If we can measure it, it worked

Weak marketing decisions look much more respectable once they put on a lab coat.

I’ll know good creative already showed what happens when taste, seniority, and internal comfort carry the review. The next move feels responsible: judge the work by the numbers instead.

If we can measure it, it worked.

The idea is seductive because it allows the business to turn ambiguity into dashboards. Dashboards can also protect people from blame, but the trap here is different: measurability starts deciding what marketing is allowed to matter.

Impressions, clicks, engagement, conversion rates, cost per lead, open rates, attributed pipeline, return on ad spend, opportunity creation, deal influence. A language that sounds harder than opinion. A way to settle arguments without having to rely on instinct, politics, or taste.

Sometimes that is exactly what measurement does.

Good measurement can stop a company lying to itself. It can expose friction, wasted spend, channel mismatch, bad follow-up, weak pages, poor data, false confidence, and activity with no visible commercial effect. Businesses should absolutely measure marketing. A book telling founders not to would be a dangerous book.

The danger is that companies rarely stop at measuring. They start over-believing whatever is easiest to measure neatly.

The B2B Institute has reported that 96% of B2B marketers in a LinkedIn study expected the main effect of ad campaigns within two weeks. That expectation naturally over-rewards fast signals and under-rewards compounding ones.

Once that happens, the dashboard stops being a tool for learning. It starts shaping what the organisation treats as real. Some effects count more because they show up faster. Some channels gain prestige because they produce cleaner lines to response. Some activities become politically safer because they generate numbers leadership can point to in a board deck.

What began as accountability turns, slowly, into a culture that mistakes visible traceability for total contribution.

That is how marketing gets narrowed without anybody needing to say they are narrowing it.

The logic usually sounds impeccable.

If a channel produces leads, it must be working. If a campaign gets engagement, people must care. If the conversion rate is improving, the marketing must be better. If branded search rises after a campaign, the campaign must have driven it.

If nothing measurable happened, the activity probably was not worth doing.

Every one of those statements contains a little truth. None of them is safe enough to build a marketing culture on.

The numbers are not fake. They are partial.

They tell you what happened in the part of the system that was instrumented, on the timescale the company knows how to read, using the metrics the organisation has learned to respect. That makes them useful. It also makes them dangerous when they are mistaken for the whole story.

A company can become obsessed with what it can see and oddly casual about what it cannot.

This is one reason the phrase data-driven deserves more suspicion than it usually gets. In practice, many businesses are driven by the subset of signals that arrive quickly, fit inside attribution tools, and flatter the company’s existing preferences for immediacy, precision, and control. That is dashboard-driven judgement wearing a lab coat.

The distinction matters because markets do not organise themselves around the convenience of internal reporting. Buyers notice things before they click. They hear of companies before they search. They absorb cues without remembering the moment they absorbed them.

They compare options using memory, familiarity, hearsay, prior exposure, and accumulated impressions that no single system can fully track. Some of this is the slow work of building mental availability around category entry points.

Then the firm gives the eventual form fill to the last visible touchpoint and concludes it has found the engine. That is a common risk in attribution systems: last-click models, for example, give all credit for a conversion to the last-clicked ad and corresponding keyword.

This is ordinary marketing self-deception.

The channel nearest the action gets the cleanest credit. The work done earlier and more broadly gets treated as background. The business then reallocates spend towards the thing that looks measurable and starts starving the thing that helped make that measurable thing work.

That pattern has been described in different ways for years. Performance drives revenue already showed how clean short-term attribution can starve longer work in budget meetings. The measurement version is the same bias on a finer clock: the visible effect gets mistaken for the whole effect.

A simple company scene makes the problem clearer.

A campaign has run for eight weeks. The dashboard shows solid click-through rates on retargeting, a decent cost per lead from search, modest engagement on organic social, and weaker-than-hoped early results from a broader awareness push.

The founder looks at the chart and reaches a conclusion that feels hard-headed: the work closest to the conversion point is working, the broader work is underperforming, so the obvious move is to reallocate more budget into the former.

The conclusion makes sense up to a point.

What if the retargeting worked partly because the broader campaign had made more of the audience recognise the company? What if branded search improved because repeated exposure had made the name easier to recall?

What if the awareness activity is affecting future shortlist entry, not current leads? What if the social engagement is telling you more about platform behaviour than buyer value? What if the weak early numbers are weak mainly because the company is asking a longer-term effect to justify itself on a short-term clock?

Those questions are irritating because they reopen uncertainty. Dashboards feel so comforting precisely because they seem to close it.

That is one reason measurement becomes political so quickly inside companies.

Sales wants metrics tied to pipeline. Finance wants metrics tied to spend efficiency. Founders want metrics that look serious to boards. Marketing wants metrics it can actually influence. Agencies want metrics that prove their channel matters. Product wants metrics tied to usage or activation.

Everybody wants numbers that make their preferred activity look commercially adult.

Soon enough, the debate is no longer about what best reflects how growth happens. It is about which metrics the organisation will treat as legitimate.

That has real consequences.

Trust demand-capture metrics alone and the company overfeeds capture. Trust lead volume or attributed pipeline alone and future-shaping work gets undervalued. Trust only what shows inside one quarter and marketing trains itself against compounding.

