You Know What's Working in Your Marketing. You Have No Idea Why.

Your best Google Ads campaign is printing leads at $28 each. Your worst is limping along at $140. The obvious move? Shift budget from the loser to the winner.

That is precisely how most businesses manage their marketing. And it is precisely why most businesses plateau.

The problem is not that the $28 campaign is working. The problem is that you cannot explain why it is working. You have a correlation. You do not have a cause. And until you know the cause, every optimisation you make is an educated guess wearing a lab coat.

The Most Expensive Question Nobody Asks

Mark Ritson has spent years arguing that marketing doesn't start with creativity but with diagnosis. His framework is blunt: diagnosis first, strategy second, tactics last. Never in another order. He describes brand diagnosis as the "number one job" any marketing manager will have throughout their tenure.

Most small businesses skip it entirely. They jump straight to tactics. New ad copy. Different keywords. Higher budget. Lower budget. Broader targeting. Narrower targeting. Each change is a tactical guess without a diagnostic foundation.

The data backs him up. According to Funnel.io's 2026 State of Marketing Intelligence report, nearly eight in ten marketers say they do not have a clear signal on what is truly working. Not what appears to be working in a dashboard. What is actually driving business outcomes.

That is a staggering number. Eighty percent of marketers are making daily decisions about where to spend money without understanding what those decisions produce. They have reports on what happened, but they lack the diagnostic intelligence to understand why it happened or what to do next.

Correlation Feels Like Insight. It Is Not.

Here is the trap. Google Ads tells you Campaign A converts at 12% and Campaign B converts at 3%. You feel like you know something. You have data. You have a spreadsheet. You have a chart trending in the right direction.

But attribution models, by design, tell you correlation, not causation. They identify patterns. They cannot tell you why a conversion happened or what the true incremental impact of each ad interaction was. The scale of the problem is significant: accurate causation analysis can reveal a difference between 95% accuracy versus 30 to 60% for traditional attribution models.

This is the diagnostic gap. And it shows up everywhere.

When your tracking is wrong, your Smart Bidding is optimising toward the wrong signals. Your reports are lying to you. And every decision you make is built on contaminated data. We have written before about why your marketing dashboard is lying to you and the compounding damage that produces.

But here is the thing most people miss: even when your tracking is technically correct, you still have a diagnosis problem. Knowing what converted does not tell you why it converted.

The $300 Million Difference Between "What" and "Why"

Sam Tomlinson, one of the sharpest media buyers working today, shared a story in his newsletter that crystallises this perfectly.

His team was managing an account where a single creative, a founder video, was outperforming everything else by roughly 3x. The logical, scientific response? Produce more founder videos. More of what works.

They did not do that. Instead, they spent a week analysing customer interviews, screen recordings, and performance data trying to answer a different question: Why did this specific video resonate so deeply with this specific audience?

The answer had nothing to do with "founder videos work." It was about a specific fear the CEO articulated on camera. A fear that every other active creative in the account never once addressed. While every other ad talked about features and benefits, this one spoke to the anxiety of making the wrong choice. One customer described it as: "I watch this and think, she's saying what I feel in my gut, but I'm afraid to say."

That single diagnostic insight did not just produce better video ads. It restructured the entire messaging strategy for the account, changed the landing page approach, and ultimately led to a 40%+ reduction in CPA across the board.

The creative underperformance was the symptom. The insight was the asset. And extracting that insight required asking "why," not "what."

If they had followed the standard playbook, they would have made more founder videos, gotten mediocre results, and concluded that "founder content is losing steam." They would have optimised the tactic while missing the principle. This is the same dynamic we explored in how buyers actually make decisions: the surface-level behaviour tells you almost nothing about the underlying psychology.

Your Customers Cannot Tell You Why Either

Here is where it gets uncomfortable. If you ask your customers why they chose you, the answer they give is almost certainly wrong.

Daniel Kahneman's research on System 1 and System 2 thinking explains why. System 1, the fast, intuitive brain, makes 19 of every 20 decisions before the rational brain even notices. That includes high-stakes choices like which tradesperson to call, which agency to hire, which supplier to trust.

When customers are asked directly why they made a choice, the explanation they give is a System 2 narrative constructed after the fact. It is not an accurate account of what actually drove the behaviour. They will say "you had good reviews" or "your website looked professional" or "I liked the price." These are post-hoc rationalisations, not diagnostic truths.

The real reason is usually something they cannot articulate: a feeling of familiarity. A sense that you understood their problem before they explained it. A reduction in anxiety at a moment when they felt uncertain. System 1 responds to emotional triggers, not feature lists. If your best ad happens to hit one of these triggers, you need to know which one, or you cannot replicate it.

This connects directly to Byron Sharp's work at the Ehrenberg-Bass Institute on Category Entry Points. CEPs are the specific needs, occasions, or situations that cause a buyer to think about your category. "I'm stressed about my ad spend." "My website is not generating leads." "I just lost a big customer." Each of these is a mental door through which someone enters the market.

Sharp's research shows that large brands are linked to a broader range of CEPs than smaller brands. Most small businesses cover one. Maybe two. The rest of the market, the other four or five doors through which customers enter, remains invisible to them. Not because they lack the ability to address those needs. Because they have never diagnosed which doors exist. We covered this in depth in why there are six reasons someone needs you but your marketing covers one.

The Eurostar Problem: Solving the Wrong Thing Brilliantly

Rory Sutherland's famous Eurostar example puts this in perspective. Engineers spent £6 billion building new tracks to shorten the London-to-Paris journey by 40 minutes. They solved the engineering problem brilliantly. Fast trains. Better tracks. Measurable time savings.

