What Your Biodiversity Audit Framework Assumes About Recovery: A Workflow for Testing Successional Logic

You hand a biodiversity audit report to a client. They glance at the recovery score—green, 72 out of 100—and nod. But what does that number actually mean? Most audit frameworks assume a particular story about recovery. That species return in a predictable order. That soil microbes follow vegetation. That if you wait long enough, the system will converge on a historic baseline. These aren't neutral technical choices; they're bets on how succession works. And in my experience auditing post-mining sites in Western Australia and coastal restoration in the Gulf of Mexico, those bets often lose.

This article is a field guide for testing the successional logic inside your audit framework. Not to replace it, but to know when it's lying to you. We'll walk through a workflow—from spotting hidden assumptions to running counterfactuals—and I'll show you where most frameworks get recovery wrong.

When the Scorecard Ignores What Actually Happens in the Field

The Scorecard Says 68 — The Field Says Zero

I once stood on a Pilbara rehabilitation plot where the audit dashboard glowed green. Score: 68 out of 100. Vegetation cover? Check. Species richness? Above threshold. The framework declared the site on track. But I was ankle-deep in bare red dust, and the only perennial grasses I could find were dead stumps. The audit counted presence — a few tufts clinging to gullies counted as a hit. It missed function: no root binding, no litter trapping, no seed bank. The scorecard logic had assumed that if indicator species show up, succession follows. Wrong order. That plot had been stuck in an annual-weed cycle for four years, and the audit never knew.

The catch is systemic. Most frameworks score what is easy to count — number of native species, percent cover, structural layers — and then weight those counts against a reference benchmark. That sounds fine until you realise the benchmark itself is a static photograph. It assumes that if you hit enough targets on the checklist, the ecosystem will self-assemble. But recovery is not a points game. It's a trajectory, and trajectories bend, stall, or reverse. I have seen scores of 72 on a site where the only woody recruit was a single invasive acacia. The audit gave credit for “woody cover”. The field showed a monoculture creeping in.

Indicator Presence ≠ Functional Recovery

The gap between what a scorecard measures and what actually regenerates is where most teams lose years. A framework that awards points for “three reptile species recorded” doesn't capture whether those reptiles are breeding, feeding, or just passing through. A score for “leaf litter depth” doesn't tell you if the litter is decomposing into humus or just blowing away each dry season. Quick reality check — one team I worked with celebrated hitting 90% of their floristic targets only to discover that 60% of those plants were short-lived colonisers that would die before the next audit cycle. The framework rewarded early arrival, not persistence.

What usually breaks first is the assumption of linearity. Scorecard logic often embeds a hidden belief: that succession proceeds in deterministic steps, that each indicator tick moves the site closer to a predetermined endpoint. But recovery pathways are messy. They skip steps. They loop back. They depend on rainfall timing, disturbance memory, and soil microbial legacies that no spreadsheet captures. A single wet season can mask failure; a single drought can expose it. That's not a data problem — it's a design problem. The framework was built to compare, not to predict.

“We scored 82. Then a flood wiped the topsoil. The score meant nothing — the assumptions behind it had no tolerance for disturbance.”

— Mine closure lead, reflecting on a three-year audit cycle

Why Deterministic Succession Still Haunts the Audits

Most teams skip this: the classic Clementsian climax model — orderly, predictable, human-friendly — still underpins scores that pretend to be neutral. It shows up in how reference conditions are set (single static endpoint), in how recovery windows are defined (year five = year ten’s mini-me), and in how failure is diagnosed (missing species = missing recovery). The problem is not that succession theory is wrong — it's that scorecards reduce it to a caricature. They assume recovery follows a script. The field, as always, improvises.

The Three Hidden Succession Assumptions That Derail Most Frameworks

Assumption 1: Recovery follows a single, knowable trajectory

Most frameworks plot a tidy line from disturbance to reference condition. You calibrate against an endpoint, score progress, call it done. The catch is that real ecosystems don't move in straight lines—they branch, stall, and sometimes reverse. I have watched teams spend three seasons chasing a benchmark that the site had already abandoned. The R-squared looked great on paper. On the ground, the seam blew out because a flood bypassed the plot, depositing cobble where the model expected fine sediment. That single-trajectory assumption forces every anomaly into the error term. It doesn't ask why the plot diverted; it penalizes the diversion. The hidden cost is audit scores that tell you everything is fine until the next disturbance proves it isn't.

