The short version: influencer marketing still works, but “post a nice photo and hope” stopped working years ago. What works now is smaller creators, longer contracts, and tracking actual sales instead of likes. I used to be paid to post content that I know for a fact never sold a single thing, so I’m not going to pretend the old model deserves defending.
Related reading: How to Measure the Success of Influencer Marketing.
What I saw from the inside, as the influencer
Between roughly 2015 and 2019 I was on the receiving end of influencer marketing budgets, not just advising on them. Brands paid me flat fees, four figures for a single Instagram post more than once, in exchange for a photo, a caption, and a hashtag. Nobody asked me for a conversion rate. Nobody checked whether the affiliate link I’d been given got clicked. The brief was reach, the invoice was reach, and the report at the end was reach.
I checked some of those affiliate links myself out of curiosity. A few converted well. Several converted at close to nothing, single-digit clicks over a month, on a post that had “reached” forty or fifty thousand people according to the platform. That gap between reach and result is the whole story of why people now ask if this stuff still works. It never worked the way the invoices implied. What’s changed is that brands finally have the analytics to notice.
So does it still work, or not
It works, but the mechanism has moved. In 2016 to 2019 the model was one big post from one big name, treated like a billboard. By 2026 the model that moves product is a smaller creator, often under 50,000 followers, who is trusted enough by their audience that a recommendation reads as a favour rather than an ad. Nielsen and various platform studies have found repeatedly that audiences rate recommendations from people they follow above traditional advertising, and that trust holds up even at small scale, sometimes better at small scale, because the audience feels the creator uses the product rather than being handed it by an agency.
The uncomfortable bit, the part most brand decks skip over, is that a huge share of influencer marketing spend still goes to the wrong size of creator for the wrong reason: because a name is recognisable to the marketing director’s teenager, not because anyone has checked whether that creator’s audience buys anything. I’ve sat in meetings where the shortlist was picked entirely on follower count with zero mention of engagement rate, past campaign performance, or audience overlap with the target customer. That’s not a strategy. That’s a guess wearing a strategy’s clothes.
The number that changes the conversation
Here’s a specific comparison worth sitting with. A mega influencer with two million followers might get an engagement rate of 1 to 2 percent on a sponsored post. A micro influencer with 10,000 to 30,000 followers in a genuine niche, running gear, home baking, small business finance, whatever it is, will often sit at 4 to 8 percent engagement, sometimes higher. Multiply that out and the smaller account is frequently putting a sponsored message in front of more attentive people per pound spent, not fewer, even though the raw reach number looks tiny by comparison.
I ran a small test for a client last year with a home and garden product. We split a modest budget across one creator with 400,000 followers and four creators averaging 15,000 followers each in the same niche. The single big name generated more impressions on paper. The four smaller creators generated more click throughs to the product page and, when we tracked it with unique discount codes, close to three times the redemptions. Same total spend, roughly. That’s the pattern I’d expect most businesses to find if they measured it instead of eyeballing follower counts.
Why the old model broke
- Audiences got wise to sponsored content and started scrolling past anything that reads as an ad, which pushed brands towards creators who blend product mentions into normal content rather than obvious posts.
- Platforms changed their algorithms so organic reach for any single post, sponsored or not, dropped hard, meaning a one-off post buys far less visibility than it did in 2018.
- Regulators in the UK and US tightened disclosure rules, so the lazy “just don’t mention it’s an ad” trick that inflated a lot of early influencer results is largely gone.
- Younger audiences, particularly on TikTok, now favour creators who show process and mistakes over creators who show a polished result, which rewards long-term partnerships over single sponsored posts.
What works right now
If you’re deciding whether to put budget into this, here’s the version I’d run today, step by step.
- Pick the audience before the creator. Define exactly who buys your product, then find who that audience already trusts, rather than starting from “who’s popular in our industry.”
- Favour smaller creators in tight niches over one big name. Three to six micro creators in a relevant niche will usually outperform one mid-tier generalist for the same total spend.
