What to Fix First When Your Sustainability Metrics Become Performance Theater

You're sitting in a quarterly review. The sustainability dashboard glows green across every KPI. Carbon intensity down 12%. Water use per unit flat for three years. Someone from marketing is already drafting the press release. But you've got this knot in your stomach—the numbers look too clean, too polished, like a stage set where everything is bolted down so nothing can wobble. That knot is your best diagnostic tool.

I've spent the last decade building and auditing sustainability analytics pipelines—first at a Fortune 500 energy firm, then as an independent consultant for mid-market manufacturers. What I've learned is that the prettiest dashboards often mask the ugliest data. When metrics become performance theater, the first thing to fix isn't the metric itself—it's the incentive structure that turned it into a prop. But since you can't fix culture overnight, here's where to start cutting.

Where This Performance Theater Shows Up in Real Work

The quarterly review that felt off

You sit through it every three months. Someone projects a slide deck — green arrows everywhere, a carbon intensity line that slopes down like a gentle ski run, and a box labeled 'on track.' The room nods. Then you walk back to your desk and notice the raw data feed has been returning nulls for the last six weeks. The green line? A flat interpolation from a stale baseline. Nobody caught it because nobody looked past the dashboard. That's performance theater: the meeting feels productive, the metrics look reassuring, and the actual pipeline is hemorrhaging accuracy. I have seen teams burn three quarters chasing a 'net-zero operations' badge while their underlying compute efficiency flatlined. The theater works because it rewards the signal of progress — not the grind of fixing broken instrumentation.

The odd part is — the people running the show are rarely malicious. They inherited a dashboard built by an intern who left, or they optimized for a single stakeholder who asked for 'fewer red numbers.' So the data gets smoothed. Outliers vanish. The story tightens up. And the real problem — a metastasizing data gap — stays invisible behind a perfect PowerPoint arc.

How a green dashboard hid a data pipeline failure

I once consulted for a SaaS company whose sustainability page bragged about a 12% emissions drop. The CEO tweeted it. The press picked it up. Then the engineering team admitted they had changed their cloud billing integration — and the new connector silently dropped half the regions. The dashboard never flagged it. No alert fired. The graph simply looked better. That's the signature of a green dashboard concealing rot: it never shows a breakdown by region, by service, or by time lag. It shows one big happy number going down.

'We were celebrating a reduction that existed only because our data stopped traveling.'

— VP of Engineering, after the post-mortem

The trap is seductive. A cleaner dashboard means fewer questions in exec reviews. Fewer questions mean less friction. Less friction means faster approvals for next quarter's budget. But what you're actually buying is a delay — the pipeline failure will surface eventually, usually during an audit or a public sustainability report. By then the trust is gone, and the fix costs twice as much because you have to rewind eighteen months of corrupted baselines.

The moment you realize your baseline was cherry-picked

You compare this month to 'the baseline year' — 2021, say. But 2021 was a pandemic year. Office buildings sat empty. Travel evaporated. Data centers ran at 40% utilization. Using that as your reference point is like measuring a marathon runner's time against a sprinter who pulled a hamstring. That hurts — not because the numbers are wrong, but because they mislead honestly. The baseline was chosen because it looked good, not because it was representative.

Most teams skip this: they never ask who set the baseline, what data they excluded, and whether the methodology would survive a skeptical journalist's scrutiny. The catch is — once you publish a baseline, changing it feels like admitting fraud. So you lock it, polish the trend line, and hope nobody recalculates from raw logs. Performance theater thrives on that locked-in inertia. The fix is brutal but simple: re-baseline every two years with transparent criteria, and publish the exclusion list. Ugly numbers upfront save you from a prettier lie later.

Foundations Readers Confuse: Metrics vs. Indicators vs. Goals

Why 'metric' and 'indicator' are not synonyms

Most teams use these words interchangeably. That's the crack where performance theater seeps in. A metric is a raw number — kilowatt-hours per server rack, liters of water per batch. It describes what happened. An indicator tells you whether that number means something good or bad relative to a threshold you actually care about. I have watched engineering leads present monthly energy drops as proof of sustainability progress, only to discover the drop came from shifting compute to a dirtier grid at night. The metric improved. The indicator — the thing that signals real environmental impact — got worse. The catch is that nobody defined the indicator in the first place.

