In short
- Score each creator after every campaign on outcomes (40), delivery (15), reliability (15), content (20) and audience fit (10), with pass-or-fail gates for compliance and conduct on top.
- Judge outcomes against peers in the same tier, platform and format, and pull the score toward the middle when a creator has few tracked results.
- Use bands: 75 or more with enough data is a priority rebook, below 40 is a pause. A flag adds a named fix to the brief, not a penalty.
- Separate brand-side causes from creator-side ones, or the scorecard punishes creators for late products and broken links.
A creator performance scorecard rates each creator you have booked on the same five dimensions after every campaign: outcomes against peers, delivery against forecast, reliability, content quality and audience fit. Two pass-or-fail gates, compliance and conduct, sit on top. A correction for small samples keeps a lucky creator from topping the list, and bands turn the score into a decision: rebook, test again or pause.
The creator table in our influencer campaign report template records what happened in one campaign, and the creator KPIs in our KPI guide define the measures. This post is for a repeat programme that needs one comparable view across campaigns. It sits within the complete guide to influencer marketing in India, uses the definitions in how to measure influencer marketing ROI, and follows influencer cost per acquisition on the cost side. The need is real: about 83% of surveyed Indian brands say they struggle to find the right creators, per a 2025 Kantar-backed study covered by Campaign India, and a scorecard is how a brand stops starting from zero each time.
Three scores, three moments
| Tool | When | Question | Where it lives |
|---|---|---|---|
| Vetting score | Before booking | Is this creator real, safe and a fit? | Our Gully Score, or the influencer vetting checklist |
| Content score | At draft and at day 7 | Is this post good? | How to evaluate influencer content quality |
| Performance scorecard | After each campaign, then rolling | Should we rebook, and on what terms? | This post |
The Gully Score is our 100-point vetting score: audience authenticity 30, engagement quality 25, buyer fit 20, content craft 15 and brand safety 10, described on our how we work page. The scorecard uses the content score as one input and checks the vetting score's prediction against what happened.
The five dimensions and two gates
These bands are our working choices, not published thresholds, and we found no India benchmark for any of them. Set your own once you have two or three campaigns of data.
| Dimension | Points | Measure | How points are awarded |
|---|---|---|---|
| Outcomes | 40 | Creator cost per result ÷ peer median | At or below 0.70: 100% of points. 0.70 to 1.00: 80%. 1.00 to 1.30: 60%. 1.30 to 2.00: 30%. Above 2.00: 0 |
| Delivery | 15 | Day-30 views ÷ the creator's forecast | 100% or more: 15. 80 to 99%: 12. 60 to 79%: 8. 40 to 59%: 4. Below 40%: 0 |
| Reliability | 15 | On-time posting (6), approval rounds (5), responsiveness (4) | On time: all 6, one late 3, more 0. One round 5, two 3, three or more 0. Always responsive 4, usually 2, rarely 0 |
| Content | 20 | The day-7 content score | Content score ÷ 100 × 20 |
| Audience fit | 10 | Share of the creator's audience in your target market and age band | 60% or more: 10. 40 to 59%: 6. 20 to 39%: 3. Below 20%: 0 |
| Gate | Pass | Flag | Fail |
|---|---|---|---|
| Compliance | Correct label at first publish | Label missing or wrong, fixed within 24 hours | Not fixed, or repeated |
| Conduct and brand safety | No issues | A minor issue, handled | A serious issue, or a breach of the morality clause |
A fail pauses the creator whatever the score. A flag adds a condition: rebook only if the next brief names the fix. In ASCI's 2025-26 annual report, 97.3% of the 1,609 influencer ads it processed needed modification (ASCI), which is why compliance is a gate and not a few points among many. This is general information, not legal advice.
Judge outcomes against peers
Cost per result means little on its own. Divide each creator's all-in cost per result, with fee and rights, by the median of creators in the same tier, platform and format in the same campaign period. A nano Story set and a mid-tier Reel have different economics, and ranking them against each other produces a list of tiers. If a peer group has fewer than five creators, widen it by dropping one criterion, such as language, and note that you did.
