Costs and pricing
Part of A UK scoping framework for social media scheduling
Social media scheduling return on investment: what to claim
Social media scheduling return on investment depends on the evidence you hold: match the claim to your counterfactual, attribution share and switching value.
The ROI claim a scheduling tool can support depends on the comparison you hold, not on the dashboard. With a matched comparison group, a project ratio is defensible. With only before-and-after figures, the honest claim is narrower. The tool, whether Buffer, Hootsuite, Later, Sprout Social, SocialBee or Metricool, does not change that. Dashboards also cannot settle attribution methods.
What to take away
- Define the project and the counterfactual before measuring social scheduling ROI.
- The claim you may make shrinks as the comparison weakensmatched group, intervention group only, dashboard only.
- Attributable contribution needs a supported counterfactual, not a last-click label.
- Stress-test the ratio with sensitivities and state the switching value for the decision.
- Privacy, PECR, accessibility and security gates cannot be traded for a positive ratio.
Name the counterfactual first
A counterfactual is what would have happened without the tool. The Magenta Book describes it as an unaffected group acting as proxy for what would have happened to the affected group in the absence of the intervention. For scheduling, the stand-in is business as usual: the current process.
That is government evaluation guidance, so a private firm should adapt the discipline and not claim formal compliance with it. Start by fixing the project: the exact plan, seats, connected channels and post limits, the England operating team, the affected accounts, the workflow change, the start date and the measurement period.
Measurement card fields
- Event and population
- Numerator and denominator
- Unit and source
- Query version and period
- Timezone and exclusions
- Owner
Choose outcomes before launch. Each needs an event, population, numerator, denominator, unit, source, query version, period, timezone, exclusions and owner. Keep the organic and paid metrics a scheduler reports for Instagram, TikTok, LinkedIn or X separate from internal orders, contribution and cash receipts.
Freeze that measurement card before the workflow changes, and keep the same eligibility and observation window for the baseline and the intervention. Note seasonality, paid-media changes, promotions and product availability, which can all move the outcome. If the groups differ materially, report the limitation rather than forcing one effect estimate.
What each level of evidence lets you claim
Match the sentence in the report to the design behind it. HM Treasury's quality in policy impact evaluation guidance says causality can rarely be confidently attributed without a comparison or control group, matched on relevant variables or randomly allocated.
| Evidence you hold | Claim you can make | Claim to avoid |
|---|---|---|
| Matched comparison or control group, same window | An estimated incremental difference, with the attribution share the design supports | A precise result that goes beyond the design |
| Intervention group only, no comparison | The tool was introduced, an outcome moved, and interviews or process records support an explanation | A causal percentage |
| Scheduler dashboard only | Platform-reported activity, such as posts, reach and clicks | Any ROI figure, because a last-click label or post timestamp is an association |
In the middle row, the team may report that the tool arrived, that an outcome moved and that process records fit a contribution story. It should not turn that sequence into a number. Record alternative explanations, and the evidence that would weaken the preferred account.
From gross value to attributable contribution
Start with observed gross value. Remove cancellations, refunds, variable fulfilment costs and other agreed deductions to reach contribution. Estimate the incremental difference against the counterfactual, then apply only the share of it that the evaluation design supports.
Attributable contribution steps
- Start with observed gross value
- Remove cancellations and refunds
- Deduct variable fulfilment costs
- Estimate incremental difference vs counterfactual
- Apply supported attribution share
attributable contribution = observed incremental units versus counterfactual x verified contribution per unit x supported attribution share
Every field stays blank until evidence exists. An empty field is a finding: it shows exactly which piece of the claim the project cannot yet support, and it tells the next study what to measure.
The cost side, kept short
The denominator is a budgeting job in its own right. Take it from the budget framework, which keeps cash spend, VAT, full economic cost, gross value and attributable contribution in separate fields, and bring only the total across.
Full economic cost items
- Subscription and add-ons
- Implementation and migration
- Integrations and training
- Content work and licences
- Accessibility, privacy and claim review
- Security assurance and operations
- Support, incidents, renewal and exit
full economic cost = supplier cost excluding confirmed recoverable VAT + internal resource cost + contractor cost + risk and exit cost
State currency, period and tax treatment. Do not use a generic wage rate or assume VAT recovery: HMRC's business-expense guidance makes recovery conditional on registration, business use, records and other rules. Include the workflow change itself, since training and migration are costs of the project.
Calculate, then stress-test
project ROI = (attributable contribution - full economic cost) / full economic cost
This is an internal project ratio, not accounting profit, a tax result or a market benchmark. Report the numerator, denominator, currency, period and uncertainty beside it. Do not calculate when full economic cost is zero or incomplete.
Run sensitivities for lower contribution, weaker attribution, higher operating cost and delayed outcomes. The switching value is the point at which the decision changes, such as the contribution below which expansion stops. Name who can stop expansion if evidence falls below the agreed threshold.
Present the result as a range or as scenarios when input uncertainty is material. Keep the base inputs available for reproduction, and review once enough eligible observations exist. A positive early result does not override a small denominator, an incomplete cost record or an inaccessible measurement method.
Privacy, PECR and UK GDPR, accessibility to WCAG 2.2, advertising, IP and security gates do not become financial benefits. The ICO enforces the UK privacy rules behind them. This method offers general evaluation information, not investment, accounting, tax or legal advice; qualified reviewers must approve the real model.
Before you act
- Name product, plan, team, accounts, workflow change, start date and measurement period.
- Freeze the measurement card before the workflow changes.
- Say which row of the evidence table you are in before writing the claim.
- Calculate full economic cost including internal hours and exit costs.
- Separate platform-reported clicks from internal orders and cash receipts.
- Run sensitivities and state the switching value.
Common questions
What can I claim without a comparison group?
That the tool was introduced, that an outcome moved and that interviews or process records support an explanation. HM Treasury's guidance says causality can rarely be confidently attributed without a comparison or control group, so do not turn the sequence into a causal percentage.
What is a counterfactual in social scheduling ROI?
It is the result expected if the organisation kept its current process. The Magenta Book calls an unaffected group a proxy for what would have happened to the affected group without the intervention. A platform dashboard cannot supply that comparison on its own.
What is a switching value?
It is the point at which the decision changes, for example the level of attributable contribution below which expansion should stop. State it beside the ratio, with the sensitivities for lower contribution, weaker attribution, higher operating cost and delayed outcomes, so reviewers can see how fragile the result is.
