River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Training Effectiveness Measurement Template

Three documents and three sheets that read a business metric two ways and report the gap between them as a range, not one number.

Free download  ·  No account needed

Business Metric Correlation

Programme:

The metric read two ways: a same-period comparison and a whole-population before-and-after. The gap between them is reported as a range, not smoothed into one number.

PeriodWaveStatusOpportunitiesWonRate
P-1
P-2
P-3
P-4
P-5

Attribution

ReadingResult
Naive whole-population before/after
Same-period cross-group gap
Reported range

No range appears here until the rollout's own comparison weeks are found and the metric is read both ways against them.

Every training evaluation guide names the same four levels: reaction, learning, behaviour, and business results, the ones the federal government's own training evaluation field guide names in that order. Almost every guide stops at the fourth level with a single instruction: measure the business metric before training and after. A business metric moving after training is a fact about a calendar window. It is not automatically a fact about the training, and most programmes report it as one anyway.

Isolating a training's own share of that number is a documented problem, not an afterthought. GAO's own guide to assessing federal training programs calls it especially difficult and names control groups, trained employees compared against untrained ones inside the same window, as one of the few ways to reveal a real difference. Most rollouts already contain that comparison and never use it. Cresswell Analytics trained 26 account executives in two waves six weeks apart, and its qualified-opportunity close rate rose afterward, 16.0 percent before training to 26.6 percent once both waves were trained.

A raw before-and-after reads that as a 10.6-point lift. Reading the six weeks where Wave 1 was trained and Wave 2 was not yet gives one estimate, a 6.0-point gap. Adjusting Wave 2's own before-and-after for what Wave 1's unchanged group drifted by across the same two windows gives a second, 5.5 points. The remaining 4.5 to 5.1 points traces to a lead-scoring change and a quarter-end pull-forward, both dated inside those same six weeks. Training Program Pack is where the evaluation plan behind this reading was designed, before delivery even started.

5.5 to 6.0 points, not 10.6

The rollout's own waves, the metric read two ways, and the isolated range against the naive number.

Rollout & Waves

Two waves, six weeks apart

Cresswell Analytics, 26 account executives. The gap between the two training dates is the comparison the rest of this sheet runs on.

WaveRepsAvg. TenureTrainedChecklist BeforeChecklist After
Wave 1143.2 yrsJan 5, 202632%78%
Wave 2120.9 yrsFeb 16, 202630%74%

Both waves start within 2 points of each other on the behaviour checklist and gain within 2 points of each other once trained, the evidence that the cross-wave gap on the next tab is comparing like with like.

Business Metric, Two Ways

Qualified-opportunity close rate, by period and wave

The same six-week windows the rollout schedule already provides, read before touching a single average.

PeriodWaveStatusOppsWonRate
PreCompany-wide (26)Not yet trained4006416.0%
Wk 1-6Wave 1 (14)Trained2105023.8%
Wk 1-6Wave 2 (12)Not yet trained1803217.8%
Wk 7-12Wave 1 (14)Trained, unchanged2055526.8%
Wk 7-12Wave 2 (12)Trained since wk 61754626.3%

Weeks 1-6: Wave 1 already trained, Wave 2 not yet, same calendar weeks. Weeks 7-12: a lead-scoring change and the fiscal quarter-end both land here, for both waves equally.

Attribution & Range

5.5 to 6.0 points, not 10.6

Two independent readings, half a point apart.

ReadingResultWhat it is
Naive whole-population before/after+10.6 pts16.0% to 26.6%, the dashboard number
Same-period gap, weeks 1-6+6.0 ptsWave 1 (trained) minus Wave 2 (not yet)
Wave 2 before/after, confound-adjusted+5.5 pts8.5-pt raw drift minus Wave 1's 3.0-pt drift
Reported range+5.5 to +6.0 ptsWhat the sponsor presentation actually claims

The 4.5-to-5.1-point gap between the naive number and the range traces to a lead-scoring change and a quarter-end pull-forward, both dated inside the same six weeks.

What's in the pack

01

Business Metric Correlation

Reads the metric two ways, a same-period comparison and a whole-population before-and-after, and reports the gap between them as a range.

02

Method and Caveat Note

States exactly what comparison the rollout provided, every confound dated inside the same window, and the one thing the method cannot rule out.

03

Reaction and Learning Results

The two levels that need no comparison group at all, a relevance rating and a pre-to-post assessment score, for every cohort separately.

04

Behaviour Change Indicators

A specific, countable checklist scored from real call or work-sample data, before training and after, for every cohort separately.

05

Evaluation Report

Opens on the business result as a range with both readings named, then the behaviour, learning and reaction evidence behind it.

06

Sponsor Presentation

Nine slides built from the report's own numbers, ending on what would strengthen next year's estimate rather than a bigger claim.

How to use it

  1. 1

    Open it in River, or download it

    Open the pack in River and have the agent find your rollout's own comparison group, or download the six blank Word and CSV files with no account.

  2. 2

    Send the rollout schedule and the data

    Who was trained and when, down to the wave, site or cohort level, plus whatever reaction, learning, behaviour and business-metric data already exists.

  3. 3

    Get the comparison the rollout already gives you

    The exact calendar weeks where one group was trained and another was not yet, named before a single business metric gets read.

  4. 4

    Get the business result as a range

    The metric read two ways, the gap between the two readings named as a number, and the one thing the method cannot rule out.

Frequently asked questions

Is this template free?

Yes. Download all six files as Word documents and CSV sheets with no signup and no card. Edit with AI is the optional path, where the agent finds your rollout's own comparison group and reads your business metric against it. The rest of the packs sit in the template library.

What format are the downloaded files?

Word documents (.docx) for the method note, the evaluation report and the sponsor presentation, and CSV (.csv) for the three data sheets, zipped together. They open in Word, Pages, Google Docs, Excel, Numbers and Sheets. In River the same content opens as live Docs and Sheets.

How does this find a comparison group if we never designed one?

Read the rollout schedule at the individual or team level, not the programme's overall dates. A wave gap, a pilot group, or a site trained later than the rest each puts some people on one side of a training date and others on the other. That gap, inside the same calendar weeks, is the comparison this pack looks for first.

What if everyone was trained on the same day, with no group held back?

Say so on the Method and Caveat Note rather than forcing a split that is not there. The fallback is a participant's or manager's own estimate of how much of a change belongs to the training, discounted because the person answering has a reason to overstate it. Report it as weaker evidence than a same-period comparison, in those words.

Does this replace Kirkpatrick's four levels or the ROI methodology?

No, it organizes around the same levels most evaluation guides already use, reaction, learning, behaviour, business results, in the order the federal government's own field guide names them. What it adds is the step most guides stop short of. Isolating the fourth level's number from everything else that moved is a named, expected technique, not an optional refinement, per the ROI Institute itself.

How is this different from a completion or compliance record?

A completion or compliance record proves the training happened and the right document is on file. Compliance Training and Evidence Pack builds that record. This pack starts after delivery and asks a different question: whether the training changed a real business number by more than everything else moving in the same window would have anyway.

What if we're not even sure training was the right fix?

Before this. Job Task Analysis and Competency Pack works out whether a task is frequent, critical or difficult enough to justify training at all, and Training Program Pack designs the modules and the evaluation plan once it is. This pack runs that plan once delivery is underway and completion data exists to read.

Find out how much of your lift is actually the training

Download the blank pack as Word and CSV files, or open it in River and have the range computed from your own rollout schedule and business metric.

Edit with AI