How to Vet an Influencer's Audience Before You Pay

Followers are a cumulative record of past interest. Views measure current attention, and attention is what you are renting. Take the last ten to fifteen comparable posts from a creator, drop the single highest and lowest, and take the median of what remains. Divide that median by the follower count. That one ratio is the most informative number you can get without asking anyone for anything, and it is the number a media kit is least likely to volunteer on its own.

Baselines move by platform, format and category, so build your own rather than importing a rule from a blog post. Sample five to eight creators in the same category publishing the same format, and treat the middle of that set as normal for your niche. A creator sitting far below their peers has an audience that stopped watching, whether or not it was ever real. Then convert the ratio into something you can price: median views multiplied by your rough estimate of the share who could actually buy your software.

Some platforms do not publish view counts at all. Newsletters, private communities and professional networks show the author far more than they show you. For those, use reactions and comment volume as a weak proxy and treat the analytics conversation as mandatory rather than optional. Tiptop scores audience signals on the same 0-100 scale it uses for every other channel, so the evidence behind one creator stays comparable with the next one when you revisit the roster months later.

Open the last four or five posts and read thirty to fifty comments end to end. A healthy comment section is specific. People name the tool they are replacing, the step where their process breaks, the size of their team, or the constraint that makes the advice hard to follow. Questions ask how rather than whether. Somebody disagrees, and the creator answers with something more than gratitude. That texture is difficult to manufacture at volume, and it is the closest free proxy for trust you will find.

The unhealthy version is uniform. Short praise in the same three or four shapes, emoji strings, compliments aimed at the creator rather than the content, and remarks that would sit equally well under any post in the category. Check the handles: if the same accounts appear on every post within minutes of publication and never say anything specific, you are looking at an engagement pod rather than an audience. Check the timing too, because a burst of comments in the first ten minutes followed by silence is a distribution problem dressed as popularity.

Read the creator's own replies as well. Someone who answers implementation questions in public has an audience that keeps asking them, which is exactly the behaviour you want when your product is the subject. It also tells you what happens after your post ships, because the questions a sponsored post generates land in that same comment section and somebody has to answer them. If the creator has never replied to anything, agree in advance who monitors the thread and how product questions get handled.

Ask for a full-screen capture of the creator's analytics rather than a number in an email. The capture should show the account name, the metric selector and the date range, and it should cover the same window you asked every other candidate for, normally the last 28 or 90 days. Request per-post views for that window, the audience country breakdown, the audience language, and whatever role or industry data the platform reports. A media kit is a marketing document. The dashboard is the record.

Verify before you decide. Pick one post that appears in both the screenshot and the public feed, then compare the public like or comment count against what the screenshot claims. Ask for five minutes of screen share and have the creator change the date range while you watch. Most agree without friction, and the ones who hesitate usually explain why. Refusing a screen share is not automatic disqualification, but it does mean you fall back to public evidence and size the first deal to match what you were able to confirm.

Keep the evidence attached to the offer. Note which numbers were verified, which were taken on trust, and the date you collected them, because creator performance drifts inside a single quarter. Tiptop's Influencers workspace holds the rate card, the audience evidence and the outreach thread on one creator record for that reason. Whatever tool you use, the rule is the same: a figure without a source and a date is a figure you will have to check again before the next campaign.

Plot follower count month by month for the last two years and put median views on the same timeline. Real growth is uneven but continuous, with one or two spikes you can trace to something identifiable: a launch, a podcast appearance, a post that travelled beyond the usual audience. Ask about every visible spike. A creator who remembers which video did it and can name the week is describing something that actually happened to them rather than something that was bought on their behalf.

A step change looks different. Thousands of followers arrive within a few days, the line goes vertical, and nothing else moves. Views stay where they were. Comment volume stays where it was. That shape is the signature of purchased followers, a giveaway, or a follow-for-follow campaign, and the practical consequence is identical in all three cases: the new accounts do not watch. Treat it as purchased audience until the creator gives you an explanation that the view line actually supports.

The slower failure deserves as much attention. Followers keep climbing at a normal rate while median views decline quarter over quarter. That creator is growing on paper and shrinking in practice, usually because the platform stopped distributing their work or the audience drifted to a different topic. Price on recent views rather than accumulated followers, and ask what changed. The answer tells you whether this is a format problem they are already fixing or a slide that started a year ago.