Small numbers: pull toward the middle
A creator with nine tracked orders and a great cost per result has not proved much. A count of nine has a 95% range of about ±65%, as the sample-size rule in the CPA guide shows. So shrink the outcome points toward half marks when results are few.
Outcome points = 40 × (w × band share + (1 − w) × 0.5), where w = n ÷ (n + 15), n is the number of tracked results and 15 is our working constant.
| Tracked results | Weight on the creator's own band |
|---|---|
| 5 | 0.25 |
| 9 | 0.38 |
| 15 | 0.50 |
| 25 | 0.63 |
| 46 | 0.75 |
| 64 | 0.81 |
Tag confidence separately: high if there are at least 25 tracked results across at least two posts, low otherwise. A low-confidence creator can score well, but cannot reach "priority".
Five creators scored (made-up numbers)
The creators are labelled A to E. The figures are invented to show the arithmetic.
| Creator | Results | Outcome (40) | Delivery (15) | Reliability (15) | Content (20) | Audience (10) | Total | Gates | Confidence |
|---|---|---|---|---|---|---|---|---|---|
| A | 46 | 35.1 | 15 | 15 | 16.4 | 10 | 91.5 | Pass | High |
| B | 9 | 27.5 | 12 | 15 | 15.6 | 10 | 80.1 | Pass | Low |
| C | 31 | 22.7 | 8 | 8 | 14.0 | 6 | 58.7 | Pass | High |
| D | 64 | 13.5 | 15 | 13 | 12.8 | 3 | 57.3 | Pass | High |
| E | 22 | 27.1 | 4 | 2 | 9.6 | 6 | 48.7 | Flag | Low |
Take creator B. Cost per result was 0.58 of the peer median, a full-marks band, but with nine results the weight is 9 ÷ 24 = 0.375. Outcome points are 40 × (0.375 × 1 + 0.625 × 0.5) = 27.5. Without the correction, B would have scored 92.6, nearly level with A on a fifth of the evidence.
Bands and decisions
| Score and confidence | Decision | Terms |
|---|---|---|
| 75 or more, high | Priority rebook; consider a multi-month deal | Hold the fee; ask for rights options |
| 75 or more, low | Rebook as a volume test | Same fee, more posts, to build evidence |
| 55 to 74 | Rebook with named fixes | Negotiate on the weakest dimension |
| 40 to 54 | One more test, smaller scope or lower fee | Set a stop rule |
| Below 40, or any gate fail | Pause and review | No new booking |
Applied to the five: A is a priority rebook, and a candidate for the creator ambassador programmes. B is a volume test. C is a rebook with fixes: one late post, extra revisions and only 52% of the audience in the target market. D delivered every view and cost 1.85 times the peer median per result, with only 38% of the audience in the target market, so rebook only at a lower fee or with regional targeting. E scored 48.7 with a flag for a label that was missing at first publish and fixed in 24 hours: one smaller test, with the label step written into the brief.
Read the trend, not only the latest score
Keep the last three campaigns' scores and weight them 50, 30 and 20 from newest to oldest. Take a creator F, outside the table above, who scored 88, then 79, then 70. The weighted figure is 76.3, which still clears 75, but the 18-point fall in three campaigns is the story. Check delivery against forecast and the content score for the cause. Repeated sponsored posts to the same followers can tire an audience, so look at whether views per post are falling across bookings and consider rotating the format.
Separate brand-side causes
A scorecard that ignores why something went wrong teaches creators to distrust it.
| Cause | Example | Treatment |
|---|---|---|
| Late product delivery | Creator posts late because the box arrived late | Exclude the post from reliability |
| Slow approvals | Brand took six days to approve | Exclude the delay; note it |
| Broken link or code | Orders not tracked | Exclude the affected post from outcomes |
| Stock-out during the live window | Orders capped | Exclude or re-weight the post |
| Brief changed mid-way | Extra revision rounds | Do not count the extra rounds |
Mark each exclusion on the sheet with a reason. Score only what the creator controlled, and use audience insights, not personal attributes, for the fit dimension.
