Influencers who posted a very low or very high proportion of sponsored content tended to receive more engagement on later sponsored posts than those in the middle, according to a study combining social media data with experiments.
The research, published online in January 2026 in the Journal of Marketing, analyzed posts on Instagram and Douyin and tested audience responses in controlled settings.
For brands choosing creators, the results suggest that counting paid posts alone may not explain how an audience will respond. The study measured engagement with sponsored content, not the sales or profit generated by a campaign.
What the researchers measured
Yijing Li, Angela Yi Gao, Flora F. Gu and Fine F. Leung define an influencer’s endorsement rate as the proportion of posts sponsored by brands. Someone with two sponsored posts out of ten has an endorsement rate of 20% for that set of posts.
One analysis covered 6,869 sponsored Instagram posts from 4,908 influencers over one month. A separate analysis examined 533,092 Douyin videos from 1,000 influencers, including 147,877 sponsored videos, over nearly five years.
In both datasets, the estimated relationship between sponsorship frequency and engagement formed a U shape: engagement tended to be lower in the middle of the observed range than at its low and high ends. The field analyses are observational, so other differences between creators and audiences may contribute to the pattern.
The researchers also ran two experiments. In one Instagram-based experiment, participants shared the sponsored post more often in low- and high-sponsorship conditions than in the intermediate condition. The authors report that 28 of 67 participants in the low condition, 15 of 69 in the middle and 32 of 72 in the high condition shared it.
Why more ads can send mixed signals
The authors propose two competing perceptions. Frequent sponsorship can make a creator appear recognized by brands, while also raising concern that the creator is trying to manipulate the audience. Their experiments examined those perceptions as possible explanations for the engagement pattern.
A creator’s previous unsponsored product mentions and repeated endorsements of the same brand also affected the estimated relationship in the researchers’ analysis. These results do not establish that any particular posting schedule will work for every creator.
Audience interests, product category, platform, creative quality and the commercial terms of an agreement can all affect a campaign. The Instagram and Douyin samples also came from different markets and periods.
Likes are not sales
Engagement can show whether people interact with a post, but a like or share does not by itself establish that someone purchased a product. A campaign intended to introduce a brand may be judged differently from one intended to produce immediate orders.
Our guide to measuring additional sales from marketing explains why comparing exposed and unexposed audiences can answer a different question from counting interactions.
Brands assessing sponsored content can review the creator’s audience, prior posts and campaign outcomes together, while keeping disclosures clear. The Federal Trade Commission’s influencer guidance says material connections should be disclosed clearly and conspicuously.
The study’s observed turning points are estimates for its samples, not universal thresholds. A sensible campaign assessment still requires evidence about the audience a business wants to reach and the outcome it is paying for.
