TANKE reference guide
Which Influencer Marketing KPIs Should You Track? The Complete Guide
More than 150 metrics organized by objective, with formulas, denominators, sources and caveats.
Direct answer
The right influencer marketing KPIs depend on the objective and campaign mechanism. A reliable system connects the business objective, creator and content role, KPI, denominator, data source and attribution rule. It reads operational, media, brand and commercial evidence together instead of collapsing performance into one vanity score.
The TANKE KPI map
Start with the job the campaign must do, then define the measurement contract. This table summarizes the families before the detailed library.
| Objective or family | Indicative KPI | What it measures | Main caveat |
|---|---|---|---|
| Delivery and exposure | Reach | Potential and delivered exposure | Not the same as creators who publish |
| Attention and consumption | Qualified views and watch time | Actual content consumption | A start, 2-second view and qualified view are not equivalent |
| Engagement and resonance | Engagement rate with named denominator | Audience response and resonance | Publish the included action set |
| Traffic and consideration | Landing-page sessions | Movement to owned properties | Platform clicks may include repeats and non-landing clicks |
| Conversion, revenue and economics | Conversions, ROI and incremental revenue | Commercial outcomes and economics | Define deduplication and qualification |
| Paid amplification and media efficiency | CPM, CPV, CPC and CPA | Paid distribution efficiency | Not comparable if impression quality differs |
| Brand impact and incrementality | Brand lift and cost per incremental outcome | Attitudinal and causal impact | Sample and question wording matter |
| Product seeding and gifting | Activation rate and content yield | Gifting output and efficiency | Remove duplicates and invalid contacts |
| Programme operations and creator relationships | Acceptance, cycle time and retention | Programme execution and relationships | Automated replies and duplicates need rules |
| Content asset and rights value | Usable assets and reuse rate | Content and rights value | Define uniqueness, quality and rights |
| Community, advocacy and long-term value | Retention, advocacy and repeat performance | Long-term relationship value | Identity resolution is partial on platforms |
| Measurement and data quality | Tracking coverage and confidence | Reliability of every other KPI | Tags do not solve identity or consent loss |
The six-field measurement contract
Selection is not a universal funnel. Awareness campaigns, product seeding, paid amplification and ambassador programmes use different mechanisms and standards of proof. TANKE recommends one primary KPI, two to four diagnostic KPIs and at least one data-quality KPI.
- Business objective
- Creator or content mechanism
- KPI and unit
- Denominator
- Data source
- Window and attribution rule
This extends TANKE’s view on vanity metrics in influencer marketing: expose, qualify and connect every number to a decision.
Complete KPI library
Every row states the unit or formula, denominator, best use and what the KPI does not prove.
Delivery and exposure
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Creators contracted | Count | None | Capacity and commitment | Not the same as creators who publish |
| Creators activated | Count meeting activation rule | None | Actual programme participation | Define activation before launch |
| Deliverables planned | Count | None | Scope baseline | Plans can change; version the baseline |
| Deliverables delivered | Count | None | Production output | Delivery does not imply approval or publication |
| Deliverables published | Count | None | Live output | Remove duplicates and expired stories |
| Delivery or completion rate | Delivered deliverables / planned deliverables × 100 | Planned deliverables | Scope control | State whether partial delivery counts |
| Publication rate | Published deliverables / approved or planned deliverables × 100 | Approved or planned deliverables, explicitly stated | Execution reliability | Denominator choice changes the result |
| On-time delivery rate | On-time deliverables / due deliverables × 100 | Deliverables due in period | Operational predictability | Use agreed deadlines and approved extensions |
| Reach or unique reach | Unique accounts or people exposed | Unique eligible people/accounts | Audience exposure | Platform estimates may not deduplicate across creators |
| Estimated reach | Modelled unique exposure | Model-defined audience base | Planning where observed reach is unavailable | Label as estimated and disclose model |
| Impressions | Total content displays | None | Delivery volume | One person may generate several impressions |
| Frequency | Impressions / reach | Reach | Exposure intensity | Invalid if reach and impressions use different universes |
| Follower or subscriber base | Account followers or subscribers | None | Screening and potential audience context | Followers are not reach |
