Krista Hooper, the creator behind @walmartfinds, curates her favorite picks from Walmart. With nearly 400,000 Instagram followers, she has built a macro-tier account and strategically incorporates paid media into her business.
According to impact.com’s internal platform data, she drives more value than most creators with ten times her following. Krista built a successful business through her content, and paid media helped her amplify that growth. That is the first lesson in how to choose the right influencer for your brand.
The brands still building shortlists around follower count are the ones who keep getting burned. The creators actually converting are the ones who know how to turn any size audience into sales.
“Don’t confuse followers with reach,” says Jeff Melton, Senior Director of Global Creator Strategy at impact.com. “On platforms built around a content graph, algorithms decide who sees your content. Having followers doesn’t give creators control over reaching them organically, and paid media can buy distribution regardless of follower count.”
Melton has seen a creator with just 500 followers put a few hundred dollars of paid media behind a post and drive meaningful business. For him, the question is what a creator can deliver against a brand’s objectives.
You’ve probably run into these objections already. Fake followers or a creator who never converted. Spend nobody can defend after the fact. The first two take a bit of homework. You can pull the media kit, or check the conversion history. You should always screen for bots before you commit budget. The third takes a habit: log every test so the spend is never a bet you can’t explain later.
Once you know who to select, pay is the next decision—start with how brands pay creators.
The follower-count blind spot: why size and sales don’t always correlate
One pattern holds: audience size and sales don’t move together, so reach tells you little about what a creator will sell.
According to impact.com’s “The Global State of Affiliate Marketing in 2025” report, 59% of brands plan to dedicate a quarter or more of their affiliate budget to creator partnerships. Of that group, 18% plan to dedicate more than half their budget to creators in the coming year. Every dollar that shifts toward creators raises the cost of picking them on reach, and most programs are about to shift more of them.
Melton says, “If you look at our internal data, follower count is a weak predictor of sales. Two creators with the same follower count can have very different results, so it shouldn’t be used alone to shortlist creators.”
If followers and sales do not move together, follower count measures something other than sales.
Paid media proficiency is a key indicator of sales ability
Most campaigns’ largest creators use paid media to amplify their existing content and drive more conversions. For selection, ask how a creator gets people to buy. Paid media is one answer.
A creator who spends $200 boosting a post and drives 10 more sales earns a $1,000 bonus and nets $800. Evaluate creators on that return profile rather than on follower count.
Reach can also fail from the other direction, and Melton has watched it happen up close. While consulting for a brand years ago, he saw it pay a well-known creator with millions of followers a six-figure fee for a single post. This is the kind of placement that looks like a guaranteed win on paper. The post drove virtually no sales.
The brand had expected a hit. Its own verdict afterward was that the post was the biggest waste of money it had ever spent.
| Reach vs results: two creators, two outcomes | ||
|---|---|---|
| Approach | Well-known creator | Krista (Walmart Finds) |
| Follower scale | Millions of followers | Nearly 400K Instagram followers (macro tier) |
| Result | Virtually no sales | Outearning creators with larger followings |
| Paid media behind the content | None, relied on organic reach of a single post | Used on some of her content |
| Brand spend | Six-figure fee for one post | Ongoing, incremental ad spend |
“If you look at our internal data, follower count is a weak predictor of sales. Two creators with the same follower count can have very different results, so it shouldn’t be used alone to shortlist creators.”
– Jeff Melton, Senior Director of Global Creator Strategy at impact.c
How to choose the right influencer for your brand: media kits, track record, and tested results
If audience size cannot predict sales, a creator’s sales record has to, and the media kit is where creators keep theirs.
Throw out the demographic assumptions
Throw out what you thought you knew about a creator’s audience. Brands have historically been picky about demographic and psychographic fit, down to requests for moms on Long Island aged 25 to 44. The question to ask first is why that fit matters for sales.
Demographic fit still counts, but it is not the first filter. If a cowboy in Texas loves your product and can sell a lot of it, there is no reason to exclude him.
Use fit to choose between creators who’ve already proven they can sell. Avoid screening anyone out before you’ve tested them.
Treat the media kit as a claim, not a guarantee
Ask every creator you consider for a media kit. Creators increasingly add their ability to drive sales to those kits. A creator who has worked with many brands will pitch a new one on stats from the earlier brands.
Read each stat for the brand and period it covers, then ask for the same measure for your brand. Results for one brand rarely carry over to another, which is why the test is needed.
