on demand.
That expectation still governs a lot of boardrooms today, even though the environment has moved in the opposite direction.
Why it breaks now
The modern measurement environment is increasingly defined by partial visibility.
Privacy regulation, consent barriers, ID fragmentation, walled gardens, and black box algorithms have narrowed what any one team can actually observe.
Even when each platform can tell a compelling story about performance inside its own walls, the moment you try to reconcile performance across ecosystems, confidence drops. The single customer journey becomes a set of overlapping partial views stitched together with assumptions.
This is the core tension of the 2020s marketing function.
Observability is declining, so certainty becomes theatre.
And it is not happening in a calm environment. Budgets are flat or under pressure, scrutiny is higher, and expectations remain aggressive. That combination turns measurement from a learning tool into a survival tool.
So organisations do what organisations always do under uncertainty.
They ask for more proof.
They tighten reporting cadences.
They expand dashboards.
They demand cleaner attribution.
Which is understandable.
And it is exactly how the trap closes.
What attribution theatre looks like inside companies
Attribution theatre has a very specific vibe. You can feel it in the weekly meeting.
The room is full of people who did not used to attend marketing discussions. Finance is there. RevOps is there. Someone from Data is there. The CEO is there “to stay close to performance.”
Marketing gets ten minutes.
Not ten real minutes. Ten minutes as a test.
You put up the slide everyone expects. A clean line. A tidy breakdown. A pie chart that suggests the world is knowable.
And then you start speaking, and the caveats arrive immediately.
“Organic is up, but branded search is doing some heavy lifting.”
“Paid improved, but we tightened targeting after CPCs spiked.”
“The nurture sequence is contributing, but it’s multi touch.”
You are being accurate. But accuracy is expensive in a room that wants certainty.
Then someone else speaks, usually sales, sometimes product, occasionally the CEO.
They do not have a chart. They have a narrative.
“Pipeline rose because we pushed the accounts we were sitting on.”
It is not false. It is simply cheaper. And cheap stories win when the organisation is tired.
This is how attribution theatre becomes more than a marketing issue. It becomes an organisational behaviour pattern.
Because attribution is not just measurement anymore.
It is status.
Who gets to tell the story of why the company is winning or losing.
Who looks like the engine, and who looks like the cost.
Who gets budget protection, and who gets cuts.
So marketing responds the way any function would respond.
It builds defence documents.
More dashboards. More models. More confidence in language. More claims that “we know this drove X,” even when everyone in the team quietly knows the model is built on partial views and assumptions.
Attribution becomes less about learning and more about appearing in control.
That is the theatre.
The organisational costs
Attribution theatre is not harmless. It does not simply create a little noise in reporting. It changes what the company invests in, how teams behave, and which kind of marketing survives.
1) It turns marketing into an internal politics game
When budgets are allocated based on what the model can see, teams begin to fight for credit, not outcomes.
You see it as subtle territorial behaviour.
Channel owners protecting their numbers.
Arguments over attribution windows.
Debates that sound technical but are really about control.
And because visibility is partial, these arguments are rarely resolvable. They become proxy wars.
The result is a slower organisation with more friction and less trust.
It becomes harder to collaborate across functions because every discussion has an implied subtext:
“If we do this together, who gets credited?”
The tragedy is that growth is increasingly cross functional by default, so the behaviour is the opposite of what the environment requires.
2) It bakes in a short term bias
Attribution theatre changes what “good” looks like.
It rewards the activity that produces immediate, attributable movement. It punishes investment that works upstream, works slowly, or works through messy interactions the model cannot capture cleanly.
Over time, the organisation learns a very specific habit:
If it cannot be proven this quarter, it becomes politically fragile.
That is how demand creation gets underfed.
Brand work is reduced to a soft nice to have.
Customer research gets cut because it does not “show up” in the dashboard.
Distribution leverage gets ignored because it pays off later.
And then, six to twelve months later, performance becomes more expensive and less reliable, and everyone asks why the engine feels weaker.
