Watch what happens in a planning meeting.
For the first twenty minutes, people talk about customers. Someone quotes a support ticket. Someone describes a call where a buyer got stuck on the third step and never came back. The conversation is specific, a little messy, and full of actual people.
Then a number goes on the board. Activations up 30% by Q3.
From that moment, nobody mentions the buyer again. The rest of the meeting is about ownership, dependencies, and reporting. The person who got stuck on the third step has left the room, and no one noticed them go.
I have watched this happen at every company I have worked at, and I have done it myself. It is not a character flaw in the people around the table. It is a predictable swap, and it has a name.
The swap has a name
Michael Harris and Bill Tayler wrote about this in Harvard Business Review in 2019. They call it surrogation: because strategy is abstract, people mentally replace it with the metric meant to measure it. The metric stops representing the goal and quietly becomes the goal.
Their example is Wells Fargo. The strategy was long term customer relationships. The metric was cross sales. Employees opened 3.5 million accounts without customer consent, which destroyed the exact relationships the strategy was built around. Nobody in that building woke up wanting to defraud customers. They were chasing the number that stood in for caring about customers.
The mechanism underneath comes from Kahneman and Frederick’s work on attribute substitution. When a question is hard, we answer an easier one and rarely notice the substitution. “Are we making it easier for dealers to sell a car?” is hard. It needs judgment, context, and someone willing to disagree. “Are activations up 30%?” is easy. It has an answer, and the answer is in a dashboard.
Goodhart’s law describes what comes next, when a measure becomes a target and people game it. Surrogation happens earlier and quieter. Nobody games anything. The team simply starts answering a different question than the one it set out to answer, and everyone stays busy.
Why empathy goes first
Empathy is the most expensive thing in the room. It is slow, it resists summary, and it usually complicates the plan. The number is cheap. It fits on a slide, it settles arguments, and it survives being forwarded to someone who was not in the meeting.
There is lab work pointing the same way. Small, Loewenstein, and Slovic ran field experiments in 2007 showing that people primed to think in calculations gave less to an identified person in need than people primed to feel. Worth flagging that a preregistered replication in 2023 found no support for the effect, so treat it as suggestive rather than settled.
You do not need the lab version anyway. You can watch it happen in your own reviews. The moment a number exists, it becomes the shortest path to sounding competent. Talking about a buyer’s problem takes ninety seconds and invites questions. Reporting a percentage takes five and closes the topic.
The clearest test I know is one question. Ask the team who is worse off if you miss this target. If the room goes quiet, or if the honest answer is “our quarterly review,” the number is already running the project.
What the swap actually costs
Three things break, and they break in order.
The roadmap tilts toward what moves the number. Measurable beats important every time there is a target attached. The fix that would help most sits below the line because its impact is hard to attribute, and three smaller items that reliably nudge the metric get built instead.
The messaging gets written for the metric. This one is mine to own as a product marketer. Write copy aimed at a signup target and you will hit it, because overpromising works. Then retention drops two quarters later and someone blames onboarding. The message map exists to stop this, but only if the pillars are tied to a buyer’s problem rather than to the funnel stage you need to move.
Reviews turn into activity reports. Everyone hit their number. The launch shipped. The campaign ran. Nothing changed for the customer, and nobody in the review is positioned to say so, because saying so means arguing with a green dashboard. A win/loss retrospective is the cheapest way back, since it puts a real deal and a real reason back on the table.
Keep a person attached to the number
None of this is an argument against targets. Work without numbers drifts, and product marketing that cannot prove impact gets treated as a support function. The point is to stop the number from eating the reason it exists. Five things that hold the line.
1. Write the goal as a sentence before you write it as a number. I use one structure: for [this person], with [this problem], we believe [this change] will [this outcome], and we will know because [this number]. The number lands last and stays in the evidence position. If the sentence is hard to write, the goal is not ready, and finding that out in planning is cheap.
2. Pair every target with a counter metric. Research on surrogation shows people substitute less when they are held to several metrics of a strategy instead of one, because no single metric can plausibly stand in for the whole thing. Pair activations with what happens after activation. Pair pipeline with win rate. Pair adoption with the support volume it creates.
3. Let the people who execute help set the target. Same source, same finding: involving implementers in shaping the strategy makes them less likely to swap it for a metric. In practice this means the team that will hit the number needs to have argued about it first. A target handed down complete gets optimized. A target the team helped build gets questioned, which is what you want.
4. Loosen the link between one metric and money. Tying compensation tightly to a single number raises the odds of surrogation. That is a pay design problem more than a marketing problem, but it is worth naming when you see it, because no amount of good messaging survives a comp plan pointed the other way.
5. Put one customer voice in every review. A call recording, a verbatim from a lost deal, a support thread. Thirty seconds is enough. It costs almost nothing and it restores the question the number replaced, which is the whole job.
What it looks like when it works
The pricing and packaging redesign that raised average revenue per user by 15 to 20% did not start as a revenue target. It started as a problem: buyers were picking between tiers that both fit badly, and sales spent half of every call explaining a catalogue instead of a fit. The revenue number was how we knew the fix had worked.
Run it the other way and the outcome changes. Open with the revenue target and the obvious move is to raise prices and repackage the same confusion at a higher price point. The number goes up either way for a quarter or two. Only one version leaves the buyer better off.
Same principle behind cutting a portfolio from seven products to three. Nobody sets a goal of owning fewer products. You set a goal of letting a buyer understand what they are buying in one conversation, and the portfolio math follows from that.
The part that is easy to miss
Empathy in goal setting is not a soft addition to the plan. It is the only thing that tells you if the number was worth hitting. Strip it out and you still get a result, still get a green dashboard, and still get a team that worked hard. You just lose the ability to tell a win from a very well executed mistake.
Put the person back in the sentence. The number will still be there when you need it.
Most go to market problems I get called into started as goal problems. The numbers were set before anyone agreed who the work was for, and the positioning, the launch, and the sales story all inherited that gap. I help B2B software companies build or fix their go to market when positioning is unclear, launches don’t land, and sales can’t explain what makes them different. Contact me at zackalami.com/#contact.
Sources
- Michael Harris and Bill Tayler, “Don’t Let Metrics Undermine Your Business,” Harvard Business Review, September 2019. Source of surrogation, the Wells Fargo case, and the three countermeasures on multiple metrics, implementer involvement, and incentive design.
- Daniel Kahneman and Shane Frederick, “Representativeness Revisited: Attribute Substitution in Intuitive Judgment,” in Heuristics and Biases: The Psychology of Intuitive Judgment, Cambridge University Press, 2002. Cited by Harris and Tayler as the mechanism underneath surrogation.
- Deborah Small, George Loewenstein, and Paul Slovic, “Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims,” Organizational Behavior and Human Decision Processes, 2007. Free PDF here. A 2023 preregistered replication by Maier, Wong, and Feldman, published in Collabra: Psychology, found no support for the effect, so it is treated here as suggestive.
- Charles Goodhart, “Problems of Monetary Management: The U.K. Experience,” 1975, generalised by Marilyn Strathern in 1997 into the phrasing most people quote. Background on both.