| Share of voice | Brand mentions or exposure / category mentions or exposure × 100 | Defined category conversation | Competitive presence | Sensitive to query, channels and weighting |
| Creator or category presence | Qualified creator mentions / monitored creator mentions × 100 | Defined monitored creator set | Cultural presence | Define the creator universe before comparison |
Attention and consumption
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Video starts or views | Platform count | Platform-specific start or view event | Top-line consumption | A start, 2-second view and qualified view are not equivalent |
| Qualified video views | Count meeting named threshold | Views or impressions eligible for threshold | Comparable attention within one definition | Name seconds, completion or interaction rule |
| View-through rate (VTR) | Qualified views / video impressions × 100 | Video impressions | Creative attention efficiency | Platforms use different qualified-view rules |
| Video completion rate | Completed views / video starts or qualified views × 100 | Starts or qualified views, explicitly stated | Depth of video consumption | Short videos naturally complete more often |
| Average watch time | Total watch time / video plays | Video plays | Average attention depth | Replay and autoplay handling varies |
| Total watch time | Sum of watched time | None | Aggregate attention delivered | Scale metric, not individual quality |
| Audience retention curve | Viewers remaining at each timestamp / starters × 100 | Video starters | Find drop-offs and strong moments | Requires comparable length and format |
| Dwell time | Time with page, post or experience in view | Sessions or eligible exposures | Reading or experience depth | Background tabs and passive exposure can inflate it |
| Hook rate | Users reaching named early threshold / video starts × 100 | Video starts | Opening creative strength | Not standardized; name threshold |
| Hold rate | Users reaching later threshold / users reaching hook threshold × 100 | Hook-qualified viewers | Retention after the hook | Custom definitions are not cross-platform standards |
| Thumb-stop rate | Named early view or stop event / impressions × 100 | Impressions | Paid-feed stopping power | Use only with documented platform definition |
Engagement and resonance
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Total engagements | Sum of included actions | None | Interaction volume | Publish the included action set |
| Engagement rate by followers | Engagements / followers × 100 | Followers at a stated date | Creator screening | Followers may not have seen the content |
| Engagement rate by reach | Engagements / reach × 100 | Reached accounts or people | Response among exposed audience | Reach may be estimated and unavailable |
| Engagement rate by impressions | Engagements / impressions × 100 | Impressions | Interaction per delivered exposure | Repeat impressions lower the rate |
| Engagement rate by views | Engagements / named view count × 100 | Views under named definition | Video response | View thresholds differ |
| Likes or reactions | Count | None | Lightweight resonance | Low effort and easy to decontextualize |
| Comments | Count | None | Conversation volume | Count does not show quality or sentiment |
| Shares or reposts | Count | None | Distribution intent | Private shares may be incomplete |
| Saves | Count | None | Future utility or reference value | Not available consistently across platforms |
| Comment quality rate | Meaningful comments / coded comments × 100 | Comments reviewed under a codebook | Depth and relevance | Requires a reproducible coding rule |
| Meaningful interaction rate | Weighted or qualified actions / chosen exposure base × 100 | Named reach, views or impressions | Prioritize higher-intent actions | Weights are analytical choices, not facts |
| Share rate | Shares / chosen exposure base × 100 | Reach, impressions or views, named | Content transmission | Denominator must be stated |
| Save rate | Saves / chosen exposure base × 100 | Reach, impressions or views, named | Reference or utility signal | Unavailable or differently exposed by platform |
| Sentiment | Positive, neutral and negative coded mentions / coded mentions × 100 | Mentions with enough context | Response direction | Sarcasm, language and sampling reduce accuracy |
| Conversation themes | Coded theme incidence | Coded comments or mentions | Understand why people respond | Themes need a stable codebook |
| Earned mentions and UGC response | Qualified organic mentions or assets | None | Secondary amplification | Do not assume every mention was caused by the campaign |
| Audience growth associated with campaign | End audience minus baseline audience | Baseline audience | Directional community growth | Temporal association is not causal attribution |
Discuss your measurement system
Traffic and consideration
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Link clicks | Count | None | Traffic intent | Platform clicks may include repeats and non-landing clicks |