The answer is not mysterious.
You optimised the visible part and starved the invisible part.
3) It drives underinvestment in real demand creation
The most damaging effect is that attribution theatre teaches the company to confuse capture with creation.
Lower funnel activity looks clean in a model. It is closer to the conversion event. It is easier to count. It produces graphs that move.
Upstream demand creation is messy. It works through memory, recognition, preference, trust, availability, and long term compounding. It rarely offers clean causal proof on demand.
So the company drifts toward what it can defend.
And marketing slowly becomes a function that captures demand created elsewhere, rather than one that confidently builds future demand.
This is one of the central forces behind the “decay” feeling many teams report in the 2020s.
More effort. More output. More reporting.
Less clarity. Less confidence. Less real leverage.
4) It expands complexity and burns teams out
Attribution theatre makes the dashboard grow.
A new question appears, so a new metric is added.
A new metric is added, so a new meeting appears.
A new meeting appears, so a new deck appears.
The team spends more time building readiness than building advantage.
And then AI arrives and accelerates the expectation trap.
If AI makes marketing faster, leadership assumes certainty should follow. If you can generate more variants, surely you can prove which one worked. But certainty does not scale like output. Often, it scales slower, because evaluation and interpretation become the bottleneck.
So teams ship more, defend more, and learn less.
That is not a talent problem.
It is a system problem.
The alternative: decision quality measurement
The way out is not to abandon measurement. It is to change what measurement is for.
The shift is from attribution certainty to decision quality measurement.
Decision quality measurement has one job:
Help the organisation make better investment decisions under uncertainty, without pretending the fog is a straight line.
That sounds abstract until you see what it changes in practice.
1) It changes the leadership question
Instead of “What caused this result with certainty?”
The question becomes:
“What should we do more of, less of, or differently, and why is that the most defensible bet?”
This is a profound shift because it makes honesty survivable. It reduces the incentive to manufacture precision.
2) It introduces a hierarchy of evidence
Causal where feasible.
Triangulation everywhere else.
In practice:
- Use experiments and incrementality testing when you can.
- Use modelling and econometrics when user level visibility is weak.
- Use brand and demand proxies as directional signals, not courtroom proof.
- Use qualitative insight to interpret why numbers moved.
- Accept that no single dashboard is “truth,” and build governance around combining signals.
This is not less rigorous. It is more mature.
It matches the reality of complex systems.
3) It reduces theatre by changing incentives
Attribution theatre thrives when marketing is forced to prove itself with a single number.
Decision quality measurement reduces the single number obsession.
It also shrinks the dashboard, because the goal becomes “few credible proxies that change decisions,” not “everything we can possibly count.”
And it protects long term work from short term panic by making time horizons explicit.
Now: convert demand.
Next: build demand.
Later: build leverage.
Once the organisation acknowledges these horizons, it becomes much harder to accidentally starve the future while celebrating short term efficiency.
4) It restores marketing’s role as a growth system, not a reporting system
The quiet goal here is credibility, but not the fake kind.
Not credibility built on confident numbers that are partially invented.
Credibility built on a consistent, repeatable commercial narrative supported by honest measurement.
That is how marketing stops being a function that produces defence documents, and becomes one that produces direction.
A final scene: the same email, a different response
Let’s go back to the finance email.
“Which channel drove last week’s revenue lift?”
In the old posture, you reply with a pie chart and a false calm.
In the decision quality posture, you reply with a narrative that does not pretend.
“Demand was higher than expected and we captured more of it because we reduced checkout friction and clarified the offer. Paid helped, but we think the category moved too. We are treating that as an external factor until we can validate it. Next week we are doubling down on the conversion fix and testing whether the message change holds in a controlled way.”
This does two things.
It protects truth.
And it teaches the organisation how to think.
Because the real cost of attribution theatre is not that it gets the numbers wrong.
It is that it trains the company to make worse decisions, more confidently, faster.
In a declining signal world, that is how you quietly destroy a marketing function.
Not with incompetence.
With certainty.