A media kit can’t be independently verified because creators have always been able to overstate figures like engagement rate. Use your judgement rather than taking every number at face value.
The stronger move is to put the task on the creator to prove the same result for your brand. Every stat in a pitch should clear that same bar before it earns a place in your shortlist.
Screen creators twice before you test
Screen every creator twice before you test.
- Brand-safety screenings check what a creator already posts for politics, profanity, or adult topics that would sit badly beside your brand.
- Authenticity screenings check who follows, looking for fake followers and bought engagement, and the table below names one telltale sign.
Melton’s bar for testing is deliberately short: “I think by far the most important piece is as long as they’re brand safe, and they don’t look like they buy followers or engagements, then just test with them.” Nothing past those two checks has to be settled before a creator earns a test—that’s what the test itself is for. Keep engagement rate as a secondary signal.
The table below sets each familiar signal beside the reason it falls short and the evidence to ask for in its place.
| Vanity metric vs conversion signal | ||
|---|---|---|
| Old signal | Why it fails | Check this instead |
| Follower count | Doesn’t link to sales. Top-performing creators on impact.com’s platform have driven eight-figure GMV without large followings. | Documented conversion history and revenue behavior in the creator’s media kit |
| Demographic and psychographic fit | Demographic fit excludes creators who convert but don’t match the audience profile a brand imagined. | Actual sales and content-to-conversion performance, regardless of audience profile |
| Engagement rate alone | Purchase-intent clicks don’t always register as likes, comments, or shares | Engagement rate as one input among several, with conversion data making the call |
| Follower authenticity assumptions | Bots and bought followers can sit behind a normal-looking profile. | Screening for bot signals, such as generic emoji comment patterns and unnatural engagement spikes |
When a media kit arrives, hold each claim in it to the same standard and ask for the evidence before the claim earns a place in your test pool. Once a creator clears both screens, the question changes from who looks right to who converts, and no marketplace can answer that for you.
“I think by far the most important piece is as long as they’re brand safe, and they don’t look like they buy followers or engagements, then just test with them.”
— Jeff Melton, Senior Director of Global Creator Strategy at impact.com
You can’t search a marketplace by sales, so test at scale to find out who converts
A wide pool is how you find out who converts, because follower count says little about sales and sales data isn’t searchable before you spend.
Look for context clues, then test wide
Sales data stays out of reach in a creator marketplace. The search has to start with other signals.
To spot a likely mogul, look for clues you can read without sales data.
- Affiliate links in their content.
- A link in bio that looks legitimate.
- A storefront that looks legitimate
- Use of affiliate tools like LinkDM
Together, those signal a creator who knows how to sell.
Context clues only qualify a creator for a test. Invite a broad pool to test with your brand, then track who converts rather than judging fit by profile alone. The pool should run to hundreds or thousands of nano-creators and micro-creators rather than a handful of expensive names with millions of followers.
Testing at scale spends budget on creators who will not convert, which is the price of finding the ones who will. A test also catches the creator whose media kit looks strong and whose results on your offer do not. Set the test window and the conversion bar before it starts, so the same numbers are used to assess every creator.
In scaled creator communities, Melton has found that about 10% to 15% of creators drive about 80% of revenue. Confirm the ratio against your own program before you plan around it. If it holds, most of your test pool will account for a small share of revenue. You can’t tell in advance which creators belong to the group that drives most of it, so the width of the pool decides what you find.
Turn your test results into a leadership-ready record
A leadership-ready record is a log of every creator you tested and how each one performed. Keep five things in it.
- Log every creator you test, along with the context clues that qualify each one (legitimate affiliate links, a real storefront, a credible media kit claim).
- Track actual conversion data for each creator during the test window, in place of engagement or reach.
- Flag which creators cleared your conversion bar and which did not, with numbers beside each name.
- Record what every converting creator delivered and why the result looks repeatable.
- Present the full process to leadership, including the creators that did not convert, because a system that finds converters outweighs any single result (and is a stronger case for continued spend).
The same report found that only 20% of brands track customer acquisition cost (CAC) and 18% track average order value (AOV). The figures don’t say anything about creator performance, but they show that most brands don’t track the metrics a leadership conversation needs. A sales count alone leaves a finance lead asking what each customer cost and what each order was worth, so put CAC and AOV beside each creator’s sales count.
In-house vs. managed services for influencer programs: How to decide what yours actually needs
Deciding who owns which part of the creator program matters more as you expand. This guide outlines best practices and draws on expert insights from impact.com’s creator services team.