| Click-through rate (CTR) | Link clicks / impressions × 100 | Impressions | Creative traffic efficiency | Use link clicks, not all clicks |
| Landing-page sessions | Attributed tracked sessions | None | Arrivals on owned property | Consent loss and redirects create gaps |
| Landing-page users | Deduplicated tracked users | None | Unique site audience | Identity and device deduplication are limited |
| Engaged sessions | Engaged sessions under analytics definition | Sessions | Visit quality | Document the analytics definition and changes |
| Landing-page engagement rate | Engaged sessions / sessions × 100 | Sessions | Post-click quality | Tracking events can inflate the result |
| Bounce rate | Non-engaged sessions / sessions × 100 | Sessions | Identify weak landing alignment | Definition differs between analytics generations |
| Exit rate | Exits from page / pageviews × 100 | Pageviews | Journey drop-off context | An exit can be a successful final step |
| Cost per click (CPC) | Eligible cost / link clicks | Link clicks | Traffic efficiency | Cost scope must include or exclude creator fees explicitly |
| Cost per landing-page visit | Eligible cost / landing-page sessions | Tracked sessions | True arrival efficiency | Tracking loss raises apparent cost |
| Lead or micro-conversion actions | Count by named action | None | Consideration progress | Do not mix sign-ups, installs and leads without labels |
| Step conversion rate | Completed next step / entrants to step × 100 | Entrants to the named step | Funnel diagnosis | State the exact step and population |
Conversion, revenue and economics
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Conversions, orders or leads | Count of named outcomes | None | Business outcome volume | Define deduplication and qualification |
| Conversion rate | Conversions / eligible visits, clicks or users × 100 | Named eligible base | Outcome efficiency | Different denominators answer different questions |
| Cost per action or acquisition (CPA) | Eligible cost / named conversions | Named conversions | Acquisition efficiency | CPA is not always full CAC |
| Customer acquisition cost (CAC) | Fully scoped acquisition cost / new customers | New customers | Unit economics | Scope shared costs and new-customer logic |
| Revenue | Currency total | None | Commercial scale | Separate gross, net, returned and tax treatment |
| Attributed revenue | Revenue credited under model | None | Channel optimization | Model credit is not causal lift |
| Incremental revenue | Treatment revenue minus counterfactual revenue | Control or modeled counterfactual | Causal economic impact | Requires adequate design, power and assumptions |
| Average order value (AOV) | Revenue / orders | Orders | Basket quality | Returns and discounts change realized value |
| Units sold | Count | None | Product movement | Stock and seasonality confound results |
| Promo-code usage | Orders or uses with code | None | Trackable response | Codes leak and miss un-coded demand |
| Affiliate sales | Approved sales credited to affiliate link or code | None | Performance compensation | Apply validation, returns and attribution rules |
| Return on ad spend (ROAS) | Attributed revenue / ad spend | Ad spend | Media revenue efficiency | Not profit and often excludes creator or production cost |
| Return on investment (ROI) | (Return minus total investment) / total investment × 100 | Total investment | Profit-oriented return | Define return, margin and all included costs |
| New-customer rate | New customers / customers acquired × 100 | Customers acquired | Acquisition quality | Identity matching can be incomplete |
| Contribution margin return | Incremental contribution margin / total programme cost | Total programme cost | Profit quality | Needs product-level margin and incremental volume |
| Customer lifetime value (CLV or LTV) | Expected discounted contribution over customer lifetime | Customers or cohorts | Long-term acquisition value | Model-sensitive and unreliable with immature cohorts |
| LTV to CAC ratio | LTV / CAC | CAC under same cohort scope | Long-term unit economics | Only useful when LTV and CAC scopes align |
| Assisted conversions | Conversions receiving non-final touchpoint credit | Conversions under model | Upper-funnel contribution | Depends on model and observable paths |
| View-through conversions | Conversions after view without eligible click | Exposed users under window | Post-view contribution signal | Highly sensitive to window and counterfactual demand |
Paid amplification and media efficiency
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| CPM | Media cost / impressions × 1,000 | Impressions | Paid exposure efficiency | Not comparable if impression quality differs |
| Effective CPM (eCPM) | Total eligible cost / total impressions × 1,000 | Impressions | Normalize blended cost | State whether creator, production and rights costs are included |
| Cost per view (CPV) | Media cost / named views | Views under named threshold | Video efficiency | View definition must match |