The temptation is obvious. A single trajectory makes scoring simple—you draw one line, compare, move on. But simplicity here is a trap. Biodiversity doesn't converge on a point; it converges on a corridor of possible states. Frameworks that ignore this produce scores that look precise but predict nothing.

Assumption 2: Disturbance resets the successional clock

This one is subtle. Many audit protocols treat fire, flood, or grazing as a clean slate—year zero, start counting again. Quick reality check: disturbance often accelerates certain functions while erasing others. A prescribed burn that clears understory might also trigger a pulse of nitrogen-fixing forbs that the pre-fire community never hosted. That's not a reset. It's a shift in functional composition that the framework misreads as failure. Most teams skip this distinction: they recalculate distance-to-reference from zero, ignoring that the site now contains species that won't appear in the reference at all. The score drops. The team scrambles to "fix" something that was actually a legitimate successional shortcut.

Wrong order. Not every pulse is a problem. But when your audit assumes disturbance wipes the tape clean, you spend money restoring conditions that the system had already moved past. That hurts budgets and, worse, it erases the natural signal of recovery.

Assumption 3: Reference ecosystems are static and accessible

Pick any reference site you used last year. Has it burned, dried out, or been invaded since then? Probably. Yet most frameworks treat the reference condition as a fixed photograph—a permanent benchmark that sits unchanged while the audited site churns through drought, flood, and species turnover. The result is a moving target that moves only on one side. The reference doesn't budge; the project takes all the variance. This creates a systematic bias: the older your reference data, the higher the chance your audit flags a "failure" that's actually just the reference drifting out of alignment with reality.

I have seen audits where the reference plot had not been visited in five years. The team dutifully scored their site against those stale metrics. What usually breaks first is the cover-abundance ratio—early-seral species dominate the project site because that's the stage it's in, but the reference data reflects a later phase that no longer exists anywhere in the landscape. The audit screams "divergence." The field team shrugs. Nobody recalibrates because recalibration is expensive and nobody budgeted for it.

The reference is not a museum specimen. It's a living plot that moves whether you track it or not.

— senior ecologist, after losing a season to stale baselines

So what do these three assumptions cost in practice? They produce audit results that are internally consistent but ecologically brittle. Teams chase static targets that nature ignores. Scores drift. Trust erodes. The fix is not to abandon reference conditions—it's to treat them as dynamic companions rather than fixed rulers. That's where the next section picks up: nested recovery windows that respect the branching, stalling, and shifting that real succession does.

Patterns That Hold: Nested Recovery Windows and Functional Turnover

Using functional groups instead of species lists to detect recovery

Species lists lie. Not maliciously—they just capture presence, not performance. I watched a restoration site in the Mediterranean score "excellent" on species richness while every oak sapling was chlorotic and dying in saline crusts. The list said recovery. The functional reality said collapse. Functional groups—rooting depth, nitrogen fixation strategy, leaf phenology—reveal what species lists hide: whether the *processes* that sustain recovery are actually running. A site can host ninety species and still lack a single deep-rooted perennial that stabilizes slopes. That's not recovery. That's a museum of relics.

The shift is painful for teams trained on checklists. "But we found the target species!" they say. Yes, but the target species are stunted, reproducing clonally, and surrounded by ruderal weeds. Functional turnover—the orderly replacement of fast-colonizing pioneers by slower, stress-tolerant species—tells you if succession is moving forward. Species lists only tell you who showed up to the party.

Nested recovery windows: soil, then plants, then animals

Recovery doesn't happen all at once. It arrives in nested windows, and the order matters more than the pace. Soil function must stabilize first—aggregate formation, organic matter cycling, microbial respiration. Then plants can recruit into a substrate that holds water. Then animals, particularly soil mesofauna and pollinators, return to a structured habitat. Flip that order and the whole system stutters. I once audited a framework that demanded bird species richness targets within three years. The soil was still compacted from heavy machinery. Birds visited, sure—they perched on dead shrubs. They didn't nest. No food base existed.

Wrong order. That hurts.

The nested window model forces auditors to ask: "Is the foundation built?" Not "Did we hit the headline number?" Most frameworks skip this entirely, scoring animals and plants on the same timeline. The catch is that soil recovery often takes 5–10 years longer than above-ground metrics suggest. When you compress windows, you either fudge the data or declare premature success. Both corrupt the audit.

