Brand Case Study — 2025

What actually drives engagement for a leading yogurt brand on Instagram?

A post-level quantitative analysis of 118 organic posts from a leading U.S. yogurt brand, using OLS regression and text-derived content indicators to identify what content characteristics are statistically associated with higher likes and comments.

Request the full brand report
118
organic posts analyzed
12
months of 2025 data
268K
brand followers
R² 0.61
likes model fit
9
content predictors tested

Key Findings

All findings are drawn from a single brand’s statistically significant regression results. Effect sizes are approximate percentage changes relative to the baseline, back-transformed from the log scale. The brand is not identified in this public summary.

01

Collaboration posts are the single largest engagement driver — by a wide margin

Posts featuring an external collaborator — a creator, brand partner, or co-promoter — dramatically outperformed solo brand content on both engagement dimensions. The effect was statistically significant at the 1% level and dwarfed every other content variable in the model.

+369% more likes (p < 0.001) +76% more comments (p < 0.05)
02

Seasonal and limited-edition content consistently outperforms evergreen posts

Posts referencing seasonal themes or limited-edition products generated a reliable lift on both likes and comments, independent of whether the post also featured a collaborator or new product announcement. The effect held after controlling for calendar month, ruling out simple timing effects.

+51% more likes (p < 0.05) +44% more comments (p < 0.05)
03

Static images outperform video for likes — despite video dominating the posting mix

One of the most counterintuitive findings: static image posts generated substantially more likes than video posts at the 1% significance level, even though video accounted for the majority of content published. Format strategy appears misaligned with format performance for at least one major brand in this category.

+354% more likes vs. video (p < 0.001) Effect specific to likes; comments gap not significant
04

A brand's core identity theme may be working against it on Instagram

For one brand in the study, posts centered on the theme most associated with its brand positioning significantly underperformed on both likes and comments. The effect was robust, statistically significant at the 1% level, and held after controlling for all other content variables. This does not mean the brand should abandon its identity — but it does suggest the current execution of that theme on Instagram is not connecting with the audience as intended.

−41% likes (p < 0.01) −42% comments (p < 0.01)
05

New product launches generate a significant lift in likes

Posts explicitly announcing a new product were associated with substantially more likes than standard content, significant at the 5% level. The corresponding effect on comments did not reach significance, suggesting that novelty reliably attracts passive appreciation more than it sparks conversation. Launch posts that pair the announcement with a direct question or invitation to respond may be better positioned to lift both dimensions.

+86% more likes (p < 0.05) Comments effect directional but not significant

Methodology

This research applies quantitative methods from business analytics to a domain where most analysis stays qualitative.

Post-level data collection

Each post manually observed and coded across 17 variables including content type, visual elements, caption characteristics, and engagement outcomes across a full calendar year.

Text-derived content indicators

Caption and visual scene text parsed into structured binary indicators, enabling content themes to be tested as quantitative predictors alongside manually coded variables.

Month fixed effects

Seasonal variation and audience growth effects absorbed via month dummy variables, isolating content-driven engagement from calendar timing confounds.

OLS regression, log-transformed outcomes

Standard regression approach for count data with right-skewed distributions. Marginal effects back-transformed into percentage changes for direct interpretability.

Methodology Note

Data collected via direct observation of public Instagram profiles (January through December 2025). All findings are original analysis by Jimmy Chen Consulting. No post content, images, or captions reproduced. Based on independent analysis of publicly available data. Jimmy Chen Consulting is not affiliated with, endorsed by, or sponsored by the brand analyzed herein.

Get the Full Brand Report

The complete analysis covers one brand in full detail: regression results, confidence intervals, residual diagnostics, and prioritized strategic recommendations.

What is included

Brand-specific findings, not category averages

The public summary above presents anonymized, category-level insights. The full report is prepared for a specific brand and includes findings directly relevant to that brand's content mix, audience, and competitive context.

  • Complete coefficient tables with standard errors and p-values
  • 95% confidence intervals on all marginal effects
  • Format-by-format engagement breakdown
  • Month-level fixed effect estimates
  • Prioritized content recommendations tied directly to findings
  • Methodology note suitable for internal presentation

Your information will not be shared. You will receive one email with the report and a brief follow-up from Jimmy Chen Consulting.

Request received. You will hear from us within one business day. Thank you for your interest.