Write down the countries where you can bill, support and legally sell before you look at anyone's audience map, then compare. A creator whose audience sits mostly in markets you do not serve is not cheap at any rate, because you are paying for reach you cannot convert. This is the most common quiet mismatch in creator deals and it is invisible on the profile: the content is in English, the creator shares your timezone, and two thirds of the viewers are somewhere you have never shipped.

Language needs its own check. Content published in English does not mean the audience reads it as a first language or searches in it. The comment section is more honest than the caption, so read what language people reply in and whether their questions suggest they followed the detail or only the headline. For B2B, extend the same scepticism to role and company size, since an audience of students and job seekers converts very differently from one of practising operators with a budget line.

Where the platform reports nothing useful, sample by hand. Open thirty to fifty followers or recent commenters, record country and job title from the public profile, and accept that the result is crude. Crude and collected beats precise and imagined. Half an hour of manual sampling has ended more bad deals than any audience analytics vendor, and the numbers are yours to defend when a colleague asks how you knew the audience was wrong before you paid.

Find the creator's last three to five paid posts. A disclosure tag, an ad label, a sponsor read or a thank-you line will identify most of them. Compare the median views and comment volume of those posts against organic posts from the same weeks. Some drop is normal and expected. A collapse means the audience has learned to skip this creator's advertising, which is precisely the thing you are about to buy, and no amount of creative work in the brief will fix it.

As a working band, sponsored posts that hold roughly 70 to 90 percent of organic views suggest an audience which tolerates the format. Below half, you are paying for a post people actively scroll past. Read the sponsored comment sections as well as the counts, because the difference between people asking what the tool does and people complaining about how many ads have run lately is the difference between a partner and a billboard you happened to rent for a week.

Count the ad load while you are there. If eight of the last twenty posts were paid, your post arrives into a stream the audience has already discounted. Check for a competitor promotion in the last six months, and for any product the creator endorsed and later criticised in public, which tells you how much of an endorsement is conviction and how much is inventory. None of this rules a creator out by itself. It changes what the slot is worth.

Booking three creators in the same narrow category often means paying three times to reach a substantially similar group of people. In B2B software niches the audiences of the two or three best-known voices can overlap heavily, because practitioners in a specific role tend to follow all of them. The symptom appears after the money is spent: the second and third posts produce a fraction of the first one's signups, and the campaign gets written off as a channel failure rather than a sequencing mistake.

Overlap can be estimated cheaply before booking. Sample twenty to forty commenters from one creator and check how many follow the others on your shortlist. Look for mutual recommendations, shared podcast appearances, collaborations, and the same handles turning up in both comment sections. None of this is precise. It is still enough to tell a shortlist where every creator sits inside one conversation from a shortlist where each of them reaches a genuinely different room.

Then design around whatever you find. Stagger bookings by three to four weeks so repeat exposure works as reinforcement rather than waste, give every creator a distinct tracking code so overlap becomes measurable instead of theoretical, and prefer a set that covers adjacent buying roles: the operator who runs the workflow, the technical evaluator, and the manager who signs. Match scoring answers which creator fits the brief. Overlap answers whether the group makes sense bought together.

Some findings adjust the price and some end the conversation. A premium rate quoted alongside a refusal to show any performance data at all. Screenshot numbers that contradict counts visible in the public feed. A follower step change the creator cannot or will not explain while the view line stays flat. Recent paid work carrying no disclosure, which shows you exactly what your own post will look like. An audience concentrated in markets you cannot sell to, priced on total followers.

Softer flags concern how the conversation runs rather than what the data says. Pressure to decide before you have finished checking, a slot another brand is supposedly about to take with no detail attached, a rate that moves between two emails, or reluctance to name the date range behind a number. Any one of these can be a busy person handling enquiries badly. Two or three together are a pattern, and a first deal with a new creator is the wrong place to absorb that risk.

Leaving is easier when it is plain and short. Say the audience data did not match the markets you sell into, thank them for their time, and keep the door open, because a creator who is wrong for this quarter may be right for the next product or the next segment. Record the reason and the date in your roster so a teammate does not restart the same conversation in six months and reach the same conclusion the slow way.