Wrong order. Teams lock in a metric, slap a goal on it, and call the work done. They never ask: what would this number have to do before we change our behavior? That question is what turns a metric into an indicator. Skip it, and you're running on vanity.

The difference between tracking and managing

Tracking is passive. You collect data, plot it, maybe share it in a dashboard. Managing means you have a lever — something your team can actually pull when the number goes sideways. I once consulted for a logistics company that tracked carbon per parcel down to the gram. Beautiful chart. But the data lagged by six weeks, and nobody on the operations team could adjust routes mid-week. They were tracking. They were not managing. The dashboard became a museum piece, not a steering wheel.

That hurts. Because the team spent months building the tracking infrastructure, and when the board asked for action, all they could show was a trailing indicator with no causal chain back to their daily decisions. The distinction matters: if you can't change the metric within one planning cycle, you're measuring for show. The fix is not better data — it's shorter feedback loops and a clear operator who owns the response.

How goals corrupt metrics unless you watch the denominator

Goals create pressure. Pressure finds the path of least resistance. In sustainability analytics, that path is almost always the denominator. A team targets reducing water use per unit of production. Smart. But production volume drops, and suddenly the ratio looks worse through no fault of the team. Or — more commonly — the denominator inflates. I have seen a factory reclassify what counts as "production" to keep the ratio flat. The metric stayed green. The absolute water draw ticked up. The goal corrupted the indicator because nobody audited what the denominator actually captured.

“A ratio that never goes red is not a measure — it's a story you have already decided to believe.”

— comment from a facilities director during a post-mortem I attended, after we found the phantom denominator.

The hard truth: absolute numbers are harder to game. They expose real strain. But they also punish teams for growth, which is why ratios dominate sustainability reporting. The trade-off is constant vigilance. Every quarter, ask: is the denominator still honest? Has our definition of the unit shifted? If nobody can answer without checking three spreadsheets, the goal has already consumed the metric. Strip it back to raw counts for one cycle, see what surfaces, then rebuild with denominators that someone outside the team can verify in under five minutes.

Patterns That Usually Work: Honest Sustainability Signals

Absolute emissions plus intensity with fixed baselines

The pairing that holds up best under pressure is simple: report your absolute metric (tons of CO₂e) alongside an intensity ratio (tons per revenue dollar, per square foot, per widget shipped). And then lock the baseline year. Don't rebaseline because you bought a fleet of electric vans and want a better denominator—teams that shift baselines every eighteen months lose the ability to detect real drift. I have watched infrastructure teams spend a full sprint recalibrating a baseline, only to discover their absolute emissions actually rose 7% while the intensity ratio looked heroic. That hurts. The catch is that intensity can hide absolute growth if your business is scaling fast. You need both numbers visible on the same dashboard, side by side, no toggle. One drives accountability; the other shows efficiency. When a team asks “which one do we optimize for?” the honest answer is both, and the tension between them is the signal.

Third-party verified scope 1 and 2 data

Scope 3 is where performance theater thrives—it's noisy, full of spend-based proxies, and easy to prune. So fix what you can actually touch. Scope 1 (your direct fuel burn, refrigerant leaks) and scope 2 (purchased electricity, steam, heat) are measurable with meters, utility bills, and a little algebra. The pattern that works: get those numbers audited or verified by a third party annually. Not a carbon consultancy that also sells offsets—an actual assurance provider. The odd part is—teams often resist because verification costs money and exposes mistakes. But a verified number that goes up year-over-year is still more useful than a glossy estimate that stays flat. I have seen two orgs share identical cloud workloads; one reported scope 2 at 40 tons, the other at 120 tons. The difference wasn't efficiency—it was that the lower number used a market-based renewable certificate trick. Third-party verification catches that. It doesn't fix the problem, but it kills the illusion.

“We dropped our carbon footprint by 60% in one quarter — by switching from actual meter data to a free online calculator. Then we had to explain why the utility bill disagreed.”

— Cloud operations lead at a mid-market SaaS firm, reflecting on a 2023 audit cycle

Year-over-year transparency even when numbers go up

Most teams publish only the good years. The pattern that builds trust is publishing every year—up years included—with a short narrative about why. Higher cooling load from a new colo facility. A manufacturing line that ran 24/7 for a product launch. A colder winter that spiked natural gas use. Those explanations matter more than a flat green arrow. The pitfall here is psychological: leadership may panic when they see a red bar and demand the metric be “adjusted.” Resist that. Instead, show the same chart with a rolling three-year average overlaid. The average smooths noise without hiding the spike. One rhetorical question worth asking your team: would you rather defend a real number that went up 9%, or explain next year why your “optimized” number no longer matches any invoice in the ERP system? Wrong order. Start with transparency, then build the narrative. Teams that do this find that even their investors stop asking for tweaks—because the data has a spine.

