Goodhart's Law in GTM: When Your North Star Becomes the Reason You Miss

There is a rule from economics that explains more GTM failures than any playbook will admit. It is called Goodhart's Law, and it goes like this: when a measure becomes a target, it stops being a good measure.

The idea started in monetary policy in the 1970s, but it describes a B2B revenue org perfectly. Pick any metric your team is rewarded on. Watch what happens to it over a few quarters. The number keeps going up while the thing it was supposed to represent quietly stops moving. The metric becomes a game, and the game replaces the goal.

Most GTM leaders are not missing their number because they ignored their metrics. They are missing it because they hit them.

How a healthy metric goes bad

Every vanity metric started life as a reasonable proxy. That is what makes this so hard to catch.

Take the MQL. It began as a sensible shorthand for marketing: a lead engaged enough to be worth a sales conversation. Useful. Then someone made it a target, set a quarterly MQL quota, and tied demand gen's budget to hitting it. From that moment the definition of an MQL started drifting downward, because the fastest way to produce more MQLs is to lower the bar for what counts as one. The number climbs every quarter. Sales trusts it less every quarter. The proxy and the reality come apart, exactly as Goodhart predicted.

CROs run the same script with a different number. Pipeline coverage was a useful sanity check until "3x coverage" became the target, at which point reps learned to stuff the pipeline with deals that pad the ratio and never close. The CRO walks into a board meeting with a healthy coverage chart and walks out of the quarter with a missed number, because the chart was being managed instead of the pipeline.

RevOps and GTM Finance see a third variant. Lead scores were a fine way to prioritize SDR outreach until they were asked to forecast pipeline, a job they were never built for, and teams started optimizing the score instead of the revenue. We wrote about that specific failure in Why Lead Scoring Fails and What's Next for GTM Teams. It is Goodhart's Law in one chart: the moment the score becomes the goal, it stops predicting the thing you actually care about.

The pattern is identical across all three roles. A proxy gets promoted to a scoreboard. The number gets gamed. The outcome stops moving. Nobody is doing anything wrong. The system is doing exactly what the incentives told it to do.

Why this keeps happening

It would be easy to blame the people gaming the metrics. That misses the point. The gaming is rational. If you reward someone for a number, they will move the number, and the easiest way to move a number is almost never the same as the hard work the number was meant to stand in for.

The real culprit is that proxies are convenient and outcomes are slow. Revenue takes two or three quarters to show up. MQLs show up today. So teams measure what is fast instead of what is true, and act surprised when the fast number and the true number disagree at the end of the year.

This shows up worst in cross-functional meetings, where the same proxy means three different things to three different people. Marketing has its number. Sales has its number. Finance has a third. None of them match, because each team built its number for the local incentive it was managing. The result is what we have called the Debate Tax: meetings stop being about what to do next and start being about whose data is right. Eventually teams stop bringing data at all, because they know it will be challenged, and the company flies blind by choice. It is more comfortable than fighting.

There is a quieter version of this that punishes the person closest to the metric. A VP of Marketing gets promoted in Q1 on an attribution model that credits her channels, then puts on a performance plan in Q3 when the model changes. Her actual work did not change. The lens did. She was not experiencing performance volatility. She was experiencing model volatility. The proxy was political the whole time, and she got judged by it as if it were physics.

The fix is not a better metric. It is a better direction of travel.

The instinct, once you spot a gamed metric, is to invent a cleaner one. That rarely works, because the new metric becomes a target too, and Goodhart starts over. You cannot out-define this problem.

What breaks the cycle is flipping the direction you measure from. Most GTM measurement starts at the top of the funnel and pushes forward: impressions, then clicks, then MQLs, then hopefully revenue. Every step is a proxy for the step after it, and every proxy is gameable. The alternative is to start at the outcome and work backward.

