Influencer marketing metrics: how to read them before and after a campaign

A profile shows 250,000 followers and one video with 900,000 views. The brand assumes a partnership will perform similarly. Across the last ten videos, however, median views are 34,000 and most of the viral audience comes from another country.

No metric is inherently bad. The mistake is using it without context.

Audience is not follower count

Followers describe accumulated account size. They do not show how many people will see the next asset or whether those people belong to the client's market.

Before commissioning, request recent aggregate data relevant to the campaign: country, language, age where relevant and available, topic affinity, device or format when it affects conversion, and trends rather than one snapshot.

You do not need personal follower data. You need to know whether the overall composition matches the intended audience.

Reach, impressions, and views

• Reach: unique accounts or people exposed, under the platform definition.
• Impressions: times content was displayed; one person can create several.
• Views: plays meeting the platform's criterion.
• Unique viewers: estimated distinct viewers where available.

Platforms do not always count a “view” in the same way. Record the source and avoid treating unlike figures as equivalent.

Use a distribution rather than one viral asset

Request a coherent sample, such as the last ten comparable posts, excluding giveaways or unrelated viral outliers.

Review median views, normal range, sponsored versus organic results, performance by subject and format, publishing frequency, and unexplained spikes.

There is no universal “good” engagement percentage. It changes by platform, size, format, subject, country, and formula.

Useful engagement

Likes and comments help explain reaction, but they do not have equal value.

Look for comments related to the content, product questions, saves or shares where available, creator responses, repetitive generic activity, and reaction to previous partnerships.

For a consideration campaign, one specific product question may be more valuable than one hundred emojis.

Retention and integration placement

Retention shows how much of the audience remains throughout an asset.

When a sponsorship appears at minute eight in a ten-minute video, total views do not show how many people reached it. Request retention around the integration and average duration when the creator can share them.

One screenshot supports an estimate; it does not guarantee future performance.

Clicks and conversions need tracking

Before publication, prepare a UTM link, identifiable code, matching landing page, agreed attribution window, tested analytics events, and a person responsible for preserving results.

A short link can measure clicks but not prove sales. Store, CRM, or client analytics must complete the journey.

Separate the costs

• Creator fee.
• Product or travel.
• Agency commission.
• Paid media.
• Usage rights.
• Exclusivity.
• Additional production or adaptation.

Combining them prevents meaningful comparison.

Using one consistent definition, calculate cost per thousand impressions or views, cost per click, cost per registration or sale, attributed revenue, and the cost of reusable assets.

Do not turn those metrics into an automatic score. A creator can sell while harming brand perception, or produce a valuable reusable asset with few direct conversions.

Post-campaign record

Preserve the publication screenshot, date, URL, agreed reporting milestones, client clicks and conversions, relevant comments, incidents, rights, expiry, and a recommendation to repeat, adjust, or stop.

Keep brief, contract, versions, published asset, and report together. The next selection can use evidence rather than a memory that “it went well.”

Use the influencer brief for deliverables and rights. Creators presenting their own data can use the media kit guide.