| Cost per completed view | Media cost / completed views | Completed views | Deep video efficiency | Completion can favor shorter creative |
| Cost per engagement (CPE) | Eligible cost / included engagements | Included engagements | Interaction efficiency | Publish action set and cost scope |
| Paid CPC | Media cost / link clicks | Link clicks | Paid traffic efficiency | Use landing visits for a stricter quality view |
| Paid CPA | Media cost / conversions | Conversions | Paid acquisition efficiency | Attribution model and window drive the result |
| Paid amplification lift | Amplified result minus organic baseline | Comparable organic baseline | Value of amplification | Selection bias if only best posts are boosted |
| Creator-handle ad performance | Standard paid KPIs segmented by identity | Named paid population | Whitelisting or allowlisting evaluation | Compare like for like creative, audience and spend |
| Organic versus paid split | Organic result and paid result reported separately | None | Channel clarity | Avoid blended totals that hide media contribution |
| Earned media value estimate | Documented rate card or model × eligible outcomes | Model-defined outcomes | Directional media-equivalent comparison | Not revenue, profit or universal value |
Brand impact and incrementality
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Aided awareness | Respondents recognizing brand when prompted / valid respondents × 100 | Valid respondents | Brand recognition | Sample and question wording matter |
| Unaided awareness | Respondents naming brand unprompted / valid respondents × 100 | Valid respondents | Mental availability | Category and coding rules affect results |
| Ad or brand recall | Respondents recalling exposure / valid respondents × 100 | Valid respondents | Memory effect | Self-report and survey design limitations |
| Consideration | Respondents considering brand / valid respondents × 100 | Valid respondents | Mid-funnel movement | Intent is not behavior |
| Preference or favorability | Favorable respondents / valid respondents × 100 | Valid respondents | Attitudinal movement | Scale and baseline must be stable |
| Purchase intent | High-intent respondents / valid respondents × 100 | Valid respondents | Directional demand | Does not equal purchase |
| Message association | Respondents linking brand to target message / valid respondents × 100 | Valid respondents | Communication takeout | Avoid leading questions |
| Search lift | (Treatment searches minus expected/control searches) / expected/control × 100 | Control or baseline searches | Interest generated beyond clicks | Seasonality and other media need control |
| Brand lift | Treatment metric minus control metric | Control group metric | Causal brand effect | Needs adequate sample and confidence interval |
| Conversion or sales lift | Treatment outcome minus control outcome | Control outcome | Causal business effect | Check contamination, power and test overlap |
| Cost per incremental outcome | Incremental programme cost / incremental outcomes | Incremental outcomes | Causal efficiency | Unstable when lift estimate is imprecise |
Product seeding and gifting
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Creators contacted or invited | Count | None | Top of seeding pipeline | Remove duplicates and invalid contacts |
| Opt-in or acceptance rate | Accepted invitations / eligible invitations × 100 | Eligible delivered invitations | Offer and fit quality | Define eligibility and response window |
| Products shipped | Count of shipments or units | None | Operational volume | Shipments and units are different counting units |
| Delivery success rate | Successfully delivered shipments / shipped shipments × 100 | Shipped shipments | Logistics quality | A delivered parcel is not creator activation |
| Eligible recipients | Delivered recipients meeting activation rules | None | Activation denominator | Document exclusions and eligibility window |
| Activation or posting rate | Creators who publish / eligible recipients × 100 | Eligible recipients within stated window | Realized organic activation | No universal benchmark; gifting without obligation cannot promise posts |
| Potential activation rate | Creators signaling intent / eligible recipients × 100 | Eligible recipients | Pipeline forecasting | Keep separate from realized publication |
| Unique asset yield | Unique usable assets / eligible recipients or shipments | Recipients or shipments, named | Content production yield | Define uniqueness and usability |
| Placement yield | Placement instances / eligible recipients or shipments | Recipients or shipments, named | Distribution yield | One asset may create several placements |
| Cost per activated creator | Total seeding cost / activated creators | Activated creators | Activation efficiency | Include product, logistics and operations consistently |
| Cost per published asset | Total seeding cost / published assets | Published assets | Output efficiency | Published does not mean usable or rights-cleared |
| Cost per usable asset | Total seeding cost / approved usable assets | Approved usable assets | Content value efficiency | Define approval and rights criteria |