If the soil is still silent—no fungal hyphae, no aggregate stability—don't count the birds yet. They're just tourists.

— field ecologist, tropical dry forest audit, 2023

When these patterns actually predict field outcomes

These patterns hold best in systems where disturbance is discrete—clearcut, burn, plow—and the surrounding matrix can supply propagules. Continuous stressors like chronic grazing or acid deposition break the windows. Soil recovers partially, plants stall, animals ghost. I have seen frameworks built on neat nested logic fail entirely on industrial brownfields where pH stays below 4 for decades. But within predictable disturbance regimes—temperate forests after selective logging, grasslands after pulse drought—functional group replacement tracks recovery with 12–18 month lag times you can actually budget for. Not theoretical. Auditable.

The trade-off: this approach demands longitudinal data, not single-season snapshots. One visit can't confirm functional turnover. You need two, three, four sampling rounds. That costs money. That costs patience. But the alternative—celebrating a species list while the ecosystem degrades—costs credibility. And credibility, once lost in an audit, rarely recovers.

Anti-Patterns That Keep Teams Stuck in Static Indices

The 'Green Blob' Trap: Scoring by Cover Area Alone

I once watched a team celebrate a 95% recovery score on a site that had lost every native forb and shrub. What remained? A dense monoculture of a single exotic grass that happened to turn green at the right season and cover the plot evenly. The framework rewarded area—green area, any area—and called it progress. That's the green blob trap: you measure the surface, not the structure. A canopy of invasive black wattle can score higher than a sparse but diverse scrubland because sheer volume dominates the metric. The catch is that field ecologists know this. They write notes in the margins. Yet the audit template keeps the blob alive because it's easy to automate and compare across sites. Painful.

What usually breaks first is the link between cover and function. A blob scores well in wet season surveys and catastrophically in dry season stress tests. Teams then scramble, patching weights onto the cover score—"let's multiply forb presence by 1.2"—without asking whether the index itself needs to die. Dying is better. Swap cover-only scores for a minimum patch richness threshold: if a 10×10 meter cell holds fewer than four growth forms, flag it regardless of total cover. That one rule caught more failure modes on my projects than any composite index ever did.

How Weighting Schemes Hide Missing Functional Guilds

Here is the counterintuitive part: weighting schemes often make the blindness worse. A team detects that pollinators are absent. Their response? Boost the weight on "pollinator-friendly species" from 5% to 15% of the total score. The raw data still shows zero pollinators, but the weighted score climbs because other categories—cover, litter depth, soil compaction—now contribute less to the denominator. The framework mathematically obscures the missing guild. One audit I reviewed had a "functional diversity" sub-index that gave positive points for bare ground when combined with high litter scores. Bare ground plus litter had no mechanism—it was a statistical artifact that happened to stabilize the total. The team had spent three months calibrating those weights. Wrong order.

How do you catch this? Simple: run the audit with all weights set to 1.0 first. Compare the raw outcome to the weighted one. If the weighted version hides a missing guild—if the pollinator score rises without any new pollinators—throw out the weighting scheme and rebuild from functional presence-absence. Not from scores. Presence-absence is brutal and honest. A 0 for "no nitrogen-fixing species" is clearer than a 64.7 that buries the zero inside thirteen other metrics.

Why Teams Revert to Simple Indices Even After Bad Predictions

The organizational pull toward simplicity is magnetic. After two failed pilot seasons using a detailed functional turnover index, one client I worked with reverted to a three-variable slope: cover, native richness, soil pH. Their stated reason: "The field crews could not remember the protocol." The real reason: the detailed index produced uncomfortable answers—sites they had certified as recovering were actually losing functional groups. So they simplified until the framework stopped disagreeing with them. That hurts.

A framework that always confirms your bias is not a framework at all. It's a mirror dressed as a scorecard.

— heard at a post-audit debrief, client site, 2023

The anti-pattern here is false pragmatism dressed as operational efficiency. Teams tell themselves they're reducing noise. In practice they're removing the signal that contradicted their investment decisions. The fix is cheap but uncomfortable: forbid reversion after a bad prediction unless the new index is tested on the same three failed sites first. No exceptions. If the simple index passes those failing sites—if it actually flags them—then fine. In my experience it never does. It just moves the target.

