Most business problems announce themselves loudly. Revenue declines. Customers leave. Costs climb. The dashboard turns red, and people pay attention.
Some issues arrive disguised as success instead, and those are often the most dangerous, precisely because an improving number creates confidence, and confidence reduces scrutiny at the exact moment scrutiny would help most.
The pattern repeats across very different businesses. Traffic grows. Engagement improves. Lead volume climbs. The dashboard looks healthy, the team feels encouraged, and underneath the improvement something else is quietly deteriorating. Lead quality weakens. Retention softens. Brand perception erodes a little at a time. The visible metric goes up. The business underneath it doesn’t get healthier alongside it, and for a while nobody notices the two have decoupled.
This happens because no metric exists in isolation. It sits inside a system, and optimising one part of that system often creates pressure somewhere else that simply hasn’t shown up yet. A company doubles lead volume and marketing celebrates, while sales quietly struggles with longer cycles and weaker conversion. A platform lifts engagement, and only later discovers users are spending more time because tasks got harder to finish, not because the experience improved.
What makes an improving metric more dangerous than a declining one isn’t the direction. It’s the asymmetry in how organisations respond to each. Decline triggers a meeting. Somebody owns the problem, a root cause gets demanded, and the investigation happens whether anyone wants it to or not. Improvement triggers the opposite response: the assumption that the explanation is already understood, because the number is doing what everyone wanted it to do. Nobody schedules a root-cause analysis for a metric that’s going the right way. That asymmetry, not the metric itself, is the actual mechanism that lets a real problem hide behind a good one for months.
There’s a specific window in a company’s life where this asymmetry gets even more dangerous, and it’s worth naming directly: right after a leadership change, a new hire’s first initiative, or a freshly launched strategy. In that window, everyone in the room has a strong, mostly unconscious incentive to attribute any improving number to the new thing, because doing so validates the change that was just made and the person who made it. A new head of growth arrives, a metric ticks up three weeks later for reasons that have nothing to do with them, seasonal timing, a competitor’s outage, a delayed cohort finally converting, and the improvement gets narrated as early proof the new strategy is working. Nobody is lying. Everyone genuinely wants the story to be true, which is exactly what makes this the single worst moment to skip the causal check, and exactly the moment it’s most likely to get skipped, because questioning the win feels like undermining the person who just arrived.
The better operators I’ve watched get more curious when a number moves sharply in the right direction, not less. What’s actually driving this. Is it sustainable. What else changed in the same window that might explain it better than the story everyone’s already telling themselves. An unusually positive movement is often just as informative as an unusually negative one, sometimes more, because nobody is looking for the problem inside a number that’s going the right direction, which means whatever’s hiding there gets a longer runway before anyone finds it.
There’s a second specific version of this worth naming: an improving metric that was caused by making something else worse, where the causal link is real but invisible from where either team is sitting. A support team resolves tickets faster by closing them earlier, and resolution-time improves while the customer’s actual problem often isn’t solved, just reclassified as solved. A growth team lifts sign-up rate by simplifying onboarding, and activation quietly drops because the people who sign up now understand the product less at the point of entry. Each team’s dashboard is honestly reporting good news. The connection between the two dashboards is where the actual story lives, and it’s structurally invisible unless someone is specifically looking across the boundary between them, which almost nobody’s job description includes.
The test worth applying whenever a metric improves significantly: name the other number that should have moved alongside it if the improvement were genuinely healthy, then go check whether it did. And separately, ask what else changed in the organisation at the same moment the number started improving, whether a new hire, a new initiative, or simple timing, that might be quietly claiming credit for something it didn’t actually cause. If the second number didn’t move, or moved the wrong way, or if a more boring explanation fits the timing better than the story currently being told, the improving metric isn’t evidence the business got better. It’s evidence something shifted, and shifting is not the same claim as improving, no matter how similar the two look on a slide, and no matter who’s in the room hoping it’s true.
Businesses are naturally wired to worry about numbers heading the wrong way, and they should be. But some of the most expensive strategic mistakes happen when an improving number goes completely unquestioned, because deterioration announces itself and false confidence never does. The metric creating the most optimism in the room is sometimes also the one quietly hiding the most risk, and the only reliable defence against that is treating good news as a question rather than as a verdict, especially when someone in the room badly wants it to be one.
Every business has its own version of this story. If you're working through something similar, I’d love to hear from you. Whether it's to exchange ideas, brainstorm a challenge, or just have a thoughtful conversation, feel free to reach out at [email protected].