Honestly — most data posts skip this.

Honestly — most data posts skip this.

Anti-Patterns and Why Teams Revert to Them

The 'improvement' that comes from rebaselining every year

We fixed it. That's what the slide said. Carbon intensity dropped 14% year over year, the board nodded, and everyone moved on. Except nothing actually changed in the data center. The improvement was a mirage—we had simply rebased the baseline. The tricky bit is that rebaselining isn't always wrong; when you acquire a company or retire a legacy plant, you adjust. But I have seen teams rebase because the old number was "too embarrassing" or because a new software tool auto-set a different reference year. That hurts. You lose the thread, the year-over-year comparison becomes meaningless, and suddenly your 14% drop is just spreadsheet theater. The organizational reason is simple: nobody wants to explain to the CEO why a real metric went up. So we change what we measure against. We call it "normalization." The catch is—normalization without disclosure is just hiding. One concrete fix: publish the baseline year and recalc rules in the same dashboard. If you change it, flag it. Red text. A human note. Otherwise, your improvement is a fiction your team will have to defend—and fail to—when someone actually audits.

Baselines are commitments, not suggestions.

Scope 3 estimates that never get audited

Most teams skip this: verifying their supply chain data. Scope 3 emissions are often 80% of the total footprint, yet I see teams spend two hours polishing a Scope 1 chart and zero hours validating the supplier spreadsheet that was last updated during the Obama administration. The estimates become a black box—spend-based multipliers, industry averages, a few emails to vendors that nobody followed up on. The anti-pattern is clear: treat estimates as if they were measurements. When we fixed this at one mid-size firm, we found that their "16% reduction" in purchased goods was entirely driven by a supplier who had simply stopped responding. They had imputed zero. Zero emissions. That's not a reduction, that's a data gap masquerading as progress. The organizational reason teams revert here is that auditing costs money and creates friction. Procurement doesn't want to bother suppliers. Finance doesn't want to pay for third-party verification. And the sustainability team, already understaffed, prioritizes the dashboard that keeps the board happy over the spreadsheet that reveals the truth.

"We stopped asking because the answers were inconsistent. So we used averages. The averages made us look good. So we kept using them."

— Former sustainability analyst, manufacturing company

Wrong order. You fix the inconsistency, not the comfort.

Why teams choose a pretty dashboard over a messy truth

The dashboard was beautiful. Real-time widgets. Pulse animations. Color-coded threshold alerts. And completely disconnected from the operational reality—the data pipeline feeding it had a known bug in the emissions factor library that nobody had patched for eight months. Why? Because fixing the bug would require admitting the dashboard was wrong for two quarters. That's an awkward conversation. The odd part is: the prettier the dashboard, the harder it's to challenge. Visual polish signals competence, even when the underlying data is rotten. Teams revert to this because a clean interface buys them time. It projects control. Meanwhile, the messy truth—a clunky CSV with annotated footnotes and manual overrides—sits in a shared drive that nobody opens. I have seen this pattern across three different organizations. The fix is not to ban dashboards. The fix is to make the data lineage visible on the same screen: a small "source stale" badge, a last-audit date, a link to the raw extract. Ugly is fine. Unexplained polish is the enemy.

One rhetorical question for the road: would you rather explain an honest flat line or a beautiful trend that later collapses under scrutiny? Your board will forgive the flat line. They won't forgive the scandal.

Maintenance, Drift, and Long-Term Costs of Performance Theater

Dashboard rot: when nobody remembers how the number is calculated

It starts innocently. A junior analyst leaves.

Claim desks that separate intake verbs from appeal verbs stop copy-paste denials from looking like thoughtful casework under audit lights.

The person who knew the SQL logic for ‘carbon per transaction’ moves to a different team.

Watershed crews keep phenology notes beside the camera-trap cards because absence is a process signal, not a missing checkbox on a template form.

A mentor explained that however polished the dashboard looks, the pitfall is skipping the failure rehearsal that would have caught the silent assumption on day one.

Six months later, the number still appears green on the dashboard — but nobody alive can explain the denominator. I’ve walked into rooms where three people gave three different answers for the same KPI.

