Across the 14,763 university ads we captured, 68% carry no call-to-action button and 16% of Meta ads carry no detectable message at all. Behind those numbers sits a creative production pattern we measured across 14,126 ad images: 93% use a headline overlay, 89% show a logo, and the single most common composition is the same stat-or-text card, at 32% of our vision sample. This post is about the craft of the category’s ads, how they are assembled, and the two places the budget behind them leaks.
Scroll a student’s Instagram feed during admission season and try to tell the universities apart with the names covered. In our data, you mostly cannot, and now we can put numbers on why: on how the category’s ads are assembled, and on the two specific places where the money behind them stops working.
This is the second post in our ad intelligence series. The first mapped what 174 universities advertise, where, and when. This one is about the ads themselves: 14,126 downloaded ad images analysed for format, colour, and on-image text, plus a 180-creative sample run through visual analysis for composition and style.
Build the Category Ad Yourself
The findings read like a single design brief the whole category agreed to follow without ever meeting. The fastest way to see it is to assemble it:
Build the category ad yourself
Four components, each with its adoption rate across the corpus. Switch them on and watch the average university ad assemble itself.
(the only part that varies)
Switch each layer on. A headline overlay (93% of creatives have one). A logo (89%). A ranking badge (39%). A blue accent (51% of the vision sample). By the fourth toggle you have built the ad that thousands of teams across the category are shipping this cycle, and you built it in four clicks because it only has four moving parts.
The composition data confirms what the toggles suggest. The stat-or-text card is the most common creative type at 32% of the sample, ahead of professional portraits and student portraits. Colour families across all 14,126 images: white and grey dominate at 55%, blue and teal at 23%, black and dark tones at 10%. Formats spread across square (29%), wide (26%), portrait (17%), story (17%), and landscape (11%), so the sizes vary while the design does not. And 59% of creatives rate as polished, which is worth pausing on: the category’s problem is not production quality. Polish is not the gap. Distinctiveness is.
Why Does Sameness Cost Money?
None of these choices is wrong individually. Logos build attribution, headlines carry the offer, blue photographs well against campus skies. The problem is what happens in aggregate: when 174 institutions run creatives with the same four components, the same composition, and the same palette, into the same audience during the same eight-week window, the ads stop being distinguishable from one another. The auction charges every one of them full price for impressions that blur together.
Matching the template buys safety: no stakeholder will object to an ad that looks like everyone’s. Breaking it buys attention: in a feed where 93% of competitors run a headline overlay on a stat card, the creative without one is the only thing that looks like content instead of an ad. That is the trade-off, and the data says the category has priced it wrong in one direction.
Leak One: the Words Are in the Wrong Place
The images are doing work the platform cannot see. On-image text across the corpus mentions a deadline 5,197 times, a ranking 1,930 times, placement 1,551 times, and a scholarship 812 times. Text baked into pixels is invisible to ad-platform optimization and unreadable in several placements. Meanwhile, in the ad copy itself, where the platform can actually read and optimize, deadline urgency appears in 3% of ads and scholarship offers in 3%.
Hold those two numbers together: the category writes the deadline into the picture 5,197 times and into the offer almost never. The strongest conversion levers it owns are decorating the creative instead of driving the campaign.
Leak Two: the Ad Never Asks
An ad does three jobs: it reaches someone, it says something, and it asks for something. The reach costs money every time. Across the corpus, the third job is mostly not happening: 68% of ads carry no call-to-action button. On Meta, where adding one is a dropdown, the omission is not a technical constraint. It is a habit, repeated across thousands of ads, each one paying full price for the impression and then leaving the viewer with nowhere to go.
The difference is easier to feel than to describe:
The same ad, with and without its next step
The “before” is what 68% of the category ships: creative, headline, and nowhere to go. Toggle to “after” and the only changes are the three things the data says are missing.
Toggle between the two states. The “before” is what 68% of the category ships: a polished creative, a campus photo, a headline, and then nothing. The viewer who is genuinely interested has to screenshot it, remember the name, and find the admissions page on her own. Every step she has to invent is a step where the funnel loses her. The “after” adds the three things the data says are missing: a CTA button, a message with actual admission intent, and urgency worth acting on. Same creative budget, same auction price, different job done.
Underneath the missing button, the pattern deepens: 16% of Meta ads carry no detectable message at all, not admission, not brand, not discovery. Roughly one Meta ad in six is reach with nothing attached to it. And at the far end of the same spectrum, 20 universities run no ads on any platform while their competitors launch thousands a month.
One honest caveat, because it changes what you fix: a CTA narrows an ad’s audience to people ready to act, which is exactly why brand and discovery ads legitimately run without one. The failure in the data is not that some ads skip the button. It is that two-thirds do, including the admission ads whose entire purpose is the click.
The One-Afternoon Audit
Here is the check we would run on any university’s ad account, in order, cheapest first:
Step 1. Export your active ads and count the CTA buttons. If more than a handful lack one, you have found budget that can start working this week. Adding a button is a settings change, not a creative project.
Step 2. Read each ad’s copy and name its job in one word. Admission, brand, discovery, placement, event. If you cannot name it, the algorithm cannot either, and neither can the student. Rewrite or retire it.
Step 3. Move the hooks out of the image. Deadline, scholarship, ranking: into the copy and the CTA where the platform can optimize on them. Keep the image for the thing only images do, which is making a viewer feel something about the place.
Step 4. Take one running ad and strip it. No overlay, no badge, one human, one line of copy in the caption instead of the pixels. Run it against the template version with the same budget and read the cost-per-result after two weeks. You are testing distinctiveness, and the test costs nothing but a variant.
Step 5. Claim a colour. 55% of the category is white and grey, 23% blue and teal. If your brand palette contains literally anything else, the feed is handing you differentiation for free.
The full creative analysis, including the format mix, colour families, and the vision-sample breakdown, is in the report, alongside the money-leak section this post draws from. Read the Ad Intelligence report →
And the sameness problem extends past paid media: the digital competitors your applicants actually see, most of which are not other universities, are mapped in 7 digital competitors your university doesn’t know about →
This post draws on Thrivemattic’s analysis of 14,763 ads and 16,907 creatives from 174 of 194 Indian private and deemed universities, part of The Digital State of India’s Private Universities 2026 →
Start of this series: what 14,763 university ads reveal →