The Hidden Costs of Maintenance: Recalibrating Reference Conditions

The Undiscussed Tax of Keeping a Framework Alive

Most teams budget for building a biodiversity audit framework once. They price the field surveys, the indicator selection, the statistical handshake. What they don't price is the long-term cost of keeping that framework's succession logic honest. I have watched a well-funded project collapse not because the initial design was wrong, but because the reference conditions quietly rotted over three years of neglect. The gap feels invisible until suddenly your recovery windows no longer match what the ground actually does.

The math is brutal but rarely done. A reference condition—say, "mature grassland with 12 forb species per square meter"—drifts as the local climate shifts, as grazing pressure changes, as invasive seed banks build. Ignore it for one season, fine. Ignore it for three, and you're measuring recovery against a ghost. One client I worked with spent two years wondering why their framework flagged everything as "degraded." Reality check: the reference had not been recalibrated since the last drought cycle ended.

That hurts when you present findings to a regulator.

How References Drift: The Three-Year Gap That Breaks Your Logic

Consider a hypothetical watershed audit that set reference baselines in 2021. By 2024—without recalibration—the "reference" already describes a system that no longer exists. Early-succession species that depend on bare soil have vanished from the benchmark. Late-stage canopy cover assumptions sit 14% higher than actual regrowth rates. The framework's successional clock is now wrong at both ends. Most teams skip this entirely—they treat reference conditions as permanent, like a museum specimen, not a moving target.

The catch is that recalibration costs real fieldwork days. You can't update reference curves from a desk. Someone must go back to the reference sites, remeasure, and often discover that the old "undisturbed" plot now hosts a different functional guild entirely. I have seen teams budget zero for this step. Their frameworks become brittle within two budget cycles. That's not a failure of ecological understanding—it's a failure of operational planning.

'A reference condition is only a hypothesis until you test it against the current growing season.'

— field ecologist, after redoing 32 plots in one wet year

Cost of Field Validation: Reassessing Recovery Windows Across Seasons

Reassessing recovery windows means sampling the same sites in spring, summer, and autumn—minimum. One season's snapshot tells you nothing about whether the reference logic holds across germination windows, drought stress points, or pollinator arrival timing. The hidden tax here is compounded: each recalibration round requires travel, gear, analysts, and the coordination bandwidth that no project manager ever builds into their timeline.

Most teams skip this. Wrong order.

What usually breaks first is the functional turnover assumption—the idea that species x replaces species y at month 18. Without seasonal validation, you can't tell whether that turnover has shifted by six weeks or collapsed entirely. That's not an academic nuance. That's the difference between a framework that predicts recovery accurately and one that generates false positives every reporting period.

When Maintaining a Framework Costs More Than Building a New One

Here is the trade-off teams rarely face honestly: at some point, the annual recalibration cost exceeds the original build cost divided over the same period. If you spend $60,000 building the framework and then $18,000 every year keeping it current, by year five you have paid $150,000 total—and the framework still carries accumulated assumptions that may no longer hold. A fresh build, using current reference data and a simpler indicator set, might cost $45,000 and run cleaner for another three years.

That sounds fine until you realize the sunk-cost trap. Teams cling to old frameworks because the initial investment was painful. They retrofit, patch, and recalibrate until the logic resembles a scaffold built on shifting sand. I have seen organisations spend more on maintenance than on two complete rebuilds. The reason is never ecological—it's emotional and institutional. Breaking this cycle requires a hard rule: if recalibration costs exceed 30% of the original build in any single year, audit the framework's viability before you touch the reference conditions again.

One practical next action: build a cost log from day one. Track every hour, every sample, every recalibration trip. When the ledger hits that 30% threshold, convene a one-day review with an ecologist who was not part of the original design. Fresh eyes are cheaper than another year of patching a drifting reference.

When Not to Use This Approach: Novel Ecosystems and Managed Analogs

Novel ecosystems: no historical baseline exists

Some sites have crossed a threshold so far that the past is irrelevant. Think abandoned agricultural land colonized by species that never lived there before—or urban woodlands built on rubble heaps. No remnant patch, no paleoecological record, no old photograph tells you what “recovery” should look like. The successional framework I’ve described assumes a target state drawn from historical reference conditions. That assumption collapses here. You can't calibrate nested recovery windows against a baseline that never applied. What happens instead? Teams force-fit a trajectory—they pick a nearby forest and call it the endpoint. Wrong move. The site will never track that model. In novel ecosystems, the audit must shift from measuring return to documenting emergence: what assembles, how fast, and what functions appear.