Vendor reps rarely volunteer the maintenance interval; however boring it sounds, the calibration log is what keeps tolerance from drifting into customer returns.

Kill the silent step.

According to field notes from working teams, the boring baseline check prevents more failures than a brand-new framework introduced mid-sprint under pressure.

One cited an email from 2022. Another guessed it was an API field. The third shrugged. That's not a measurement system. That's a fossil.

The decay accelerates. Someone adds a new data source but forgets to adjust the join. A filter drifts off — accidentally excluded scope 2 emissions for two quarters. The number stays flat, the board applauds, and your infrastructure silently lies to itself. The worst part? You can't spot the rot by looking at the dashboard. It looks fine. Green arrow up. Gold star. The cost is invisible until your auditor asks: “Show me the lineage for this cell.”

Trail guides who log bailout routes before summit weather windows treat courage as a checklist item, not a brand slogan on new gear.

I fixed one of these by tracing a single metric through five handoffs. It took three weeks. The team had been celebrating ‘stable’ energy intensity for eighteen months. The number had been frozen — literally hardcoded in a spreadsheet — for fourteen of those months.

Not every data checklist earns its ink.

Not every data checklist earns its ink.

‘We thought it was steady because the chart didn’t move. We didn’t realize the chart had stopped moving for the wrong reasons.’

— engineer from a logistics firm, reflecting on a scope 3 recalculation

The hidden cost of audit failures

Performance theater doesn't stay inside your company. It leaks. An auditor finds one unexplained jump — maybe your carbon intensity dips 12% while production rose 8%. Easy to celebrate. Hard to defend when the auditor asks for the month-over-month reconciliation and discovers the dip came from a data cutoff error, not efficiency. That single finding can cascade: extended audit scope, flagged internal controls, and a footnote in the sustainability report that investors actually read.

Wrong sequence entirely.

It adds up fast.

The odd part is — the fix is cheap. A lineage document. A weekly drift check. A single cron job that flags when a metric changes more than 5% without a human comment. Most teams skip this. They invest in the dashboard paint job, not the plumbing. Wrong order. The audit failure cost is rarely the penalty. It's the months of trust rebuilding, the redirected team time, and the quiet question every future report carries: “What else did they miss?”

One logistics firm I worked with lost a year of credibility over a single misaligned unit conversion. kWh versus MWh. A decimal slip.

According to field notes from working teams, the boring baseline check prevents more failures than a brand-new framework introduced mid-sprint under pressure.

Kill the silent step.

Three days of auditor embarrassment. Two years of enhanced scrutiny.

Most teams miss this.

Most teams miss this.

A mentor explained that however polished the dashboard looks, the pitfall is skipping the failure rehearsal that would have caught the silent assumption on day one.

The theater was free. The cleanup was not.

How theater becomes institutional memory and poisons future data

Here is the real trap: repeated manipulation doesn't vanish. It hardens into precedent. New hires see the old reports, assume the methods are correct, and repeat them. A metric that started as a quarterly estimate becomes accepted as ground truth. A five-year trend line built on fudged scope 3 data becomes the baseline for net-zero targets. That hurts. You're now aiming at a moving target that was never real to begin with.

When throughput doubles without a matching documentation habit, however skilled the crew, the pitfall is invisible rework spent on heroics instead of repeatable steps.

The pattern is self-reinforcing. Teams stop questioning the number because the number has always looked that way. Questions sound like attacks. “Are you suggesting our carbon intensity never actually improved?” — yes, that's exactly what I am suggesting. The theater has memory. It remembers the shortcuts, the manual overrides, the ‘temporary’ fixes that ran for three years. You can't clean it with a one-time data sweep. You need to rebuild the culture of what counts as evidence.

What breaks first is the ability to set honest goals. If you can't trust the historical baseline, your future targets are theater too. A 30% reduction by 2030 means nothing if 2022 was a fiction. I have seen teams spend six months arguing over a baseline, not because the data was complex, but because no one wanted to admit the old number was wrong. The cost of performance theater compounds — every decision built on rotten data creates another layer of rot.

Try this quarter: pick your most celebrated sustainability metric. Find the person who can explain exactly how it's calculated, end to end, without opening a terminal. If that person doesn't exist, you have found your first fix. Not a new dashboard. Not a prettier chart. A single document and one honest conversation about what the number actually means. That's the maintenance your infrastructure needs. Not more theater.