| Time to post | Publication timestamp minus delivery or activation date | Published creators | Momentum and planning | Start event must be consistent |
| Non-activation or product-waste rate | Eligible recipients without target activation / eligible recipients × 100 | Eligible recipients | Waste and fit diagnosis | Non-posting is not automatically waste if insight or relationship value exists |
| Rights-eligible asset rate | Assets meeting rights conditions / reviewed assets × 100 | Reviewed assets | Reuse potential | Eligibility is not actual utilization |
| Seeding-attributed traffic or sales | Tracked sessions or sales under declared rule | None | Observable downstream action | Organic seeding often has incomplete tracking |
| Seeding ROI | (Defensible return minus total seeding cost) / total seeding cost × 100 | Total seeding cost | Economic evaluation | Only valid when return and costs are defensibly defined |
Discuss your measurement system
Programme operations and creator relationships
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Outreach response rate | Responses / delivered outreach messages × 100 | Delivered outreach messages | Outreach relevance | Automated replies and duplicates need rules |
| Creator acceptance rate | Accepted offers / eligible offers × 100 | Eligible offers | Offer competitiveness and fit | Separate interest from signed agreement |
| Shortlist-to-contract rate | Contracted creators / shortlisted creators × 100 | Shortlisted creators | Selection efficiency | A low rate may reflect deliberate optionality |
| Contracting cycle time | Contract signature date minus offer date | Signed contracts | Planning risk | Segment by market and complexity |
| Brief acceptance rate | Briefs accepted without material objection / briefs issued × 100 | Briefs issued | Brief fit | Acceptance does not prove comprehension |
| Brief comprehension score | Checklist or review score | Reviewed briefs or concepts | Reduce avoidable rework | Use a stable rubric, not intuition |
| Revision rounds | Average approved revision rounds | Approved deliverables | Process friction | Creative complexity changes the baseline |
| Rework rate | Deliverables needing material rework / reviewed deliverables × 100 | Reviewed deliverables | Brief and quality diagnosis | Define material rework |
| Approval turnaround time | Approval timestamp minus submission timestamp | Reviewed submissions | Brand-side responsiveness | Pause clocks for creator-side changes |
| Creator retention rate | Creators retained / creators eligible for renewal × 100 | Creators eligible for renewal | Relationship health | Eligibility must reflect strategy changes |
| Repeat-collaboration rate | Creators with another collaboration / eligible prior creators × 100 | Eligible prior creators | Partnership durability | Repeat work is not automatically superior performance |
| Creator satisfaction | Survey score or favorable responses / valid responses | Valid survey responses | Partner experience | Response bias and anonymity matter |
| Brand-safety incident rate | Incidents / active creators or placements × 100 | Active creators or placements, named | Risk control | Define severity and exposure consistently |
| Disclosure compliance rate | Compliant commercial posts / reviewed commercial posts × 100 | Reviewed commercial posts | Legal and ethical execution | Apply market-specific rules and review protocol |
| Rights secured rate | Assets with required rights / assets requiring rights × 100 | Assets requiring rights | Reuse readiness | Rights scope and duration must match intended use |
| Budget pacing | Actual spend to date / planned spend to date × 100 | Planned spend to date | Delivery control | Timing shifts can be planned, not failures |
| Cost variance | (Actual cost minus budgeted cost) / budgeted cost × 100 | Budgeted cost | Budget control | Explain scope changes and currency effects |
| Creator fee mix | Fee category / total creator compensation × 100 | Total creator compensation | Commercial model visibility | Separate cash, product, affiliate and rights |
Content asset and rights value
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Unique usable assets | Deduplicated approved assets | None | Content inventory | Define uniqueness, quality and rights |
| Placement instances | Count of published placements | None | Distribution volume | Do not treat placements as unique assets |
| Asset approval rate | Approved assets / submitted assets × 100 | Submitted assets | Creative quality and brief fit | Approval criteria must be stable |
| Asset acceptance rate | Accepted usable assets / reviewed assets × 100 | Reviewed assets | Portfolio quality | Separate publication approval from reuse approval |
| Content reuse rate | Assets reused at least once / rights-eligible assets × 100 | Rights-eligible assets | Realized content value | Rights eligibility is not reuse |
| Reuse placement count | Paid, owned, retail, CRM or ecommerce reuse instances | None | Cross-channel utility | One asset can have many placements |
