Work Output - Common Misinterpretations
Objective. Carry a quick-reference catalogue of the ways Work Output numbers get misread - grouped by theme, each with the wrong reading, why it is wrong and the correct reading - so a rep or CSM can catch the misread in the room before it becomes a customer commitment.
Work Output is powerful precisely because it reduces delivery to a few sharp numbers, and sharp numbers are easy to misuse. This reference collects the recurring misreads drawn from the app's workflow interpretation checks and the grounding guardrails, organised by theme: counting and coverage, comparison and normalisation, standards and attainment, quality and cost, and projection versus value. Use it as a catch-sheet, not a lecture.
See it in the demo
Counting and coverage misreads
Misread: 'The transaction total is our output.' Why wrong: a count is only as complete as the sources feeding it; manual or off-system work is invisible, so the total can be understated or inconsistent. Correct: a transaction total is completed-work volume in one mix, from the connected sources - trustworthy only after counted-work coverage is confirmed.
Misread: 'A high count means a strong team.' Why wrong: volume ignores complexity and quality - a high count of simple or reworked units is not strong delivery. Correct: read the count with its complexity mix and the quality filter before drawing any conclusion.
Misread: 'The number is the number - it is measured, so it is objective.' Why wrong: the number is objective about what was counted, not about reality, if coverage is incomplete or the completion definition is loose. Correct: treat every figure as inheriting the trust of its completion definition and coverage.
Comparison and normalisation misreads
Misread: 'Team A does half the volume of Team B, so A is underperforming.' Why wrong: raw counts across different work type, complexity, period and tenure are not comparable - this is the single most common error in the app. Correct: normalise first (work type, complexity, period, tenure) and apply the quality filter; only then compare, and compare clean output per paid hour, not raw counts.
Misread: 'This team is in the bottom-left quadrant, so it is our worst team.' Why wrong: a quadrant position depends on the axes and thresholds chosen, both configurable, and a low corner often reflects work mix, complexity or an uncounted-work gap. Correct: read a quadrant as a configurable lens that flags where to look next, never as a verdict on people.
Misread: 'Benchmark says we are below peer, so we are behind.' Why wrong: a benchmark is only valid if the compared populations and definitions match. Correct: confirm like-for-like definitions before a benchmark becomes a conclusion.
Standards and attainment misreads
Misread: 'The team hit 120% attainment, so they are outperforming.' Why wrong: attainment is meaningless if the standard is arbitrary; 120% against a too-soft standard means the standard is wrong, not that the team is exceptional. Correct: read attainment and the standard together, and never present attainment without being able to defend the standard behind it.
Misread: 'Low attainment proves the team is underdelivering.' Why wrong: a standard that does not match complexity condemns a team doing hard work; the shortfall may be in the standard, the mix or coverage. Correct: check the standard is agreed and complexity-matched, then rule out mix and coverage before attributing a shortfall to the team.
Misread: 'Standards are set, so attainment is reliable from now on.' Why wrong: standards drift as the operation changes and are assumptions until re-agreed. Correct: treat standards as governed, periodically re-validated agreements.
Quality, cost, and projection misreads
Misread: 'Output per hour is up, so the team improved.' Why wrong: output per hour without the quality filter flatters a team that is fast and wrong. Correct: read clean output per paid hour - volume that met quality - or the gain may be rework in disguise.
Misread: 'Cost per unit is our unit cost.' Why wrong: cost spread over raw output counts rework and defects as good work, understating true cost. Correct: use cost per clean unit as the honest figure.
Misread (the signature one): 'Add the recoverable value and the capacity value for our total savings.' Why wrong: recoverable value is an opportunity, capacity value is freed time, billed value is invoiced and collected value is cash in - four different things that must never be summed or swapped. Correct: name the value type every time and keep them in separate columns; only billed and collected are realised money.
Misread: 'The AI-impact ROI figure is our expected savings - put it on the proposal.' Why wrong: a projection or AI-impact estimate is opportunity under assumptions, not a realised result or booked cash; change an assumption and it moves. Correct: label it a projection, rebuild it on the customer's baseline, and claim value only from a measured before-and-after elsewhere.
Misread: 'Root cause / AI impact says effort caused the change.' Why wrong: these surface candidates and correlations, not proven causality. Correct: treat them as leads to validate, ruling out mix, complexity, backlog and coverage first.
How to use this catalogue in the room
When a customer or colleague states a conclusion from a Work Output screen, run three fast checks before you let it stand: what is the denominator and is it clean? (coverage, quality filter, paid hours); is this comparable / is the standard agreed? (normalisation, standard credibility and complexity match); and is this opportunity or realised, and which value type? (projection versus result, and the four value types kept apart). If any check fails, reframe to the correct reading in the section above before the statement becomes a commitment. The cost of catching a misread in the room is a sentence; the cost of letting it reach a slide is a credibility problem later.
Interpretation check
Almost every Work Output misread fails one of three checks - a dirty or unconfirmed denominator, an unnormalised comparison or undefended standard, or opportunity presented as realised with value types blended; run those three checks before any conclusion stands.
In the field
In the KnackGlobal displacement case the dangerous misread is quoting raw 'claims worked' as output, or lifting a projected cost-per-clean-claim saving as if it were booked - the correct readings (clean, complete, complexity-aware claims counts, and recoverable opportunity kept separate from collected cash) are exactly what make the displacement argument survive scrutiny.
Practical task
Pick three screens in the demo, state one plausible conclusion a customer might voice on each, classify the misread using the five themes here, and write the one-sentence correct reading you would offer in the room. At least one of your three must involve either the value-type separation or the projection-versus-result line.
Self-check
A deck sums recoverable value and capacity value into one 'total savings' figure. Name the misread and the correct reading.
The misread is blending value types: recoverable is an opportunity and capacity value is freed time, neither is cash, and they cannot be summed. The correct reading names each value type separately and treats only billed and collected as realised money - kept in separate columns.
'Output per hour rose 15%, the team improved.' What is the hidden risk and the fix?
The risk is reading raw output per hour without the quality filter, so a rise can be fast-but-wrong rework. The fix is to read clean output per paid hour - volume that met quality - before claiming improvement.
A manager ranks five teams by last month's transactions. Which misread is this and what must happen first?
The comparison trap - raw counts across different work type, complexity, period and tenure are not comparable. Normalise on those dimensions and apply the quality filter first, then compare clean output per paid hour rather than raw counts.
A team shows 118% attainment and the sponsor wants to praise them. What do you check?
Whether the standard is agreed, credible and complexity-matched. 118% against a too-soft standard means the standard is wrong, not that the team is exceptional - attainment and the standard are read together and the standard must be defensible.
Someone wants to put the AI-impact ROI number on a proposal as expected savings. Why is that a misread?
An AI-impact estimate is opportunity under assumptions, not a realised result or booked cash; change an assumption and the figure moves. It must be labelled a projection, rebuilt on the customer's baseline, and claimed as value only from a measured before-and-after.
Verified vs demo. The misreads and correct readings are grounded in the app's workflow interpretation checks on the seeded demo; whether a given customer's figure is actually affected depends on their confirmed completion definitions, coverage, standards, quality signal and cost basis, which are confirm-with-product.