— field ecologist, personal correspondence, 2024

— paraphrased from a site manager who spent three years chasing the wrong reference condition.

Managed analog sites: recovery trajectory is engineered, not natural

Mitigation banks, mine reclamation plots, and roadside revegetation zones look like restoration projects. They're not. The recovery arc is preordained by heavy machinery, seed mixes, irrigation schedules, and herbicide applications. Successional logic treats turnover as an emergent property—species replace each other through competition, dispersal, and disturbance. In managed analogs, turnover is scripted. I have watched crews plant late-successional tree species directly into bare subsoil, skipping the pioneer phase entirely. Does that violate successional rules? Yes. Does it work for the permit? Often yes. But the audit framework designed for natural recovery will flag these sites as failures because species replacement happens out of order. The hidden cost is false negative reports—teams spend months explaining why a site that looks good on paper fails the metric. You need a separate validation path for engineered trajectories: check structural outcomes (canopy cover, root depth) instead of compositional succession.

Rapid regime shifts where successional models fail entirely

Fire-adapted shrublands that convert to grassland after repeated burns. Coral reefs that flip to macroalgae beds. Mangroves drowned by sea-level rise faster than sediment can accumulate. In these cases, succession doesn't happen—the ecosystem reorganizes so quickly that the concept of a recovery window loses meaning. The catch is that many audit frameworks still apply successional logic out of habit. “We’ll just extend the timeline,” teams say. That rarely works. Regime shifts produce alternative stable states; the system won't revert even with aggressive intervention. What usually breaks first is the turnover metric: functional groups disappear entirely instead of replacing each other. If your data shows a flat line for three consecutive audits and no pioneer species appear, successional assumptions are probably the wrong lens. Switch to regime-shift indicators—threshold proximity, feedback strength, propagule pressure—or stop auditing recovery entirely and start tracking adaptation.

That sounds harsh. But forcing successional models onto non-successional systems wastes field days, inflates false alarms, and erodes stakeholder trust. The trade-off is clear: you lose comparability with other sites, but you gain honesty about what the data actually says.

Open Questions and Practical FAQ for Audit Teams

How to choose between frameworks without field testing them all?

You can't. That's the honest answer most vendors skip. Every biodiversity audit framework arrives pre-loaded with successional assumptions—about how fast a logged forest recovers, which species mark “done,” what soil chemistry should read at year five. Field testing every candidate on your site would take longer than the audit itself. So you compress: pick two or three that explicitly name their reference model (old-growth? pre-colonial? functional target?) and run a quick desktop stress test. Grab plot photos from a single 50×50 meter quadrat you know well. Plot the framework’s predicted recovery trajectory against what that quadrat actually did over, say, three sampling seasons. The one that diverges least in the first five years usually wins—not because it's “correct,” but because its hidden assumptions align with your local disturbance regime.

Wrong order costs months. I have seen teams choose a framework purely on index breadth—42 metrics!—then discover it assumes closed-canopy recovery within eight years on a site that naturally stays open for twenty. The audit looked great on paper; the field rejected it.

What to do if your framework’s recovery score says one thing but field data says another?

Pause the score. Don't force the frame onto the field—that's how you get a certified “recovering” site that's actually a novel thicket with zero functional turnover. Start by listing where the divergence sits. Is it in the structural layer (canopy height lagging) or the compositional layer (wrong species arriving)? Most audit frameworks collapse these into a single weighted score; that aggregate hides the friction. The catch is that once you separate them, you often discover the field is recovering on a different pathway, not failing to recover at all. Quick reality check—plot your field metrics against the framework’s published recovery curve for that biome. If the shape is similar but offset by a few years, you can shift the reference window without abandoning comparability. If the shapes don't match at all—field data shows a plateau while the framework expects a steady climb—then the successional model inside the scorecard is wrong for your system.

That hurts. But pretending otherwise inflates every audit downstream.

Can you modify a framework mid-audit without breaking comparability?

Yes, but only at specific seams. Modifying the calculation engine—changing weights, swapping indicator thresholds—mid-audit poisons any before-after comparison. What works is adjusting reference conditions as long as you flag it explicitly. Say your framework expects 80% native woody cover by year ten, but your site is a pyric system where native grasses dominate that decade. You can recalibrate that benchmark, provided you document: (a) the original expectation, (b) the field evidence for divergence, and (c) the new reference with a justification. The comparability problem is not the change—it's the silence around the change. Most audit teams skip the documentation step. Then next year’s team sees two different scores and assumes the site degraded.