Diagnosing “it didn’t convert”: often it’s the format, not the creator
A weak result often traces back to the format or the timeline instead of the creator. Choosing on reach is the first mistake, and misreading the result is the second.
The format problem
Take the TikTok link in bio. A viewer has to open the profile, tap the link in bio, and then reach the storefront, which adds multiple clicks before checkout. A link sticker on a story puts the link right in front of the viewer instead.
Some brands running 8 months to a year of weak results never considered the format—a blind spot that’s invisible from inside a program but obvious the moment someone checks the click path. impact.com’s creator services team runs every audit against a full rubric, and a common finding is that formats place the link several clicks away.
The timeline problem
The other misdiagnosis is time. Creator marketing behaves like paid social and paid search, channels that brands keep testing and tuning for years, and it has its own levers to adjust.
Testing also costs budget, and by Melton’s estimate, you may need 8 to 12 months to break even. That timeline is the same as the one brands already use for paid social and paid search, and creator marketing is no exception.
To tell the two apart, count the click path before you count the months.
- A long-click path is a formatting problem at any point in the timeline.
- A clean path with results still short of the 8 to 12-months is a timeline problem.
Choosing on follower count and running a bad format are two separate problems, and a weak early result makes them look like one. Retest with the link in plain view before you conclude that the creator, or the channel, does not work. Keep the test open through the break-even window.
FAQs
You can spot fake followers by reading a creator’s comments, because bots and bought followers leave telltale patterns. One pattern is a string of comments that read “Cool, cool, cool, love, cool.” Brand-safety screening is a separate filter that looks at what a creator posts, while fake-follower screening looks at who follows. Treat the follower screen as a check on honesty, since a clean comment section tells you the audience looks real and only a test tells you if it buys.
Instead of follower count, choose creators by what they have sold before and by how they perform in a test on your own offer. Ask for a media kit with stats from other brands, and treat it as a claim to test. Compare a creator’s engagement rate with the typical rate for creators of a similar size, because a poor rate for that size can be a warning sign. Keep engagement as a secondary signal, because conversion data should make the final call.
A creator campaign that doesn’t convert is often misdiagnosed. Two problems get mistaken for a bad creator: a high-friction format, such as a TikTok link in bio that adds several clicks before checkout, and an unrealistic timeline. Creator marketing behaves like paid social or paid search. Count the click path first. If it’s clean and the test is still inside that window, the channel needs more time, not a different creator.
To justify influencer marketing spend to leadership, bring a documented test instead of a bet on one creator’s follower count. Show every creator you tested with the conversion numbers beside each name, and show the process that found them, including the test window and the conversion bar you set. Set the timeline up front as well. Once your team hears that break-even can take 8 to 12 months, your day-to-day contact can brief a manager, who can brief the next manager up.
Build your selection process around conversion, not followers
Winning programs 12 months from now will be the ones that already know which creators convert and are working to keep them. If the 10% to 15% ratio holds in your program, that retained group is most of the job.
Melton has seen this play out directly. One global consumer brand he worked with asked what success actually looked like for its program. Revenue and return on ad spend (ROAS) mattered, but what the brand actually needed was a healthy community: creators who kept coming back to post and engage, not just a bigger following. That answer was enough to make the brand throw out its entire pyramid of mega, macro, and micro creators, a structure it now considers irrelevant to how creator marketing actually works.
Its replacement segments the community into champions and the rest of the creator base. In your own program, the equivalent is sorting creators by who converts and who keeps returning, with follower tier left out of the columns.
Begin with three moves today:
- Audit your program against documented conversion data, and count how many creators you chose on follower size and how many have a conversion record.
- Ask every creator under consideration for a media kit with real campaign history, naming which brands and over what period, and treat the stats in it as claims to test.
- Build a wide net to test creators, and let the data pick your champions.
Follower counts mattered more in the early days of creator marketing and matter far less now. The biggest outdated assumption still costing brands money is paying based on followers.
Discover more content below:
- How do brands pay influencers? A decision tree for matching compensation to goals [blog]
- Fewer sales, bigger stakes: First half of 2026 data shows how brands captured a pickier shopper [research]
- In-house vs. managed services for influencer programs: How to decide what yours actually needs [blog]
- The creator mistake that pays well and costs everything
- The Partnership Economy podcast: AI, Creators, and What Comes Next
