Work Output - Concepts and Key Terms
Objective. Hold the mental model Work Output is built on - the time-to-outcome chain from hours to completed work to attainment, quality and cost - and the 8-10 terms you must use precisely, so that output is never confused with activity and the four value types are never blended.
Before any screen makes sense, you need the model. Work Output is built on one idea: busyness is not delivery, and the only honest way to know whether paid teams are producing the expected work is to count completed work, give the count meaning with complexity and an agreed standard, filter it for quality, and price it. This module defines the terms that carry that model and shows how they chain - and names the one concept people most often get wrong.
See it in the demo
The core concept: output is not activity
Activity and attendance tell you people were present and busy. Work Output tells you whether the expected work got done, to standard, at acceptable quality and cost. The whole application exists to close the gap between those two statements, because a team can be fully utilised and fully present while delivery slips.
The spine of the model is the time-to-outcome chain: paid hours in, work units completed, attainment against an expected standard, and then the quality and cost of that output. Each link adds meaning the previous one lacked - hours alone say nothing about output, output alone says nothing about whether it was enough, attainment alone says nothing about whether it was clean, and clean output alone says nothing about what it cost. Reading any link without the others is where misreads begin.
Key terms, defined plainly
- Completed work (transaction) - one unit of work the system counted as done, at a grain the customer recognises. The raw material of everything downstream.
- Completion definition - the agreed rule for what counts as a completed unit in a given process. Loose here, loose everywhere.
- Counted-work coverage - the share of real completed work that is actually counted. Gaps (manual or off-system work) distort every derived figure.
- Complexity - how hard a unit of work is; the weighting that stops ten simple units being read as ten complex ones.
- Standard (expected value) - the output a process at a given complexity is expected to produce. An assumption until agreed and credible.
- Attainment - actual output against the standard. Meaningful only when the standard is defensible and complexity-matched.
- Quality / clean output - completed work that met the required quality; 'clean' output is volume with the quality filter on.
- Clean output per paid hour - the headline operating metric: quality-passed output over the paid hours that produced it.
- Cost per clean unit - the cost of good output, spreading cost over clean units rather than raw ones.
- Value types - recoverable (an identified opportunity), billed (invoiced), collected (cash in) and capacity (freed time): four distinct things, never summed or swapped.
How the concepts chain together
The terms are not a flat glossary; they build on each other in order, and skipping a link breaks the meaning:
- Completed work is only trustworthy once the completion definition is agreed and counted-work coverage is confirmed.
- A trustworthy count becomes meaningful only with complexity attached, so units of different difficulty are not treated as equal.
- Attainment needs a credible, complexity-matched standard to measure the count against - otherwise it is arithmetic without meaning.
- Quality turns raw attainment into clean output, so speed-without-quality cannot flatter the picture.
- Clean output per paid hour brings in the paid-hours denominator (from Time and Activity) to express productivity honestly.
- Cost per clean unit prices that clean output for the business case.
- Value types then sit on top - and this is where the discipline bites: a recoverable or capacity figure is an opportunity, only billed and collected are realised money, and a projection (for example an AI-impact estimate) is never a realised result.
Each step depends on the one below it being honest. That is why a CSM validates from the bottom of the chain up.
The concept people get wrong most often
The single most common conceptual error is treating a Work Output number as a performance verdict rather than a reading that depends on its denominator, standard and value type. It shows up three ways:
- Reading a raw count or output-per-hour as 'this team is good/bad' without normalisation, the quality filter or coverage confirmation.
- Reading attainment as performance when the standard behind it was never agreed or does not match complexity.
- Reading a recoverable or capacity figure, or a projection, as realised savings - blending opportunity with cash.
The corrective habit is to say the denominator, the standard and the value type out loud before stating any conclusion. If you cannot name all three and defend them, you do not yet have a reading - you have a number. Hold this and most of the app's misinterpretations never happen.
Interpretation check
The time-to-outcome chain only holds if each link is honest: a count needs coverage and a completion definition, attainment needs an agreed complexity-matched standard, output per hour needs the quality filter and a trustworthy paid-hours denominator, and value needs its type named - with opportunity and projection never read as realised cash.
In the field
The KnackGlobal case is the model in one line: raw 'claims worked' (activity-flavoured output) is replaced by clean claims per paid hour and cost per clean claim (the honest links of the chain), with recoverable opportunity kept as opportunity - which is why the output model, not an activity count, is what displaces an incumbent.
Practical task
Draw the time-to-outcome chain from memory as a labelled line from paid hours to cost per clean unit, annotate each link with the one term it introduces and the one thing that must be true for that link to be honest, and add a side-note listing the four value types with 'opportunity' or 'realised' against each.
Self-check
Explain in one sentence why 'output is not activity' is the founding idea of the app.
Because a team can be fully present and fully utilised while delivery slips, so busyness cannot answer whether the expected work got done to standard, quality and cost - which is exactly what Work Output measures.
Put the core terms in dependency order from most foundational to most derived.
Completion definition and counted-work coverage, then completed work, then complexity, then standard and attainment, then quality and clean output, then clean output per paid hour, then cost per clean unit, then value types and projections on top. Each depends on the ones below being honest.
What is the difference between clean output per paid hour and output per hour, and why does it matter?
Clean output per paid hour applies the quality filter so only quality-passed work counts, and uses the paid-hours denominator. Plain output per hour can be inflated by fast-but-wrong rework, flattering a team that is producing quickly and badly.
Name the four value types and say which are realised money.
Recoverable (an opportunity), billed (invoiced), collected (cash in) and capacity (freed time). Only billed and collected are realised money; recoverable and capacity are opportunities and must never be summed with cash or with each other.
What is the single conceptual error people make most, and the habit that prevents it?
Treating a number as a verdict rather than a reading that depends on its denominator, standard and value type. The habit is to say the denominator, the standard and the value type out loud and defend all three before stating any conclusion.
Verified vs demo. The chain, terms and relationships are the app's model as shown on the seeded demo; a given customer's completion definitions, complexity model, standards, quality signal, paid-hours source and cost basis are confirm-with-product and determine whether their actual numbers behave as the model predicts.