“We changed the tree-cover threshold from 70% to 45% because the reference plot was a plantation. No one logged the reason. Now the archive looks like a collapse.”

— Field coordinator, tropical dry-forest audit, 2023

How to detect hidden assumptions in scorecards you didn’t build?

Run the framework against a known failure. Take a site you already know is not recovering—a repeated-burn edge, a compacted skid trail—and see what the scorecard says. If it still gives that site a passing recovery score, you have found an assumption blind spot. Common culprits: the framework assumes linear recovery (it never plateaus), it weights structural metrics over functional ones (tall weeds = forest), or it uses a single reference state when your ecosystem has multiple successional end-points. Another tell—check the metric list for absence. No soil indicator? The framework assumes recovery happens above-ground first. No pollination or seed-dispersal proxy? It assumes dispersal is not the bottleneck. Most teams never stress-test a framework this way because they're anxious to start collecting data. That anxiety is exactly what keeps static indices alive. Break the scorecard before the scorecard breaks your site.

One concrete next step: take your three candidate frameworks and run each against a single “failed” quadrat. The one that correctly flags it as failing—while still passing a known recovering site—deserves your trust. The others get parked until their assumptions are revised.

Next Steps: Build Your Own Successional Stress Test

Template for a 3-step stress test: assumptions, counterfactuals, field check

Pull your current framework—any framework—and set a timer. Step one: list every assumption your scoring logic makes about how species return after disturbance. Not the generic ones like 'ecological succession occurs.' The specific ones: that pioneer species will be present by month six, that soil seed banks remained intact, that dispersal corridors weren't severed. I have watched teams discover their entire audit rested on a single unverified assumption—that mycorrhizal networks survived a fire. They hadn't checked. The assumption was just inherited from the previous year's template. Step two: build a counterfactual. What would your score look like if that one assumption were false? If pioneers show up late? If the seed bank was sterilized? Run the numbers. The gap between your original score and the counterfactual score is your framework's fragility index. Most teams skip this—they fear the delta. Step three: go to the field. Not next quarter. Tomorrow. Pick three plots where your framework predicted high recovery and three where it predicted failure. Dig. Count functional groups, not just species presence. That hurts when the data contradicts your dashboard. But that's the point.

One afternoon. Three steps. You will find the seam where your logic blows out.

'We ran our stress test on a post-mining site. The framework gave it an 82. The counterfactual gave it a 34. The field data gave it a 29. We had been tracking the wrong recovery signal for two years.'

— Restoration ecologist, private conversation after an audit workshop, 2024

Experiment: compare framework score vs. functional group recovery at 1 year

Here is a direct challenge for your next monitoring window. Run your standard audit on a site twelve months post-disturbance. Then, separately, measure functional group recovery: nitrogen fixers, deep-rooted perennials, litter decomposers, pollinator host plants. Not abundance—just presence or meaningful functional activity. The catch is that most frameworks score structural recovery—canopy cover, species richness—while ignoring whether the system actually works. I have seen a site score 'high recovery' because it had thirty plant species, but fifteen of them were non-native annuals that wouldn't hold soil through a dry season. The audit said success. The field said failure. Run the comparison. If your score and the functional group map align within 15 percent, your assumptions are holding. If they diverge—and they will—you have found exactly which assumptions need recalibration. Quick reality check: don't average the scores. Plot them. One point per plot. The spread tells you more than the mean ever will.

That divergence is not noise. It's your next audit improvement.

Share findings to improve collective audit logic

Most teams treat their framework failures as proprietary secrets. Wrong order. The frameworks that hold up across ecosystems are the ones that have been stress-tested by fifty different teams in fifty different disturbance regimes. Share your counterfactual gap. Share the functional group that broke your score. Share the assumption you had to kill. I would rather read your failure report than your success slide deck—the failure report shows me exactly where my own framework is vulnerable. Start a shared log: one spreadsheet, three columns (assumption, site condition where it failed, corrected logic). No branding. No pride. Just the raw data of what succession actually does versus what we assumed it would do. Teams that contribute get access to the growing pool of counterfactual tests. That is the exchange. Your honesty buys you everyone else's mistakes. The goal is not a perfect framework—there is no such thing. The goal is a framework that fails transparently, early, and in ways that teach the next team something useful.

Build your test. Share your seam. Repeat next season.