How to vet an influencer before you pay them

Price on a typical post, not the best one

Vet a creator on what you are actually buying: what a typical recent post gets, who sees it, and whether that audience is real. Follower count alone tells you little.

Open their recent posts in the format you are buying, Reels for example, as many as they have posted in the last few months, and note the views on each. Take the median, the middle value, rather than the average, because one viral post can pull an average far above what a typical post gets. For example, if a creator has 80,000 followers but the median of their last 15 Reels is 3,000 views, a typical post of theirs gets about 3,000 views, and that is what you are paying for.

Then look at likes and comments against views, and compare with creators of a similar size on the same platform. HypeAuditor now measures engagement per post against views rather than followers for accounts with enough Reels, and TikTok One lets brands filter creators by median video views.

Check where the audience lives

Followers in countries you do not ship to are no use to you. Ask the creator for their own analytics for the last 30 or 90 days, ideally on a live screen share rather than a cropped screenshot. Instagram insights show the top countries, cities, age ranges and gender of the accounts a creator reached, once their posts have reached more than 100 accounts in the period; TikTok Studio shows the demographics of their viewers and followers.

Some of this is visible without asking. Brands in Instagram's creator marketplace can see a member creator's insights, such as average reach, and TikTok One shows audience age, gender and country for creators listed there. On any Instagram account, About this account shows when it was created and how many times its username has changed, and for some accounts the country it is based in.

Look for bought followers and fake engagement

Warning signs include a follower count that jumps without explanation, many followers with no profile picture, posts or bio, and comments that are mostly emoji or generic praise, which HypeAuditor says can point to engagement pods (groups that agree to engage with each other's posts) or fake activity. Read the comments on several posts: questions about the product and people describing their own experience are good signs.

Tools put a score on this. Modash works out a fake follower rate by flagging follower accounts on signals such as a missing profile picture, the following-to-followers ratio, account age, number of posts, a missing bio and unusual follower growth. HypeAuditor's Audience Quality Score combines engagement rate, the share of real followers, growth patterns and how many recent comments come from outside engagement pods, and HypeAuditor advises against choosing on the score alone.

Review their paid work

Look at what they have posted for other brands. Did sponsored posts get views close to their typical post, and were they clearly labelled? Meta's Ad Library includes a searchable database of public, active Facebook and Instagram branded content carrying the paid partnership label, created on or after 17 August 2023. You can search it by Instagram profile, and it shows the brand partner for each post. It only holds posts that carry the label, so a sponsored post without one will not appear there.

Labelling matters to you as well as to them. In the UK, the ASA and CAP say brands and influencers are both responsible for making ads obviously identifiable. In the US, the FTC expects advertisers to have reasonable programmes to train and monitor the people promoting them.

Start with one post

Before a large fee, begin with product (in the UK a post about a gifted product still needs an "Ad" label) or a single paid post, agree in writing what will be posted and when, and pay against delivery. Then judge the next deal on what that post actually did, not on the creator's best numbers.