Costs and pricing
Part of Social listening budget: the four ledgers behind the licence
Social listening return on investment begins with a causal question
Assess social listening return on investment in England with a defined comparator, verified incremental benefit, whole-life cost and uncertainty.
Social listening return on investment is calculable only when the buyer has a credible estimate of incremental benefit and a matching whole-life cost. Mentions, alerts and revenue recorded after an alert do not establish causation.
What to take away
- Define one decision, population, intervention date, outcome and comparison before calculating any social listening ROI.
- Mentions, alerts and post-alert revenue do not establish causation without a defensible counterfactual.
- State each benefit in physical units and get finance-approved values before converting to money.
- A negative or uncertain result can still support a decision to change scope or stop.
Write the causal question first
Choose one decision, eligible England population, intervention date, outcome and comparison. For example, did a new listening-led escalation route change verified resolution cost for eligible cases compared with the process that would otherwise have operated? Do not broaden the result to all customers or England.
Plan the evaluation before launch. HM Treasury's Magenta Book covers scoping, design, conduct and use of government evaluation. Its principles can inform a commercial pilot, but the book does not endorse a supplier or validate private ROI.
A typical England evaluation runs for 12 months: two weeks to write the protocol, three months to collect data, one month to analyse and one month to assure and publish.
Separate the evidence layers
Report activity first: items retrieved, reviewed alerts and referrals. Next record observed outcomes with their period and eligibility. Attribution links an outcome to a tracked pathway but may not be causal. Incrementality needs a defensible counterfactual, such as random allocation where appropriate or a well-justified comparison with documented assumptions.
Evidence layers
- Activity: items, alerts, referrals
- Observed outcomes: period and eligibility
- Attribution: tracked pathway, not causal
- Incrementality: defensible counterfactual
For an England counterfactual, buyers can use ONS consumer trends, ONS retail sales indices, NHS England waiting times, Companies House filings and the UK Data Service.
Pre-specify exclusions, contamination checks, missing-data treatment and adverse effects. If the intervention changes which cases become visible, naive before-and-after averages may compare different populations.
Build the benefit numerator
For each proposed benefit, state the physical unit before its monetary value. A verified reduction in staff minutes needs an independently measured time difference, eligible case count and finance-approved labour value. Avoided loss needs evidence that the event would have happened without the intervention; a hypothetical crisis value is not a realised saving.
ROI formula
ROI percentage = (verified incremental benefit in GBP minus matched whole-life cost in GBP) / matched whole-life cost in GBP x 100
Numerator: net incremental benefit
Denominator: full cost for same intervention, population and 12-month period
Exclude impressions, sentiment scores, gross attributed revenue, unverifiable avoidance
Use this calculation only after the evidence review:
ROI percentage for the defined 12-month England intervention = (verified incremental benefit in GBP minus matched whole-life cost in GBP) / matched whole-life cost in GBP x 100
The numerator is net incremental benefit; the denominator is full cost for the same intervention, population and twelve-month period. Include finance-valued causal outcomes and licence, usage, implementation, analysis, assurance, support and exit costs. Exclude impressions, sentiment scores, gross attributed revenue, unverifiable avoidance and benefits outside the period.
Sources are the approved evaluation dataset, documented counterfactual, finance valuation and cost ledger.
A typical England social listening licence for a small team is £12,000 to £36,000 per year, and analyst time is often £350 to £650 per day. Both depend on seats, data volume, modules and seniority.
Illustrative example: a 12-month England pilot with £18,000 licence, 40 analyst days at £450 per day (£18,000) and £6,000 assurance gives matched whole-life cost of £42,000. A verified incremental benefit of £63,000 gives ROI of 50 per cent.
The evaluation lead owns effect estimation; finance owns valuation and cost. Test comparator choice, lag, unit value, missing cases and cost changes. Stop publishing ROI when the comparison fails, contamination crosses the pre-set limit or cost scope changes; review on the protocol's dated checkpoints.
Assure the analysis
The 2025 AQuA Book is government guidance on proportionate analytical assurance, validation and uncertainty. A private team can use its disciplines without claiming government approval. Keep code or workbook versions, input provenance, checks, reviewer comments and an assumption log.
Publish ROI or descriptive monitoring
Credible counterfactual exists?
Publish ROI with sensitivity cases
Publish cost per reviewed item or cost alongside observed outcomes
Show the benefit and cost components beside the ratio. Provide sensitivity cases rather than one false-precision percentage.
Preserve a decision table with baseline, comparison method, sample, effect estimate and confidence or uncertainty statement. Include unit valuation, cost version and named approvers, and set continuation criteria before results are known.
A team might require valid data collection and an affordable cost ceiling, but must choose its own thresholds from risk and evidence. Missing outcomes, changed eligibility or a broken comparison need repair or a descriptive report, not a more favourable formula.
Named UK finance, evaluation, statistics, tax, commercial and legal reviewers should approve the method and wording. A negative or uncertain result may still support a decision to change scope or stop; the calculation is evidence for a choice, not a target to defend.
Before you act
- Plan the evaluation using Magenta Book principles.
- Separate activity, observed outcomes, attribution and incrementality.
- Pre-specify exclusions, contamination checks and missing-data treatment.
- Assure the analysis with version control and an assumption log.
- Stop publishing ROI when the comparison fails or cost scope changes.
Common questions
Who owns effect estimation and valuation in the evaluation?
The evaluation lead owns effect estimation; finance owns valuation and cost. Named UK finance, evaluation, statistics, tax, commercial and legal reviewers should approve the method and wording. The numerator is net incremental benefit and the denominator is the full cost for the same intervention, population and twelve-month period.