Using the scorecard
| Use | How |
|---|---|
| Fee negotiation | Pay for the creator's median views, as in calculating influencer CPV and CPM, and reward results above forecast with a bonus |
| Contract terms | A make-good clause for late posts; see influencer contracts in India |
| Ambassador selection | Start from high-confidence priority creators |
| Creator ads | Pick posts for amplification from creators with steady content scores; see creator ads and whitelisting |
| Casting and forecasting | Feed each creator's history into the shrinkage in forecasting influencer campaign results |
Add a cohort column when you have enough customers per creator: repeat purchase by creator from influencer customer cohort analysis tells you who brings customers who stay. Log what the scorecard teaches you in the learnings library.
The routine after each campaign
- Lock the campaign data at day 30, as in the post-campaign analysis checklist.
- Compute peer medians of cost per result by tier, platform and format.
- Fill the sheet: outcomes and delivery from the data, reliability from the approval log, content from the day-7 scores and audience fit from insights.
- Apply the gates and mark brand-side exclusions, each with a reason.
- Review with the strategist, agree each decision and write the named fix for every flag.
- Send each creator a summary of their own result and any fix before the next brief.
Keep the sheet beside the learnings library and update the three-campaign trend at the same time.
Mistakes to avoid
- Ranking on cost per result with six orders behind it.
- Comparing creators across tiers and formats.
- Giving compliance points instead of a gate.
- Scoring creators for brand-side delays.
- Changing the weights between campaigns and then comparing scores.
- Keeping the scorecard private so creators cannot see what they are measured on.
- Reading one campaign as a creator's whole record.
If you want a scorecard set up for your next creator programme, with peers and bands agreed before launch, send us a brief. Our micro-influencer marketing service is where scorecard data is most useful, because many creators mean many small samples, and the rates guide covers the fees behind every cost per result.
Sources
- 69% of Indian consumers turn to influencers for information and 63% for product discovery (consumer and brand survey coverage), Campaign India (12 Jun 2025)
- ASCI annual complaints report 2025-26: influencer ads processed and share needing modification, Advertising Standards Council of India (28 May 2026)
Frequently asked questions
What is a creator performance scorecard?
It is a fixed sheet that rates every creator you have booked on the same dimensions after each campaign: results against peers, delivery against forecast, reliability, content quality and audience fit, plus compliance and conduct gates. It produces a score and a decision, such as rebook, test again or pause, and keeps the reasoning visible.
What should a creator scorecard include?
Cost per result against peers, views delivered against the creator's own forecast, on-time posting, approval rounds, responsiveness, a content score, audience fit and a disclosure check. Our campaign report template holds the raw table. Add a confidence tag so small samples are not over-read.
How many results does a creator need before I rank them?
Our working rule is about 25 tracked results across at least two posts for a high-confidence score, in line with the sample-size reasoning in our cost per acquisition guide. Below that, shrink the outcome score toward the middle and label the creator for another test.
How is a scorecard different from the Gully Score?
The Gully Score is our 100-point vetting score used before booking: audience authenticity 30, engagement quality 25, buyer fit 20, content craft 15, brand safety 10. The scorecard rates what happened after booking. Use the vetting score to predict and the scorecard to learn. See how we work.
Should compliance be part of the creator score?
Treat it as a gate, not points. A disclosure label that is missing or wrong is a rule problem, not a 5-point deduction. Pass, flag and fail work better: a flag adds a named fix to the next brief and a fail pauses the creator. See our ASCI guidelines explainer.
Should I share scorecards with creators?
Share the criteria and each creator's own summary, not other creators' data. Creators work better when they know the measures, and it makes rebook conversations about facts. Mark brand-side delays on the sheet so the creator can see they were not scored against.