Specimen cohort report

What the institution receives.

The individual report shows what a company gets. This is the other artefact: the same results, aggregated across a cohort, at intake and again after ninety days.

Illustrative. Built from a synthetic cohort, not observed data. The structure, the arithmetic and the limits are exactly as a real report would carry them. See the individual report this aggregates from.

CohortIllustrative programme
Invited48
Baseline / endline41 / 29
Coverage29 of 41 · 71%
WindowFeb to May 2026
Method versionv1.4

What changed

Mean capability44 → 53+9
Mean momentum39 → 51+12
Companies that moved 5 or more19of 29 rescored
Companies that did not move10of 29 rescored

Movement is the difference between two measurements taken on the same method version. It is not attributed to any intervention. See the limits below.


Where the cohort is concentrated

Not 41 problems. One problem, 23 times.

Mean score per cell across the 41 companies measured at baseline, on a five point scale. Stronger blue is weaker. Every cell carries its number, so the colour is never the only thing saying it.

Product
People
Process
Repeatability
Demand
2.6
2.3
2.2
1.8
Sales Process
2.8
2.5
2.3
1.8
Closing
2.5
2.3
1.9
1.6
Expand
2.5
2.3
2.0
1.8

Repeatability is the weakest column across every stage, and Closing is the weakest row. Twenty three of the forty one companies carry Closing and Repeatability as their primary reading. That is a commissioning finding, not forty one separate coaching problems.

Cell mean across all sixteen cells is 2.20, which is the 44 shown above. The twelve companies that did not return for an endline had a baseline mean of 43, within one point of the twenty nine that did, so the matched set is not visibly different from the set that dropped out. That check is printed because a reader should not have to ask for it.


What moved, cell by cell

Nine cells moved. Three went backwards.

Change per cell between intake and rescore, across the 29 companies measured twice, on the same five point scale. Every value carries its sign, so up and down are readable without relying on colour.

Product
People
Process
Repeatability
Demand
+0.5
+0.2
+0.6
+1.1
Sales Process
+0.3
+0.1
+0.7
+1.3
Closing
+0.6
−0.5
+1.0
+1.5
Expand
−0.4
−0.6
+0.6
+0.2
Improved, larger Improved Not distinguishable from noise Declined

Process and Repeatability rose in every stage. People fell in the two later stages. That is a specific finding, and it is not the one a programme would choose. The support this cohort took up was process shaped, and process work did not hold the People reading in Closing and Expand.

What counts as movement here. At 29 matched pairs, a change smaller than 0.4 on a cell, or 7 points on the 100 point scale, cannot be told apart from measurement noise at conventional confidence. Four cells fall below that threshold and are marked as such rather than counted as movement. The calculation assumes a test retest correlation of 0.7 and the spread observed in this cohort. Both are stated so the arithmetic can be checked.

The four cells that did not move: Demand and People, Sales Process and Product, Sales Process and People, Expand and Repeatability.

StageRoot causeIntakeEndlineChangeReading
DemandProduct2.63.1+0.5Improved
DemandPeople2.32.5+0.2Not distinguishable
DemandProcess2.22.8+0.6Improved
DemandRepeatability1.82.9+1.1Improved, larger
Sales ProcessProduct2.83.1+0.3Not distinguishable
Sales ProcessPeople2.52.6+0.1Not distinguishable
Sales ProcessProcess2.33.0+0.7Improved
Sales ProcessRepeatability1.83.1+1.3Improved, larger
ClosingProduct2.53.1+0.6Improved
ClosingPeople2.31.8−0.5Declined
ClosingProcess1.92.9+1.0Improved, larger
ClosingRepeatability1.63.1+1.5Improved, larger
ExpandProduct2.52.1−0.4Declined
ExpandPeople2.31.7−0.6Declined
ExpandProcess2.02.6+0.6Improved
ExpandRepeatability1.82.0+0.2Not distinguishable

The table is on the page rather than behind a control, so the grid can be checked without a login and the figures survive if the grid fails to render.


The reporting line

The evidence is the same. The line you report against is not.