The engineers diagnosed a speed problem. Sutherland diagnosed a perception problem. The passengers were not unhappy about duration. They were unhappy about what they could do with that time. Once you understand that, the solution changes completely and costs a fraction as much.

This is what happens in marketing every day. A business sees low conversion rates on their landing page and assumes they need better ad copy. They spend months testing headlines. The real problem? The form asks for eight fields of information from someone who is not yet sure they trust the business. It is a friction problem, not a copy problem. And no amount of headline testing will fix it.

Activity Metrics vs. Outcome Metrics: The Diagnostic Split

Avinash Kaushik, the analytics evangelist behind the Occam's Razor blog, draws a critical distinction between three layers of measurement:

The diagnostic layer is where insight lives. It is the difference between "our conversion rate dropped" (activity observation) and "our conversion rate dropped because 68% of mobile visitors abandon the form at field four, which asks for a business ABN that most sole traders don't have memorised" (diagnostic insight).

One observation leads to panic. The other leads to a fix that takes ten minutes. We have seen this pattern repeatedly: the 10-minute marketing fixes that are worth more than your entire ad budget almost always emerge from diagnostic thinking, not tactical optimisation.

Kaushik's advice: torture your metrics every six months by asking one question. Have the actions taken based on insights from this metric had a direct impact on our KPI? If the answer is no, the metric is decoration.

The Diagnostic Habit: Five Questions That Change Everything

Diagnosis is not a one-off audit. It is a habit. Here are five questions that separate businesses who understand their marketing from businesses who merely run it.

Not what happened. Why. Was it seasonal demand? A competitor going offline? A new landing page that reduced friction? A referral that you cannot replicate? If you cannot explain your best month, you cannot engineer another one.

Not demographics. Behaviour. Did they visit a specific page? Did they come through a specific search term? Did they interact with a specific piece of content? The pattern in your best leads tells you which Category Entry Point is actually working.

Most post-mortems stop at "low CTR" or "high CPC." Those are symptoms. The diagnostic question is whether the campaign failed because the targeting was wrong (wrong people), the message was wrong (right people, wrong trigger), or the offer was wrong (right people, right trigger, insufficient motivation to act). Each diagnosis produces a completely different fix.

This is an incrementality question. If you turn off your best campaign and leads barely change, you are paying for customers who were going to find you anyway. If leads drop significantly, you have genuine incremental impact. Either answer is diagnostic gold. We explored this tension in why you might be paying for customers who were already going to call.

Not what information it communicates. What feeling it produces. Tomlinson's founder video worked because it created relief from anxiety. The best performing ads in our accounts almost always trigger a specific emotional response: confidence that this business understands the problem, relief that a solution exists, or urgency created by a genuine constraint. If you cannot name the emotion, you have not diagnosed why it works.

Optimisation Without Diagnosis Is Expensive Guesswork at Scale

Here is the uncomfortable truth. Every business optimises. Few diagnose. The difference is not effort. Both require time and attention. The difference is that optimisation asks "how do I get more of what is already happening?" while diagnosis asks "why is it happening in the first place?"

Optimisation without diagnosis produces diminishing returns. You squeeze another 3% out of a campaign that is already performing. You test a headline variation that lifts CTR by 0.2%. You adjust bids by a dollar. These are real improvements. They are also tactical ceilings.

Diagnosis produces step changes. A single insight about why your customers choose you can restructure your entire approach. Tomlinson's team did not get a 3% improvement. They got a 40% reduction in CPA across the board. That is the difference between asking "what" and asking "why."

The Ehrenberg-Bass Institute found that only 16% of advertising is both recalled and correctly attributed to the right brand. That means up to 84% of advertising spend could be wasted, not on bad creative, but on creative that fails to build the right mental associations. You cannot fix that with bid adjustments. You fix it by diagnosing what your brand actually means in the buyer's mind and building distinctive assets that stick.

What This Means for Your Business

Stop asking your dashboard what happened. Start asking why it happened.

If your best campaign is generating leads at half the cost of everything else, do not just give it more money. Pull it apart. Talk to the leads it generated. Look at the search terms. Read the landing page with fresh eyes. Find the emotional trigger, the Category Entry Point, the moment of friction reduction that made it work. Then apply that insight everywhere.

If your worst campaign is failing, do not kill it until you understand whether it failed because of targeting, messaging, offer, or landing page experience. The diagnosis determines whether you need a new audience, a new angle, or a new page. Without it, you are just guessing with a different budget.

Ritson puts it simply: diagnosis, strategy, tactics. In that order. Never another way. The businesses that grow are not the ones with the biggest budgets or the cleverest ads. They are the ones that understand why their marketing works, and build everything else on that foundation.

Your marketing is producing results right now. Good or bad, those results contain a diagnosis. The only question is whether you are reading it.

Further Reading

• Kill Data Pukes: Split KPIs, Diagnostic Metrics & Influencing Variables - Avinash Kaushik's framework for separating metrics that matter from metrics that decorate
• The Missing Step in Marketing Plans: Brand Diagnosis - Mark Ritson on why diagnosis must come before strategy
• Category Entry Points: The Last Two Strategic Pitfalls Unraveled - Ehrenberg-Bass Institute research on the buying situations that trigger brand recall
• Why 2026 Is the Year of Marketing Intelligence - Recent data on the gap between marketing data and marketing understanding
• The Art and Science of Media Buying - Sam Tomlinson's newsletter on separating creative judgment from analytical rigour