Playbooks · 5 min read
How to Read an Instagram Automation Case Study in 2026
We already published our position on inventing customer results, and it is short: we will not write the case study, and that post explains why in full. Nothing here repeats that argument.
This post is the other half, and it is the more useful one, because you are going to read other people's case studies anyway. Every tool in this category publishes them. Some are real. Most are a landing page with a logo on it. Telling them apart is a skill, it takes about four minutes per case study, and nobody teaches it.
Start at the denominator, always
A number without a base is a decoration.
A headline lead count is meaningless until you know how large the account was, whether paid traffic ran to the posts, and whether the month contained a festival sale. With those three facts the same sentence usually describes something ordinary. Without them it describes a miracle, which is the point of omitting them.
So the first pass is mechanical. Find the base for every number. If the case study reports a total, ask total out of what. If it reports a lift, ask lift from what starting point — a doubling from four to eight is a doubling, and it is also four more conversations.
The claim is never in the number. It is in the base the writer decided not to print next to it.
The four things a real one contains
These are not stylistic preferences. Each one costs the writer something, which is exactly why fabricated case studies skip all four.
- An account you can open. A handle, not "a leading D2C brand". You should be able to go to the profile, look at the posts from the stated period, and see the comment sections the story is about. This is the single strongest signal and it is the cheapest to check.
- A date range with a season inside it. Real runs happen in a specific October, next to a specific festival, during a specific launch. A case study with no dates is unfalsifiable, and a writer who lived through the run always remembers when it was.
- What else changed. No automation runs in isolation. There was a launch, or an ad budget, or a collaboration, or a reel that travelled. A real account of a real month names the confounders because the person running it was there and could not miss them.
- The part that did not work. Every genuine thirty-day run contains a keyword that never fired, a template that embarrassed someone, and a week where nothing happened. A story with no failures in it was written backwards from the conclusion.
What the fabricated ones look like
Once you have read a few, they stop being subtle.
What the page shows you
What is missing from it
The italic testimonial deserves its own mention. "We were drowning in DMs and now our team focuses on what matters" is a sentence no human being has ever said to another human being. It is written by the marketing team and approved by the customer, at best, and approval is not authorship.
Survivorship, and who paid
Two structural problems remain even when every fact in the case study is true.
The first is selection. A vendor with a thousand customers publishes the three best outcomes. That is not dishonest, it is what a case study is for, but it means you are reading the top of a distribution and treating it as the middle. The flat months exist. Nobody writes them up, including us.
The second is who commissioned it. Case studies are produced by the party that benefits. Sometimes the customer received a discount, a co-marketing arrangement, or an account manager who did the setup for them personally. None of those are scandals. All of them mean the described setup is not the setup you will have on a Tuesday with the documentation.
Even a true one may not transfer
Briefly, because we have covered it elsewhere: the variables that dominate an outcome are the product, the price, the audience and the season, and the tool is far down that list. A good product with an engaged audience produces impressive numbers with a mediocre setup.
So the transferable part of a case study is never the result. It is the method — the sequence of decisions the person made, in order, and what they did when something broke. That part is portable. The number at the top is not, and the number at the top is what the page is designed to make you remember.
What to do with a case study that passes
Read it for the decisions, write down the sequence, and then run it on your own account and watch your own rows. The first hundred leads have a shape that is worth knowing before you compare yourself to anybody.
And apply the same four tests to us. We publish no customer results, no logos and no revenue figures, which means there is nothing here for you to verify — that is a limitation of our marketing, not a virtue, and you should treat the absence as an absence. Our own review of the product is the closest thing we have to an honest account of what it is like to use, and it was written by the people who built it, which you should weigh accordingly.
