This question arrives in two forms. An owner asks it because they are afraid to spend, and a marketer asks it because they feel the thing is over. Both want a yes or a no, and I am not going to give one in that shape, because the honest answer has a number in it.
So this time I am not starting with my own experience. There are two published meta-analyses in a peer reviewed journal, and one of them pools 135 experimental studies. We will start there and then talk about your business.
The short answer: yes, and the effect is smaller than you are told
A 2025 meta-analysis in the Journal of the Academy of Marketing Science by Barari, Eisend and Pratap Jain pooled 71 papers, yielding 135 experimental studies and 571 effect sizes. Its stated finding is that social media influencers significantly impact both consumer engagement and purchase intention, and that they are relatively more effective than brand posts, virtual influencers and celebrities.
That sentence is the one an agency would quote at you. The numbers inside the paper are calmer.
| Relationship | Effect size | Studies |
|---|---|---|
| Influencer to credibility | 0.23 | 73 |
| Influencer to engagement | 0.14 | 40 |
| Influencer to brand attitude | 0.13 | 45 |
| Influencer to purchase intention | 0.11 | 86 |
| Influencer to attractiveness | 0.10 (not significant) | 23 |
A correlation of 0.11 with purchase intention is not a number you build a promise on. It is a number that says the thing works, really and modestly. That is the condition of most marketing when it is measured seriously. The problem is not the figure, it is the post telling you influencers will double your sales.
The line that stopped me
Attractiveness is not significant. Across 23 studies, how good the influencer looks had no effect distinguishable from zero, and the confidence interval crosses it.
Credibility, in the same table, is the largest figure at 0.23, roughly double the purchase intention effect. The paper's model has influencers working on consumers indirectly through credibility and attractiveness, so the path exists, but credibility is what carries it.
Now hold that against how influencers are actually chosen. People choose on looks and on follower count, and those are precisely the two weakest columns. None of which will surprise you if you have read why chasing followers is the wrong target, except that this time there is a number attached.
Size: where the line actually sits
The paper defines it plainly. Large means more than one million followers; small and medium means under one million. Then it finds something useful: small and medium influencers are more effective at driving engagement, while larger ones have greater impact on purchase intention.
That is the first explanation I have seen for why this argument never resolves. The people telling you small is better are measuring engagement. The people telling you large is better are measuring purchase intention. Both are right, about different things.
So the decision is not who is better. It is what you want.
- You want people talking, replying, trying. Small and medium, and more than one of them.
- You want a name that lends legitimacy to somebody who has never heard of you. Large, and one is enough.
- You want sales this month. Neither, on its own. That is a different problem.
What actually drives it, from a second analysis
A second meta-analysis in the same journal, from 2024, by Pan, Blut, Ghiassaleh and Lee, synthesised 1,531 effect sizes from 251 papers and asked a different question: what decides the outcome.
Their answer reorders the priorities. For non-transactional outcomes such as attitude and engagement, follower characteristics, specifically social identity, have the strongest effects. For purchase intention, post characteristics, specifically informational value and hedonic value, exert stronger effects. For actual purchase behaviour, influencer characteristics, specifically influencer communication, have the strongest effects.
Translate that into a decision. Social identity means: does their audience see itself in them. Informational value means: does the post say something useful rather than say buy this. You can judge both of those before you pay anybody, without a single figure from the influencer.
So what do I do with a small business
Five things, all of them derived from the above rather than from my taste.
- Select on credibility, not on looks or counts. Read the comments under their last ten sponsored posts. If people are asking about the product, credibility is there. If it is all emoji, it is not.
- Ask who else they have promoted in your category. Somebody running three competitors in two months is spending down their credibility, and you are paying the average of all their ads rather than the value of yours.
- Buy content, not an appearance. If the deal ends when the post ends, you rented a moment. If it includes usage rights in your own ads, you bought an asset.
- Do not measure it in one week of sales. Purchase intention is not purchase, and the table above lists them as separate numbers for a reason.
- Decide first whether you need advertising at all. All of this assumes a product people already want. If you are not sure yet, go back to when to start spending on ads.
And if you are looking for growth without an ad budget in the first place, influencers are not the first door. The other doors are in how to grow without paid ads, and whether social alone can carry a business is in is social media enough for a business.
The operational part the studies do not cover
If you go ahead, there is an execution detail that wastes real money and has nothing to do with selection: permissions. Running an influencer's own post as a paid ad has documented rules, permission types and durations, and the detailed read on that is Mostafa Faried's piece on Meta partnership ads, permissions and documented limits, which is in Arabic. Read it before you sign, not after, because the permission term belongs in the agreement.
Limits worth stating
Three.
First, both analyses rest largely on experiments, and the 2025 paper itself notes that many primary studies use hypothetical influencers. So the numbers measure the principle more than they measure a live campaign in Egypt.
Second, not one figure above is Egyptian or Arabic. I could not find a meta-analysis on this market, and I am not going to take a global correlation and call it ours. Use the direction, not the value.
Third, and this is the important one: the question in the title is a bad question that looks like a good one. Does influencer marketing work is the same shape as does advertising work. It depends on your offer, your selection and the price of the deal, and every number above is an average across all cases. Use them to lower your expectation, not to make your decision.
The simplest test beats any channel matrix: ask the last ten customers who paid you how they found you. The answers show which channels work, and which ones you pay for that nobody came from. That closes Part 5 of my free book Growth Engine, which places this subject inside a complete growth system: Part 5, Digital channels: how you reach them (in Arabic).