| Content lifespan | Last material result date minus publication date | Measured assets | Long-tail value | Define material result threshold |
| Long-tail performance share | Results after initial window / lifetime results × 100 | Lifetime results | Evergreen contribution | Window choice drives the result |
| Cost per usable asset | Total content-related cost / usable assets | Usable assets | Production efficiency | State whether creator fees and rights are included |
| Cost per reused asset | Total content-related cost / assets actually reused | Reused assets | Realized reuse efficiency | Can overreward repeated use without performance context |
| Production substitution value | Avoided comparable production cost under documented assumptions | Comparable production scope | Scenario planning | Analytical estimate, not guaranteed savings |
| Rights duration coverage | Rights days covering planned use / planned use days × 100 | Planned use days | Rights adequacy | Territory, channel and edit rights also matter |
| Rights utilization rate | Rights-cleared assets used / rights-cleared assets × 100 | Rights-cleared assets | Value captured from licensing | Unused rights may indicate overbuying or changed plans |
| Creator content versus other creative | Difference in matched KPI under like-for-like test | Matched creative cells | Creative effectiveness | Match audience, placement, spend, timing and objective |
Community, advocacy and long-term value
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Repeat engagement rate | Users engaging in multiple periods / identified engagers × 100 | Identified engagers | Relationship depth | Identity resolution is partial on platforms |
| Community retention | Members active at end / eligible members at start × 100 | Eligible starting members | Ongoing participation | Define activity and cohort |
| Advocacy actions | Qualified referrals, recommendations or defenses | None | Active support | Qualitative coding may be required |
| Referral rate | Referral actions or referred customers / eligible members × 100 | Eligible members or customers | Word-of-mouth contribution | Codes and links miss dark social |
| Topic association | Target-theme mentions / coded brand mentions × 100 | Coded brand mentions | Cultural positioning | Query and coding choices matter |
| Ambassador retention | Ambassadors renewed / eligible ambassadors × 100 | Eligible ambassadors | Programme durability | Renewal depends on strategy and budget |
| Ambassador repeat performance | Repeat-period KPI versus prior or matched baseline | Comparable periods | Consistency over time | Control for format, audience growth and paid support |
| Community growth quality | Qualified retained new members / new members × 100 | New members | Healthy growth | Quality criteria must be explicit |
| Event registrations | Count | None | Event demand | Registrations are not attendance |
| Event show-up rate | Attendees / valid registrations × 100 | Valid registrations | Attendance quality | Account for cancellations and capacity |
| Event sales or ticket value | Net ticket or event-attributed revenue | None | Event economics | Separate sponsorship, ticket and downstream sales |
| Sponsor value | Contract value plus documented delivered benefits | Contracted benefits | Partnership delivery | Do not replace actual revenue with media equivalence |
| Post-event content value | Usable assets and measured outcomes after event | Named asset/result base | Extended event value | Separate assets, placements and valuation estimates |
Measurement and data quality
| Indicative KPI | Formula or unit | Denominator | Best use | Main caveat |
|---|---|---|---|---|
| Tagged-link coverage | Eligible links with valid tags / eligible links × 100 | Eligible links | Traffic traceability | Tags do not solve identity or consent loss |
| Promo-code coverage | Eligible creators with unique codes / eligible creators × 100 | Eligible creators | Sales traceability | Codes leak and may not be used by influenced buyers |
| Pixel or server-side tracking coverage | Tracked eligible events / eligible events × 100 | Eligible events | Technical completeness | Coverage does not guarantee correctness or consent |
| Creator reporting completion | Complete reports received / reports due × 100 | Reports due | Data availability | Validate screenshots, periods and definitions |
| Data latency | Availability timestamp minus event timestamp | Measured events or sources | Reporting freshness | Fast data can still be incomplete |
| Data freshness | Current time minus latest complete data timestamp | Latest complete timestamp | Decision readiness | Do not mix preliminary and finalized data |
| Attribution window | Declared number of eligible days or hours | None | Interpret attributed outcomes | Longer windows capture more background demand |
| Attribution model recorded | Yes/no plus model name | Reports or campaigns | Reproducibility | A model allocates credit; it does not prove causality |