When Not to Fix Your Metrics First

If the leadership team is still in denial about baseline integrity

You can build the most elegant carbon-intensity dashboard on the planet. If the C-suite still insists that last year's spreadsheet 'seems fine' — and they have no appetite for reconciling meter data against utility bills — your shiny new metric will rot inside a slide deck. I have watched a team spend three months engineering a real-time energy tracker, only to have the CFO reject it because it showed a 14% increase versus the 'official' number they had been reporting to the board. The metric was correct. The organization was not ready. The fix was not technical; it was a slow, brutal campaign of one-on-one coffee meetings with people who had built careers on that wrong baseline. Until those conversations stick, don't touch the dashboards. You will waste your credibility on a tool nobody trusts.

The hard truth: perfect data, wrong audience. Fix the denial first. Then the metric.

Not every data checklist earns its ink.

Not every data checklist earns its ink.

What does 'readiness' actually look like? It looks like someone in the room saying, 'I think our Scope 2 numbers are off by at least 30%' — and not getting fired for it. That signal is rare. If you don't hear it, your metric work is performance theater for a theater that already has a full house. Wait.

If your data infrastructure is held together by spreadsheets

Spreadsheets are not infrastructure. They're memory foam — comfortable, moldable, and utterly incapable of structural load at scale. Yet I keep walking into organizations where the entire sustainability analytics pipeline is: an ERP export → a shared Google Sheet → a PowerPoint slide → the board. The person who built that spreadsheet left two years ago. Nobody knows which cells are hardcoded. The seam blows out every quarter when someone accidentally sorts a column.

Most teams skip this: they assume they can 'clean up the data later' while building the flashy metric layer now. Wrong order. You end up with a gorgeous visualization that reports a number nobody can reproduce. That hurts. Reproducibility is the cheapest insurance against performance theater — and spreadsheets can't provide it.

The catch is that replacing spreadsheets feels unglamorous. No one gets promoted for 'we now have a schema.' But the alternative is worse: you build a metric that shifts by 8% every time someone re-downloads the same report. Regulators haven't started looking yet — but they will. And when they ask for an audit trail, your spreadsheet-wielding analyst will cry.

‘We had to explain to a consultant why our carbon footprint dropped 20% in one month. It was a VLOOKUP pointing at the wrong tab.’

— former sustainability manager, European logistics firm

Fix the pipes. Then paint the pipes. The metric can wait.

If regulators haven't started looking yet (but they will)

This one stings because it feels like procrastination disguised as wisdom. But here is the pattern I have seen: teams that race to publish perfect-looking sustainability metrics before any enforcement body has defined a standard spend the next two years rewriting everything. The EU's CSRD, California's SB 253, the ISSB standards — they're all moving targets. If you harden your metrics now against a regulation that shifts next quarter, you build technical debt at 18% interest.

That doesn't mean do nothing. It means build traceable data, not polished data. Keep your raw inputs raw. Tag them with timestamps and source notes. Resist the urge to normalize, smooth, or 'improve' the numbers for presentation. The polish is the trap. Regulators, when they finally arrive, care about lineage — not how clean your line chart looks.

One rhetorical question, sparingly: would you rather explain why your number is messy but honest, or why your clean number was derived from a methodology you changed in secret last November? The second conversation is the one that ends careers. So wait on the glossy layer. Invest in the audit trail. Your future self — the one getting the regulatory inquiry — will thank you.

Open Questions and FAQ: What Still Bothers Practitioners

Is it ever okay to normalize without disclosing absolute numbers?

Short answer: yes, but only inside an agreed-upon frame where the audience already knows the denominator. I have seen teams publish carbon-per-transaction charts—beautiful downward slopes—while total transaction volume tripled. The normalized curve looked like progress. The absolute line told a different story: emissions actually rose. The catch is that raw numbers can leak competitive data or confuse readers who lack context. One workaround: publish both normalized and absolute views side-by-side, with a two-sentence explanation of why the normalized metric moves the way it does. That builds literacy, not theater. The moment you hide the denominator, you're asking for trust you have not earned.

The trickier edge case: internal dashboards. If your engineering leads already know transaction counts, normalized data alone can signal efficiency gains. But if the board sees only grams-per-request while headcount and cloud spend balloon, you're manufacturing goodwill. Honesty costs nothing here—add a second axis. Or a footnote. Just show the raw number somewhere humans can find it.