This is how the metric hierarchy collapses into a single scoreboard. A company can point to numbers and still be avoiding the real decision.

Imagine two firms in the same category.

Company Visible judges almost everything by what can be tied neatly to a short reporting window. It loves the channels that produce fast response. It becomes excellent at reporting on leads, cost per opportunity, attributed revenue, and campaign-level ROI. Broader work is tolerated only if it quickly generates an obvious trace.

Company Hierarchy also measures hard outcomes, but it accepts that different metrics answer different questions. It has a hierarchy rather than a single scoreboard.

At the top are commercial outcomes like revenue, customer growth, and pipeline quality. Below that are the measures that help explain those outcomes - reach, share of search, branded demand, conversion efficiency, buying-stage movement, and evidence of future market presence.

It gives each metric a job, and keeps the bottom of the funnel in proportion.

From the outside, Company Visible often looks more rigorous because its reporting is cleaner. From the inside, Company Hierarchy is usually making better decisions because it understands what each number can and cannot explain.

A click, an engagement, a form fill, a lead, and a sale are all real events. They are not equally diagnostic.

A click may tell you the asset generated curiosity. An engagement may tell you the platform found the post socially compatible. A lead may tell you the channel reached an active buyer. A sale may tell you the company eventually won.

None of these, on their own, tells you the whole causal story.

That is why the idea of a metric hierarchy matters. It is also why better teams separate reporting from evidence.

Dashboards are for operational visibility. If the business wants more confidence about cause and effect, it usually needs a stronger test: hold back a comparable audience, compare similar markets, switch activity on and off in a controlled way, or use another incrementality testing design that gives the company a cleaner comparison.

The method can vary. The principle is simple enough.

Do not ask a dashboard to prove what only a proper comparison can show.

A healthy company does not ask every metric to answer every question. It asks:

• What are the actual business outcomes we care about?
• What leading indicators help us understand whether we are making those outcomes more likely?
• Which metrics help diagnose channel or execution problems?
• Which numbers are interesting but easy to overinterpret?
• Which effects do we expect to show up slowly, and how will we avoid declaring them failures too early?

That is a much better measurement culture than “what can we prove right now?” It is also much less flattering to shallow certainty.

Take social engagement: a post may get likes because it is familiar, agreeable, or easy to endorse publicly. Useful sometimes, but companies often promote the metric above its station, as though reaction proved market effect when it often proves platform fit.

The same goes for conversion rates. If conversion rises after narrowing targeting, the uplift may be genuine. The company may simply have moved into a warmer, smaller, easier pool rather than strengthened its market position. Better targeting made that point from the angle of prioritisation.

The measurement lesson is that improvement in a metric is not always improvement in a business.

This is what makes dashboards so dangerous when used badly. They encourage local optimisation.

Each channel gets better at its own metric. Each team gets better at producing its own proof. Each agency gets better at telling its own success story. The business gets better at counting.

Nobody is fully responsible for whether the system as a whole is getting stronger.

That is how a firm can improve its numbers while weakening its growth logic.

A lot of founders recognise this too late. They begin by asking for accountability and end by receiving immaculate reports from a marketing machine that has learned exactly what kind of evidence the business rewards.

The reports can be honest and still come from a narrowed definition of value.

A practical framework helps because it keeps each metric in its lane.

A sensible measurement model for B2B should usually separate four layers.

First, business outcomes. Revenue, customer growth, retention, pricing realised, pipeline value, sales velocity.

Second, market effects. Reach, branded search, direct traffic, share of voice, awareness proxies, consideration signals, category association, distinctive asset recognition where measurable.

Third, channel and conversion diagnostics. Click-through rates, conversion rates, cost per lead, landing-page performance, email response, retargeting efficiency, search performance.

Fourth, activity and attention indicators. Engagement, opens, views, likes, time on page, downloads, event registrations.

All four matter, in different ways and at different levels of decision-making.

The danger begins when layer three or four starts impersonating layer one. A company ends up optimising what is easy to move instead of what most needs to improve.

This is one reason the phrase “what gets measured gets managed” is less wise than it sounds. In practice, what gets measured most cleanly often gets over-managed, while the more diffuse but strategically important parts of growth get starved of attention because they resist neat weekly interpretation.

A better rule is this:

That is less catchy. It is much safer.

This is also where founder behaviour matters again. Founders do not just ask for numbers. They teach the organisation what kind of numbers will be rewarded, feared, or trusted.

If the founder only leans forward when leads, attribution, or ROI are mentioned, the company learns very quickly what counts. If the founder treats broad market measures as fluffy, those measures will never shape decisions properly.

If the founder asks only for certainty, the dashboard will eventually become a theatre of certainty whether or not the market deserves it.

Marketing teams rarely choose their measurement culture alone. They inherit it.

Measurement is never neutral. It shapes what marketing is allowed to care about. It shapes what gets funded. It shapes how long activities are allowed to run. It shapes which failures are tolerated and which are punished. It shapes what the company mistakes for truth.

And once a business starts relying heavily on numbers to settle decisions, one particular kind of distortion becomes almost irresistible.

The company starts favouring the kinds of work that the numbers reward fastest - even when those numbers belong more to the platform than to the market. On social, that distortion often lands first: This post did well.