Value Leakage misinterpretations - the metric misreads to catch before they reach finance
Objective. Provide a quick-reference catalogue of the common ways Value Leakage numbers are misread, grouped by theme. For each: the wrong reading, why it is wrong, and the correct reading. Built from the app's workflow interpretation checks and the signature guardrails - capacity-vs-cash, confidence discounting, overlap, detected-vs-recovered, and projection-vs-result.
Value Leakage exists to produce honest, defensible numbers, which means its misreads are high-stakes: a blended total or a detected-as-recovered claim is exactly the inflation that destroys finance's trust in the whole programme. This is the catalogue of those misreads, grouped so a CSM or Sales person can catch them in a review or a demo before the number leaves the room.
Capacity versus cash (the signature misread)
Misread A - blending the lines. Wrong reading: 'Recovery shows capacity freed and cash recovered, so our total recovery is the two added together.' Why it is wrong: freed capacity is hours or FTE, not money; it only becomes cash if the customer redeploys or removes that capacity, which is their decision. Summing them inflates the result and destroys the credibility the confidence-discounted model was built to protect. Correct reading: read them as two separate lines that are never summed - 'this much capacity we could free, this much cash we actually recovered.'
Misread B - capacity equals cash saved. Wrong reading: 'We freed 200 hours, so we saved the cash value of 200 hours.' Why it is wrong: the cash is only realised if the customer actually removes or redeploys the freed hours; until then it is capacity opportunity, not a saving. Correct reading: state freed capacity as capacity, and only count cash when the customer has acted and it is Verified with evidence.
Detected and Open versus Verified (state misreads)
Misread C - detected means proven. Wrong reading: 'The system detected this leak automatically, so it is proven and recoverable.' Why it is wrong: detection produces a candidate surfaced from observed work; it does not establish the rate, the recoverable share, overlap or that any cash was recovered. Correct reading: a detected finding is a candidate to validate and size, not a conclusion.
Misread D - Open value is money won. Wrong reading: 'This finding is at Open with a large value, so that is money we have recovered.' Why it is wrong: Open means sized but not yet acted on - an opportunity, not a result. Correct reading: only a finding at Verified, with evidence, counts as recovered, and even then capacity and cash stay on separate lines.
Misread E - an assigned owner is progress. Wrong reading: 'This leak has an owner now, so we can report it as progress.' Why it is wrong: attribution names who can act; it is the start of accountability, not evidence of recovery. Correct reading: progress and recovered value live on the Recovery screen with status and evidence, not on Attribution.
Confidence, overlap and the ranking (aggregation misreads)
Misread F - the confidence discount is pessimism. Wrong reading: 'Confidence marks the numbers down, so it is underselling the real opportunity.' Why it is wrong: confidence is what makes the number defensible - it discounts uncertain findings so a directional guess is not counted as proven. Correct reading: the discount is the credibility; a number your CFO can argue with beats a fantasy they dismiss.
Misread G - a smaller finding outranking a larger one is a bug. Wrong reading: 'This smaller finding ranks above a bigger one, so the ranking is broken.' Why it is wrong: ranking is by confidence-discounted value, so a well-evidenced smaller finding can beat a larger but uncertain one - the model working as intended. Correct reading: rank reflects defensible value, not raw size.
Misread H - summing the top findings. Wrong reading: 'Add the top three leaks for a headline savings number.' Why it is wrong: the top findings can mix capacity and cash and may overlap on the same underlying hours, so summing blends value types and risks double-counting. Correct reading: rank is a validation order, not a banked total; check overlap and keep capacity and cash separate before any aggregate.
Scope, sample and projection (foundation misreads)
Misread I - theoretical maximum as recoverable. Wrong reading: 'The sized leak is the value we will recover.' Why it is wrong: the sized leak is a theoretical figure before the recoverable-share assumption; only a realistic share is actually recoverable, and only Verified evidence confirms what was recovered. Correct reading: distinguish theoretical leak, recoverable value and verified recovery.
Misread J - a seeded demo magnitude quoted as a customer number. Wrong reading: 'The demo shows Capacity around 78 lakh and Process around 1.12 crore, so that is the opportunity.' Why it is wrong: those are illustrative seeded magnitudes that prove the model sizes and ranks leaks, not a figure for any real customer. Correct reading: use the magnitudes to show the model works, and build the customer's own number from their measured quantity and finance-agreed rates.
Misread K - the leakage number is independent of the upstream signal. Wrong reading: 'The leakage figure stands on its own.' Why it is wrong: each finding inherits the definitions and coverage of the upstream app's signal (available time, friction, rework, outcomes); a weak upstream definition makes a weak leakage figure. Correct reading: validate the upstream signal before trusting the leakage number built on it.
Interpretation check
The headline misreads are blending capacity with cash, reading detected or Open as recovered, summing overlapping or mixed-type findings, and quoting a theoretical maximum or a seeded demo magnitude as a customer result. The correct readings keep capacity and cash separate, count only Verified-with-evidence, check overlap, and build a customer number from measured quantity and finance-agreed rates.
In the field
KnackGlobal (RCM, ~6,000 seats): the blend-and-overclaim misread is the exact failure the account cannot afford against an incumbent attack - a single inflated 'total savings' that finance writes down undoes the whole ROI case, which is why every one of these misreads is caught before the number reaches the customer's finance.
Practical task
Take three figures from the demo (one from a category tile, one from Top value leaks, one from Recovery) and for each write the most likely misread a customer would make and the one-sentence correct reading you would give. At least one must address the capacity-vs-cash separation and one must address detected/Open versus Verified.
Self-check
A customer reports 'total recovery' as capacity freed plus cash recovered. Name the misread and the correct reading.
The capacity-vs-cash blend. Freed capacity is hours, not money, and only becomes cash if the customer redeploys or removes it. Correct reading: two separate lines that are never summed - capacity we could free, cash we actually recovered.
Someone says the confidence discount is making the app undersell the opportunity. Why is that a misread?
Confidence is what makes the number defensible - it discounts uncertain findings so a directional guess is not counted as proven. The discount is the credibility; a number finance can argue with beats one they dismiss as a fantasy.
Why can you not sum the top three findings into a headline number?
They can mix capacity and cash and may overlap on the same underlying hours, so summing blends value types and risks double-counting. Rank is a validation order, not a banked total - check overlap and keep capacity and cash separate first.
A detected finding at Open has a large value. Two misreads are lurking here - name them.
First, that detected means proven (it is a candidate to validate, not a conclusion). Second, that an Open value is money won (Open is sized-but-not-acted-on; only Verified-with-evidence counts). Both inflate the result.
The demo shows Process around 1.12 crore. A rep quotes this to a prospect as their opportunity. What is wrong?
It is an illustrative seeded magnitude that proves the model sizes and ranks leaks, not a figure for any real customer. Build the customer's own number from their measured quantity and finance-agreed rates; use the demo figure only to show the model works.
Verified vs demo. The screens, states and category magnitudes that these misreads reference are demo-shown on seeded data (for example Capacity ~78 lakh, Process ~1.12 crore). The misreads themselves are a reasoning guide drawn from the app's interpretation checks and guardrails; the correct customer numbers, rates and recoverable-share assumptions are customer-specific - confirm with product and the customer's finance.