| Unmatched or unattributed conversion share | Unmatched conversions / total conversions × 100 | Total conversions | Measurement gap | Unattributed does not mean uninfluenced |
| Invalid traffic rate | Invalid interactions / measured interactions × 100 | Measured interactions | Traffic integrity | Detection methods differ and have false positives |
| Suspicious audience signal rate | Flagged accounts or actions / audited base × 100 | Audited accounts or actions | Creator screening risk | A flag is not proof of fraud |
| Deduplicated cross-platform reach | Estimated unique people across channels | Unified audience universe | Total campaign reach | Model assumptions and privacy limits must be disclosed |
| Cross-platform overlap rate | People reached on multiple platforms / deduplicated reached people × 100 | Deduplicated reached people | Frequency and channel planning | Requires a credible identity or panel method |
| Confidence interval | Estimate ± critical value × standard error | Valid sample | Uncertainty communication | Not a quality stamp if sampling is biased |
| Statistical power or minimum detectable effect | Design-specific calculation | Planned sample and variance | Test feasibility | Underpowered null results are inconclusive |
Discuss your measurement system
Metrics that should not be confused
Reach vs impressions
Reach counts estimated unique exposure; impressions count displays. Their ratio gives frequency only when both metrics cover the same universe.
Engagement-rate variants
Rates by followers, reach, impressions or views are not interchangeable. Always publish the numerator, denominator and included actions.
Views, qualified views and completions
A start, 2-second or 6-second qualified view and completion represent different attention depths. Name the threshold and platform.
ROI vs ROAS
ROAS divides attributed revenue by media spend. ROI subtracts investment from return, then divides by total investment.
Attributed vs incremental revenue
Attribution distributes credit. Incrementality compares outcomes with a credible no-campaign counterfactual.
CPM vs eCPM
CPM is media cost per thousand impressions. eCPM normalizes a broader or blended cost scope that must be disclosed.
Seeding activation vs content yield
Activation is the share of eligible creators who publish. Yield counts unique assets or placement instances produced.
Unique assets vs placements
An asset is a distinct creative item. A placement is a distribution instance. One asset can produce several placements.
Organic vs paid
Separate natural creator distribution from paid amplification, then provide a combined total only as a secondary view.
EMV vs revenue
A media-value equivalent is a methodology-dependent estimate, not cash revenue or profit.
Build a decision-ready scorecard
- Write the objective and the decision the report must enable.
- Select one primary KPI and two to four diagnostic KPIs.
- Freeze formulas, denominators, units, sources and windows before launch.
- Separate organic, paid, attributed and incremental outcomes.
- Add data quality, compliance, rights and operations.
- Segment by creator, format, platform, market and period without inventing a universal benchmark.
Connect the scorecard to your influencer marketing strategy, maturity model, pricing model and seeding programme.
Main sources and method
Definitions were checked on 15 October 2026. Platform metrics remain time-sensitive and should be revalidated when reporting.
Influencer marketing KPI FAQ
What is the most important influencer marketing KPI?
There is no single most important KPI. The primary KPI depends on the business objective and must be reported with its denominator, data source, measurement window and attribution rule.
How do you calculate engagement rate?
Divide the defined engagements by followers, reach, impressions or views, then multiply by 100. Always name the denominator because each version answers a different question.
What is the difference between reach and impressions?
Reach estimates unique people or accounts exposed. Impressions count every display, including repeat displays to the same person. Frequency is impressions divided by reach.
Are ROI and ROAS the same?
No. ROAS compares attributed revenue with advertising spend. ROI compares net return, after subtracting investment, with total investment. ROI therefore needs a broader and explicit cost and value scope.
How should seeding activation rate be measured?
Divide creators who published by eligible recipients within a declared window. Keep this rate separate from unique asset yield and placement instances.
Is attributed revenue incremental revenue?
No. Attribution allocates credit under a model and window. Incrementality estimates what would not have happened without the campaign using a control group, holdout or credible causal model.
Is EMV actual revenue?
No. EMV is a methodology-dependent media-value estimate. It must not be presented as cash revenue, profit or causal ROI.
Build your scorecard with TANKE
TANKE can turn your objectives, available data and market constraints into a decision-ready measurement system before the next brief.