‘Normalization without context is not simplification. It's selective storytelling dressed as rigor.’

— infrastructure lead, anonymous practitioner survey

How do you prove a metric is theater when the data looks correct?

This is the one that stings. You run the query. The numbers match. The trend is green. Everything checks out—yet something feels hollow. The fault is rarely arithmetic. It's almost always in the metric’s relationship to reality. Start by asking: What would happen if this metric improved but everything else stayed the same? If the answer is “nothing meaningful changes in the system,” you have found a proxy that lost its proxy status. I once audited a dashboard where “CPU utilization per workload” had dropped 40%. Great, said everyone. Except workloads had been silently migrated to less efficient instance families, and total energy spend climbed. The data was correct. The metric was theatre.

The fastest litmus test: trace the metric back to a physical or financial outcome. Can you walk from the dashboard number to a meter reading or a vendor invoice? If the chain breaks, you have a KPI cosplaying as an insight. Another signal—ask a junior engineer to explain what the metric measures in one plain sentence. If they stumble, the metric has already lost its communicative purpose. That's not a training problem. It's a design failure.

What’s the single fastest way to rebuild trust in a corrupted dashboard?

Delete something. Not a whole dashboard—one tile. One chart that everyone knows is wrong. Do it in a public channel with a note: “This line was misleading. We will replace it with actual utilization data by next sprint.” That act costs nothing but signals more integrity than any polished footnote ever could. I have seen teams restore credibility in two weeks using this move. The trick is to replace the deleted tile with something ugly but honest—a raw hourly spend graph, no smoothing, no averaging. Let people see the noise. Then invite the team to argue about what to clean up next. The fastest fix is not a fix. It's stopping the lie.

Wrong order: convene a steering committee, design a new index, validate for three months, roll out a v2 dashboard. That rebuilds process, not trust. Right order: admit the seam blew out, show the raw data, and let the team watch you fix the next seam in real time. That rebuilds trust. Then you can fix the metrics.

Summary and Next Experiments to Try This Quarter

Triage audit: start with the intensity denominator

Every sustainability metric hides a denominator — revenue, square footage, compute hours, active users. That denominator is where the rot usually starts. I have seen teams celebrate a 20% drop in energy per transaction while total energy surged 40% — because transactions exploded. The ratio improved; the planet didn’t. Pull your top five metrics this week and ask: is the denominator growing faster than we admit? Strip out the normalization and stare at the raw absolute. If that number climbs while your ratio falls, you're performing theater. The fix is brutal but cheap: report both numbers side by side for one quarter. No commentary. Just data.

The catch is — absolute numbers are uncomfortable. They expose growth. They make your carbon-per-dollar victory look hollow when total dollars double. That’s the point.

Run a 'theater test' on your top five metrics

Gather the team. Pick five metrics you currently report. For each, write one sentence about what behavior it actually rewards — not what it claims to measure. Example: “Energy per request” rewards teams that batch more requests into fewer servers (good) but can also reward teams that stop measuring idle capacity (bad). The exercise takes ninety minutes. It surfaces assumptions that have been fossilized in dashboards for years. One team I worked with discovered their “renewable energy percentage” had been tracking certificate purchases three regions away from their actual workload. Wrong order. They had been celebrating paper while burning coal.

‘A metric that can't be gamed has not been invented yet. The goal is to make gaming costly enough that teams stop.’

— sustainability engineer at a cloud provider, off the record

Publish one ugly number with a candid narrative

Pick something broken — a datacenter with rising PUE, a product line where hardware utilization is 12%, a Scope 3 estimate that you know is wrong. Publish it internally. Write three paragraphs explaining why it's ugly, what you're doing about it, and what you don’t know yet. No spin. No aspirational target. The effect is almost always restorative: people stop hiding, start asking better questions, and the political cost of reporting bad news drops. One quarter of that, and you will see which metrics were being propped up by silence.

The tricky bit is — this requires executive cover. Without it, someone will frame the ugly number as a failure rather than a diagnostic. Run it as an experiment in a single business unit first. If the reaction is blame instead of curiosity, you have a culture problem, not a metric problem. Fix that before you touch another dashboard.

Next quarter’s experiment: take one metric off your public report entirely. Replace it with a sentence: “We stopped tracking X because it rewarded Y behavior.” See who complains. See what they say. That noise is the signal you need.