Every institution that commissions this reports to somebody else, and each reports against a different published line. Nothing above this point changes when you switch. The same 29 matched companies, the same grid, the same limits. Only the mapping changes, and every mapping is published here so that any of them can be disputed.

Integrated Settlement Outcomes Framework · Outcome 1 · indicator 1.3 · supported businesses demonstrating improved practice

Of 29 businesses measured twice in this period, 19 demonstrated improved practice on at least one dimension read by this instrument, and 4 on all three.

Dimension, as publishedRead hereImprovedWhat it is read from
Engaged in new marketsYes13 of 29The Demand row, Product and People cells
Adopting new to firm technologies or processesYes16 of 29The Process and Repeatability columns, all four stages
ScalingIn part7 of 29The Expand row. Read as the capability to grow existing accounts, not as headcount or turnover, neither of which this instrument measures
With new to market products or servicesNoNot readProduct innovation sits outside what this instrument asks about
Accessing better finance and fundingNoNot readNot asked, and deliberately so. This is not a credit or eligibility signal

Nineteen companies improved on at least one dimension: 4 on all three, 9 on two, 6 on one. Ten showed no improvement on any dimension read here. Counts are of companies, not of improvements, so they do not sum to the column above.

The arithmetic a commissioner can then do. This cohort returned 19 companies demonstrating improved practice from 48 invited. A target of 500 over a settlement period would need roughly 26 cohorts run at this size and this rate. That number may be uncomfortable, and it is the point: it can only be calculated once improvement is measured rather than counted as activity. A count of businesses supported cannot produce it.

The indicator, its five dimensions and its wording are taken from the published framework. The mapping is ours, is stated so it can be disputed, and is not endorsed by any authority.


The aggregation, shown

Every figure above comes from here.

Twelve of forty one shown. Each row is one company's own result, the same one it received. Nothing in this report is calculated any other way.

RefIntakeEndlineMoveStageRoot causeSupport indicated
C-0413854+16ClosingRepeatabilityDeal control clinic
C-0174461+17DemandRepeatabilityProposition workshop
C-0295152+1ExpandProcessNot taken up
C-0083349+16Sales ProcessProcessDeal control clinic
C-03547470ClosingRepeatabilityNot taken up
C-0025668+12DemandProductSelf directed
C-0232941+12ClosingRepeatabilityDeal control clinic
C-0116264+2ExpandPeopleSelf directed
C-0464055+15DemandRepeatabilityProposition workshop
C-00435Sales ProcessProcessNo endline
C-0194863+15ClosingRepeatabilityDeal control clinic
C-0305358+5DemandProcessSelf directed

Twenty nine more rows in a real report. Company references are shown to the institution that invited the cohort. Companies outside the cohort never appear.


What was indicated, and what happened

Category of support indicatedCompaniesTook it upWhat that suggests
Deal control and qualification1811The largest single need, and provision exists
Proposition and segment definition127Well served by the existing workshop
Retention and expansion72Thin provision. Worth commissioning against
No provider needed44Acted on their own first action

The diagnosis names the category. Which provider appears within it is the institution's decision, from provision it already funds. Take-up is recorded; it is never required.


Limits, printed on the report

  • The data is self reported by each company. Every figure above inherits that
  • Twenty nine endline measurements is a small sample. No comparison between providers, programmes or periods is drawn at this size, and none should be
  • Movement is not attributed. This reports what preceded what. Companies also hired, lost customers and changed markets during the window
  • Twelve companies did not complete an endline. They are excluded from movement figures, not counted as no change
  • No peer benchmark is shown. The published gate is roughly 100 comparable observations within one business model at one maturity level. This cohort is well below it
  • This is not a credit, risk or eligibility assessment, and must not inform funding decisions, grant eligibility or subsidy control

Why this is the artefact that matters

A company acts on its own report. An institution cannot: one company's diagnosis says nothing about where provision is short, which cohort is weakest, or whether last year's spend changed anything.

This report is the same results read a level up. One instrument, two audiences, and no separate management view that can quietly disagree with what the companies were told.