- Ask for a three to six month arrangement, not a single post. Repetition builds trust with an audience in a way a single mention never does; one post is an ad, six mentions over a season starts to look like a genuine preference.
- Give every creator a unique link or code. This is the single change that fixes most of the reporting problems. If you can’t attribute sales to a specific creator, you’re paying for a feeling, not a result.
- Set a target before you launch, not after. Decide what a good click through rate or conversion rate looks like for your product before the campaign starts, so you’re not grading it against vibes at the end.
- Feed the content into your other channels. Good creator content is worth repurposing into your own content marketing and your email list, which most brands never do, so a piece of content that cost real money gets used exactly once.
The types of creator matter more than the platform
People ask me constantly whether it’s a TikTok thing now, or an Instagram thing, or whether YouTube still counts. The platform matters less than the type of creator and how their audience relates to them. A nano creator with 3,000 followers who reviews products with genuine detail can outsell a celebrity with millions, because the trust ratio is completely different. If you’re new to this, it’s worth reading through the different types of influencers and how to work with each one before you approach anyone, because a nano creator needs a completely different brief, budget, and set of expectations than a mid-tier one, and treating them the same is one of the fastest ways to waste money.
On the platform question specifically, TikTok has changed the game because shares and duets extend a post’s life well past the usual 48 hour window that Instagram content gets. If you’re running campaigns there it’s worth understanding how TikTok’s sharing and analytics work before you promise a client reach numbers you can’t verify.
Where it fails
I won’t pretend this works for everything. It falls flat for products with a long, considered sale, business software with a six figure contract value, for example, where a follower seeing a post is nowhere near ready to buy and the attribution window makes the whole campaign look like a failure even if it quietly built awareness. It also falls flat when a brand hands a creator a rigid script, because audiences can smell a read-aloud ad within seconds and the trust that made the creator valuable evaporates instantly. And it fails when brands treat one campaign as a complete strategy rather than one channel sitting alongside content, email, and paid ads.
The uncomfortable truth I’d add here, because it rarely gets said out loud in agency pitches, is that a fair chunk of the industry still runs on inflated numbers that nobody checks. Fake followers are cheaper than ever to buy, engagement pods exist specifically to fake early traction on a post, and plenty of “case studies” you’ll see in influencer marketing pitch decks are built on vanity metrics that were never tied to a sale. If a proposal lands on your desk quoting reach and impressions with no mention of conversion or attribution, ask why, because that omission is usually the answer.
What this means for a small business budget
You don’t need a five figure influencer budget to test this. A realistic starting test looks like 500 to 1,500 pounds spread across three or four micro creators in your niche, each given a unique discount code, run over six to eight weeks with content reused afterwards across your own channels. That’s small enough to survive a flop and specific enough to tell you something. Compare the redemption numbers, not the like counts, and you’ll know within one cycle whether this is worth scaling for your product.
If you’re running this alongside your other channels, it sits well next to whatever you’re already doing with social media management, since the same person who manages your organic posting is usually the right person to brief and track creators, rather than treating it as a separate silo with its own reporting.
Frequently asked questions
Is influencer marketing still worth it for a small business in 2026?
Yes, but only if you use micro creators in a specific niche rather than one big name, and track sales through unique codes rather than judging success by likes or reach.
What’s a realistic budget to test influencer marketing?
Start with 500 to 1,500 pounds across three or four micro creators over six to eight weeks, each given a unique tracking link or discount code, so you can compare actual redemptions rather than guessing from engagement numbers.
Do micro influencers really perform better than big names?
Often, yes, on cost per result. Micro creators with 10,000 to 30,000 followers commonly see engagement rates of 4 to 8 percent compared with 1 to 2 percent for accounts with millions of followers, which frequently translates into more clicks and sales per pound spent.
Why do some influencer campaigns completely flop?
Usually because the creator was chosen for follower count rather than audience fit, the brief forced a scripted read that broke trust, or the product has too long a sales cycle for a single post to influence, since nobody tracked whether the audience bought anything afterwards.
