Every month here at The Media Buyer we publish retrospective reports comparing last month to the same month the prior year.
These images are published (with some commentary) across our socials, but if you want the full images for download, be sure to subscribe to The Media Buyer. I will publish all files here.
Now we look at May 2026 versus May 2025, and what we find is most of the same challenges we’ve seen already this year. Spend increases, performance dips, and increasing inefficiencies for advertisers across the spectrum.
This is all independent Northbeam performance data, covering more than six billion in ad spend over the last 12 months. These numbers were calculated by comparing advertisers who had full sets of May 2026 and May 2025 data in Northbeam, meaning it’s a true comparison of changes for advertisers who were on Northbeam during both periods.
Let us begin the litany of suffering with some high level metrics:
Immediately we see that spend is up 10% but revenue is only up 3%. It’s more expensive to acquire the same customer, and revenue isn’t keeping up.
There isn’t one specific cause for this. Ad platforms need to increase their profits and numbers of advertisers on the platform, and both increase ad costs. Inflation is driving up spend, making your dollars less powerful. Depressed consumer confidence is holding down conversion rates. Nobody knows if the Strait of Hormuz is open or closed, and nobody knows if they should care.
All of these swirl together into a perfect storm of down-trending ad efficiency that we’ve seen for months.
A 10% increase is just the average. Let’s categorize our advertisers into more useful sections, and focus on Meta ads.
We’ve found the best way to categorize ad accounts and compare them is by their monthly spend levels. So here’s revenue changes, median and average, across categories of monthly ad spend on Meta.
Spend is up for everybody except the sub-$50k advertisers, who are mostly yanking spend back and struggling more. What’s interesting is the movement in the middle tier of advertisers. These accounts are probably still finding ad efficiency on Meta, where the upper end of spenders are probably expanding more to other channels.
I like to think most of these increases are the result of growth, and that most ad accounts using Northbeam are so efficient that they’ve increased their spend year over year as a result of scale. To be honest, that probably applies to most of our advertisers.
While a lot of people will imagine this as a decreased efficiency of Meta ads, the fact is: Meta is still the dominant full-funnel ad platform available to us. Just look at how little marketing budgets shifted year-over-year:
For all the teeth gnashing and hand wringing about Meta, nobody’s moved their spend from there, not yet. At least not in Q1. This means we get the most boring Sankey chart of all time. And yet, you can see some big year-over-year changes on some of the smaller ad platforms. Advertisers are shuffling creative testing around, trying to blow up stuff like AppLovin and TikTok.
The Meta conversation gets challenging when you chart revenue and spend next to ROAS and CAC, however.
ROAS generally worsened for Meta advertisers in May 2026 compared to May 2025. Especially when compared to the spend increases. This is a clear sign of inefficiency growing.
Worst still was customer acquisition cost. This slipped several percentage points even as spend increased by double or more.
There’s no way around it: we are paying more and acquiring fewer customers. You have to plan for this in your marketing. The Meta “free money machine” is becoming more challenging to game. This is especially true for low-AOV advertisers.
Let’s look at spend, revenue, etc. based on the AOV of the advertisers in Northbeam. This is across all ad platforms, not just Meta.
No matter your average order value, CAC got worse for you this year. Media efficiency ratio slipped as well. Low-AOV advertisers, long the one-day-last-click superheroes of growth marketing, are getting absolutely crushed - while high AOV products are more easily able to succeed in Meta’s new paradigm.
High-AOV advertisers, empowered with larger margins, seem more capable and flexible for dealing with the increased inefficiencies on these ad platforms.
How to use this data
This performance data is representative of the billions we track in Northbeam, but it’s also representative of our customers and their unique performance profiles. To humblebrag: most Northbeam users are simply better performance marketers than most of their peers. The data you see here is representative of that sophistication.
Does your performance line up with what you see here? If you are worse than the averages for your category, you likely could find improvements just turning on Northbeam for your business.
Are you preparing your budgets for increased inefficiency? If you are pretending or expecting to emulate last year’s performance on this year’s budget, then you are setting yourself up for failure. You need to be conceptualizing your ad spend expecting your performance to just be worse.
Analyze your year-over-year performance through the lens of the charts above - what do you see? Are you out or under-performing the bar? People constantly ask me for industry-level breakdowns of this data. The problem is: industries are not homogenous in a way that’s useful for analyzing ad performance. If you’re in skincare, there are other skincare brands running ads for everything from a $10 AOV to a $300 AOV. The budgets, user journeys, and ad account structures for a $10 AOV is wildly different from a brand with a $300 AOV. So compare yourself based on your spend and your AOV, not your industry.
As always, stay tuned: June is almost over, and we’re hoping to have a whole new set of visuals for that monthly report.