Costs and pricing
Part of Social media training: how to scope it and request comparable quotes
Social media training return on investment: a method
Social media training return on investment: separate activity, competence, outcomes and causal effects, then calculate full economic cost.
A completion certificate alone cannot show social media training return on investment (ROI). Measure the required workplace task, the complete economic cost and what would probably have happened without the intervention.
What to take away
- A completion certificate cannot establish financial return; measure the required workplace task and full economic cost.
- Keep activity, competence, observed outcome and causal effect as separate evidence classes to avoid overclaiming.
- Calculate whole-life economic cost using buyer records and verified quotes before valuing any benefit.
- Only value an outcome if finance documents the calculation, period, source and uncertainty.
- If benefit cannot be valued without speculation, report competence and operational evidence alongside cost.
Write the decision and causal claim
Name the England learner population, training version, delivery dates, target task and decision the evaluation will inform.
A suitable task might be producing a compliant draft that passes a defined review, approving a compliant post through a defined workflow, or handling response times for customer replies within a stated service level, not an imprecise aim such as improving social media.
Define the evaluation scope
- Name learner population and training version
- Set delivery dates and target task
- Write causal question before results
- Assign evaluation owner separate from supplier
Write the causal question before collecting results: did this version of the programme change the specified outcome for eligible learners, compared with the chosen counterfactual, over the stated period? Assign an evaluation owner who is separate from the supplier's sales evidence.
The Magenta Book, HM Treasury's impact-evaluation guidance, explains that experimental and quasi-experimental impact evaluation uses a comparison representing what would have occurred without the intervention. It is central-government guidance; it does not prove that a commercial course caused any result.
Keep four evidence classes apart
- Activity:invitations, attendance, practice or completion. State the eligible population, count, period, source and exclusions.
- Competence:accepted performance on a defined task. Retain the assessment version, assessor, rubric, attempts and accessibility provisions.
- Observed outcome:a before-and-after operational change. Provide numerator, denominator, period, source query, other changes and missing records.
- Causal effect:the difference attributable to the programme under a defensible design and stated assumptions.
Platform reach, sales, complaints or processing time may change for unrelated reasons. A supplier case study or learner opinion may help form a hypothesis, but it is not the buyer's counterfactual.
Four evidence classes
Activity
- Definition
- Invitations, attendance, practice
- Key records
- Population, count, period
- Exclusions
- State exclusions
Competence
- Definition
- Accepted task performance
- Key records
- Rubric, assessor, attempts
- Exclusions
- Accessibility provisions
Observed outcome
- Definition
- Before-and-after operational change
- Key records
- Numerator, denominator, source
- Exclusions
- Missing records
Causal effect
- Definition
- Difference attributable to programme
- Key records
- Design and assumptions
- Exclusions
- Counterfactual choice
Calculate full economic cost first
Use only buyer records and verified quotes:
Whole-life economic cost
- Supplier costs
- Internal design and procurement
- Learner and manager time
- Assessment and accessibility
- Privacy, security, IP and legal review
- Support, correction and exit
line cost = accepted unit amount x evidenced quantity
whole-life economic cost = supplier costs + internal design and procurement + learner and manager time + assessment + accessibility + privacy and security + IP and legal review + support + correction + exit
Keep currency, VAT basis, period, source date and confidence with every input. HMRC's National Minimum Wage Manual (NMWM) training-time guidance establishes a minimum-wage boundary for employer-approved training time, but payroll must calculate the actual labour cost. HMRC's VAT guidance shows why recoverability requires buyer-specific review.
Populate only supported benefit fields
Define the outcome unit before valuing it. The blank method is:
observed change = accepted outcome after - accepted outcome before
incremental effect = accepted outcome with training - accepted counterfactual outcome
accepted incremental benefit = incremental effect x finance-approved value per outcome unit
return ratio = (accepted incremental benefit - whole-life economic cost) / whole-life economic cost
For arithmetic illustration only, assume whole-life economic cost is £10,000 and accepted incremental benefit is £15,000. Then return ratio = (£15,000 - £10,000) / £10,000 = 0.5, or 50 per cent. These figures are illustrative only and not a benchmark; they depend on the buyer's own cost records and finance-approved value per outcome unit.
Do not insert a value per click, hour, error or sale unless finance documents the calculation, period, source and uncertainty. Do not count the same effect as both avoided cost and added income. Present adverse and base sensitivities using approved inputs, not invented likelihoods.
Decide what can be said
Publish the estimate with eligibility, sample, missingness, measurement changes, contamination, confidence and external-validity limits. If the design supports only association, label it attributed or observed. If benefit cannot be valued without speculation, report competence and operational evidence alongside cost.
Set a review trigger for changed course content, learner eligibility, outcome definitions or source systems, because any of them can break comparability.
Finance, evaluation and learning reviewers should sign the input register and causal boundary. The register should record each input's amount or quantity, calculation, period, source, VAT basis and confidence, with counterfactual assumptions recorded separately. Privacy and legal owners should approve data use independently. The honest result may be that value remains unknown. That conclusion is more useful than a precise ratio assembled from unsupported assumptions.
Before you act
- Name the learner population, training version, delivery dates and target task.
- Write the causal question before collecting results.
- Assign an evaluation owner separate from the supplier's sales evidence.
- Calculate full economic cost using only buyer records and verified quotes, and record each input in the input register.
- Define the outcome unit before valuing it.
- Set a review trigger for changed course content or eligibility.
Common questions
How should I calculate full economic cost?
Use only buyer records and verified quotes. Line cost equals accepted unit amount times evidenced quantity. Whole-life economic cost adds supplier costs, internal design and procurement, learner and manager time, assessment, accessibility, privacy and security, IP and legal review, support, correction and exit. Keep currency, VAT basis, period, source date and confidence with every input.
What should I do if benefit cannot be valued without speculation?
Report competence and operational evidence alongside cost. Publish the estimate with eligibility, sample, missingness, measurement changes, contamination, confidence and external-validity limits. If the design supports only association, label it attributed or observed. The honest result may be that value remains unknown, which is more useful than a precise ratio from unsupported assumptions.
