Measurement guides
Influencer marketing metrics: engagement, reach and campaign cost
Choose the campaign question first, then the metric. Views, unique reach, engagement and sales describe different things. Record the source, reporting period and calculation so another person can understand what the result does and does not show.
Explore the guideStart with this
Write the metric definition beside the number. Views, unique viewers and impressions answer different questions.
Compare the same engagement actions, costs and reporting windows. A percentage alone is not enough.
Use campaign totals to calculate campaign rates. Keep missing data visible instead of turning it into zero.
Choose the question before the metric
For an awareness campaign, you might ask whether the content reached enough of the intended audience to justify a follow-up. For a product demonstration, you might ask whether people understood the use case and raised useful questions. For a content-production project, the first question may be whether the agreed assets were delivered and can be used as intended.
Choose one main outcome and a small set of supporting measures before agreeing the work. Record where each figure will come from and when you will collect it. This keeps a campaign from becoming a search for whichever number happens to look strongest afterwards. For sales and revenue, use evidence from your brand's measurement system and explain how it links transactions to the campaign.
| Campaign question | Evidence to collect | What it cannot establish alone |
|---|---|---|
| Did people encounter the content? | Dated views; creator-supplied reach where available | That viewers noticed the product or remembered it |
| Did the content prompt a response? | Defined engagement actions and relevant comments | That the response was positive or led to a purchase |
| What did the response cost? | Matching spend, views and engagement counts | Profitability without verified revenue and costs |
| Did we receive useful content? | Delivered assets checked against the agreed brief and rights | How those assets will perform when the brand distributes them |
Keep views, impressions and people separate
A view is a platform-defined viewing event. A count of views should not be relabelled as a count of individual people. YouTube, for example, reports unique viewers as a separate estimated measure. Its thumbnail impressions also have a specific meaning: they concern eligible displays of a video thumbnail, rather than every possible encounter with a video.
Use the name on the source report. If you add views across posts, label the result total post views. Someone may watch more than one post or watch repeatedly, so adding those counts does not give you a deduplicated audience. Keep paid distribution separate where the source allows, and note when it does not.
Calculate engagement rate with a stated denominator
For the worked examples in this guide, engagement means likes + comments + shares + saves, where those figures are available. Engagement rate by views is those recorded actions divided by views, multiplied by 100. This is a count of actions relative to views; it is not the percentage of unique viewers who interacted.
You may encounter rates based on followers, reach or impressions instead. Those formulas can be useful for different questions, but their results are not interchangeable. Name the denominator, the included actions and any unavailable components. Compare like-for-like data, or explain why the comparison is incomplete.
Illustrative worked example
A post with 20,000 views
Suppose a post records 600 likes, 80 comments, 100 shares and 20 saves at the same reporting point. That is 800 recorded engagement actions.
Engagement rate by views = 800 ÷ 20,000 × 100 = 4%. If shares and saves were unavailable, the defensible result would be 680 known actions and a 3.4% likes-and-comments rate by views. It would not be a complete four-action engagement rate.
Give every cost metric a clear scope
CPM usually describes cost per thousand impressions in advertising. For a creator-post comparison based on views, write cost per 1,000 views explicitly. The arithmetic is similar, but the denominator is different. CreatorCall's existing CPM calculator uses views; it can help with that calculation when you supply the spend and view count.
Choose the cost basis before comparing creators. Creator fees alone may answer a fee-efficiency question. A fuller campaign view might also include sourcing, production, products, shipping and agreed usage rights. Keep the same basis across rows, and state whether paid distribution is included. Do not divide the cost of one subset of posts by the results of a larger set.
| Metric | Calculation | Useful interpretation |
|---|---|---|
| Cost per 1,000 views | Included spend ÷ recorded views × 1,000 | Cost relative to the amount of viewing recorded |
| Cost per engagement | Included spend ÷ included engagement actions | Cost relative to the response actions you defined |
| Cost per comment | Included spend ÷ recorded comments | Cost relative to comment volume, not comment quality |
Illustrative worked example
One spend figure, three different questions
Using an agreed cost basis of $600, a post with 20,000 views, 800 engagement actions and 80 comments has a cost per 1,000 views of $30, a cost per engagement of $0.75 and a cost per comment of $7.50.
None of those figures is a profit measure. A product-availability question, a complaint and an enthusiastic recommendation all count as comments, yet may call for different decisions.
Compare creators without averaging the wrong thing
Compare similar deliverables at a similar age after publication. A short product demo and a long tutorial have different jobs; a post observed after one day has had less time to accumulate results than one observed after a month. Record these differences before deciding that one creator performed better.
For a campaign engagement rate, add the included engagement actions and divide by the corresponding views. Averaging creators' percentages gives a different result because it treats a small post and a large post as equally weighted. Neither figure should be presented without saying how it was calculated.
Illustrative comparison · same actions and reporting window
The higher rate does not necessarily provide more response
Creator A has the higher rate; Creator B supplies more recorded actions. The campaign rate is 2,300 ÷ 100,000 × 100 = 2.3%. The simple average of 5% and 2% is 3.5%, which is not the campaign's total-action rate.
| Creator | Views | Engagements | Rate by views |
|---|---|---|---|
| Creator A | 10,000 | 500 | 5% |
| Creator B | 90,000 | 1,800 | 2% |
| Campaign total | 100,000 | 2,300 | 2.3% |
Turn the result into a useful next decision
Use your own comparable campaigns as a starting point for improvement, with the objective and context written down. A single universal engagement-rate target hides differences in platforms, formats, audiences, collection methods and campaign goals. A small first campaign is often most useful as a documented starting point for the next test.
Pair the numbers with a content review. Strong viewing but repeated questions about how the product works may suggest a clearer demonstration next time. An inexpensive stream of irrelevant comments is not automatically a better outcome than fewer, relevant product questions. State the observation, the plausible explanation and the next test separately.
- Keep the source and capture date with each metric.
- Write unavailable where data is missing; reserve zero for a recorded zero.
- Report exactly which posts contributed to each total or rate.
- Keep supplied private insights separate from publicly visible counts.
- Use verified sales evidence for sales conclusions; views and comments cannot supply it.