A demand gen leader at an enterprise search company put it well. Daniel Henderson of Glean, after he stopped trusting top-of-funnel vanity numbers, described the shift like this: "If you're skeptical about branding campaigns (like I was), flip the model. Instead of starting with impressions, RevSure starts with outcomes, meaning leads that progressed to a meeting set, and worked backwards."

That is the whole answer in one sentence. When you measure backward from the outcome you actually want, there is nothing to game, because the outcome is the thing itself, not a stand-in for it. A lead that became a meeting that became pipeline is not a proxy for revenue. It is revenue, traced to its source. Marketing, sales, and finance can finally read from the same scoreboard, because there is only one.

What this means for GTM leaders in 2026

You do not need to rip out your metrics. You need to demote them. Keep MQLs, coverage ratios, and lead scores as operational signals, the dials that help a team prioritize day to day. Just stop treating them as the scoreboard. The scoreboard should be the outcome, measured backward: which touches, campaigns, and accounts actually produced pipeline and revenue, traced from the closed deal back to the first signal.

For a CMO, that means walking into the board meeting with a number sales and finance both already agree on, instead of defending a number only marketing's model produces. For a CRO, it means a forecast that holds up when the deals are actually examined, not just when the coverage ratio is. For RevOps and GTM Finance, it means one source of truth that ends the Debate Tax for good.

The teams that win in 2026 are not the ones with the cleanest top-of-funnel numbers. They are the ones who stopped confusing the proxy with the point. Your north star should be the outcome itself. The moment it becomes anything else, it becomes the reason you miss.

Frequently asked questions

How do I tell if my GTM team is gaming a metric? Look for the gap between the number and the outcome it is supposed to predict. If MQLs are climbing quarter over quarter while MQL-to-pipeline conversion is sliding, the bar has been lowered, not the work raised. If pipeline coverage looks healthy but win rates are dropping, deals are being added to the pipeline that do not belong there. The clearest tell is when two teams that should agree, marketing and sales, marketing and finance, never do, even when the data sources are technically the same.

Should MQL targets and pipeline coverage ratios come out of comp plans entirely? No, but they should stop being treated as outcomes. Use them as operational dials that help SDRs prioritize and AEs sequence their work. Tie comp to the actual revenue and pipeline these dials are supposed to produce. The moment a proxy becomes the comp lever, Goodhart's Law starts eating the proxy.

How do CMOs and CROs stop the "whose number is right" debate in the boardroom? Pick one outcome, usually pipeline created and revenue closed, and measure both teams' contribution backward from there. When marketing's number and sales' number are derived from the same scoreboard rather than from two different models, the debate moves from credibility to action. The Debate Tax disappears when there is only one source of truth left to argue with.

What's the fastest way to spot Goodhart drift before it shows up in the quarterly number? Watch the leading indicators of the indicator. If MQL volume is the metric, watch MQL-to-SQL conversion. If pipeline coverage is the metric, watch deal slippage and average deal age. When the proxy moves but the indicator beneath it moves the wrong way, the proxy is being managed instead of the underlying motion. That gap is usually visible a quarter before the headline number breaks.

How does outcome-backward measurement change the relationship between marketing and sales? It removes the model from the argument. Most marketing-sales conflict is not about strategy, it is about whose attribution view gets to define credit. When credit is assigned from the outcome backward, using the same data both teams trust, the conversation shifts from defense to coordination. Marketing stops over-claiming. Sales stops dismissing. Both teams start optimizing for the same thing because they are looking at the same thing.

Can a platform actually fix Goodhart drift, or is this a people and process problem? Both, in that order. The people and process problem is real: someone has to be willing to demote the comfortable proxies and tie comp to the harder outcome. A platform cannot make that decision. But once the decision is made, the platform is what makes outcome-backward measurement operationally possible at scale, by unifying the data across marketing, sales, and finance so there is only one scoreboard to look at. Without the unified data layer, outcome-backward measurement is a slogan. With it, the slogan becomes a system.

‍