You uploaded the assets Google asked for. Headlines, a couple of images, maybe a video you had lying around. Ad Strength says “Excellent.” And the campaign is still spending more to enroll a student than it did three months ago.
That gap between a green Ad Strength label and a campaign that won’t scale is the most common thing I see in education accounts running Performance Max. The fix is in the creative, and specifically in how you feed it. The bid strategy and the budget are usually fine.
So here’s what actually moves Performance Max creative best practices from theory into enrollments: treat every asset as something that has to stand on its own, line up each asset group with the audience signal behind it, read Ad Strength for what it really is, and learn to spot in the reporting when the creative is the bottleneck and the bid never was. Let me walk through each one the way I’d explain it to a client over a call.
The rule that fixes most of it: every asset must stand alone
Performance Max takes your headlines, descriptions, images, logos, and videos, and it mixes them into combinations you never assembled yourself. It shows one headline with one image on YouTube, a different pairing on Discover, another on Gmail. You don’t get to decide which goes with which. Google does, per impression, per surface.
So the rule writes itself. Every asset has to make sense on its own, with any other asset, on any surface.
Most education accounts break this without realizing it. They write headlines that only work in sequence, like one that sets up a promise and a second that pays it off. Then Google shows the second one alone, on a YouTube thumbnail, to a cold prospect who never saw the first. Now you’re paying to show a punchline with no joke.
For an online school the test is simple. Take any single headline. Take any single image. Imagine a parent or a 28-year-old considering a career change sees just that one piece, in their feed, mid-scroll, with no context. Does it say something true and complete about the program? If it needs a neighbor asset to make sense, it’s a weak asset.
This is also where the offer shows up. An ad is an amplifier. If you’re a mediocre singer and you plug in an amplifier, more people just find out you’re unremarkable. The same is true here. A clean, standalone asset that carries a vague offer (“Enroll today, transform your future”) amplifies a vague offer. Lead with the actual outcome: the certification, the time to completion, the job the program leads to. The asset is the messenger; the offer is the message. Google’s own Performance Max asset best practices push the same direction, more and varied assets, but the standalone test is what keeps “varied” from turning into “weaker.”
Line up the asset group with the audience signal
People treat asset groups like folders. They’re not. An asset group is a creative-plus-signal package, and the algorithm reads them together to decide who to chase.
Here’s the part most owners miss. In Performance Max, your creative is part of your targeting. The people who engage with a given asset teach Google who else to go find. So if your asset group says one thing and your audience signal points at someone else, you’re handing the algorithm two contradictory hints and asking it to sort out the mess on your budget.
One clarification, because it trips people up. Audience signals are suggestions, not targeting. Google’s Performance Max audience signals documentation is explicit that the campaign will serve outside your signal whenever it predicts a conversion elsewhere. That doesn’t make signals useless. It makes the creative-to-signal alignment the thing that decides how fast and how cleanly the campaign finds your people during the learning phase.
For an education advertiser, that usually means splitting by intent stage, not by product line. One asset group built around cold prospecting: a free guide, a webinar, an outcome story, paired with a signal of people who behave like your audience early in their research. A separate asset group built around the warm enrollment push: the program, the start date, the call-to-enroll, paired with site visitors and prior leads. Cold creative against a warm signal, or a “book your enrollment call” CTA thrown at someone who’s never heard of the school, just slows everything down.
When the asset group and the signal agree, the early engagement data is clean, and clean early data is what shortens the learning phase. The algorithm stops guessing sooner because you stopped sending it mixed messages.
Why this matters most in the first few weeks
Performance Max needs a couple of weeks and a real volume of conversions before its model settles. Don’t judge it in week one. During that window, every contradictory signal you feed it, mismatched creative, brand traffic it shouldn’t be claiming, a bottom-funnel CTA aimed at cold viewers, is a data point it has to unlearn. Aligned asset groups mean fewer wrong turns, which means it gets to “I know who converts here” faster. That’s the whole game in the learning phase: feed it less noise, and it learns sooner.
Ad Strength is a checklist, not a scoreboard
This is the one I argue about most. Ad Strength is a completeness measure. It tells you whether you filled the slots, gave enough headlines, added the ratios, uploaded a video. It does not tell you whether any of it makes money.
I’ve watched “Excellent” assets quietly bleed budget and “Good” assets carry a campaign. High Ad Strength and high performance are different questions, and the label answers the boring one. Google itself frames Ad Strength as a measure of relevance, quantity, and diversity of your assets, which is exactly a completeness checklist, not a performance forecast.
So use it for what it’s good at. Run it as a fast QA pass: did the team actually upload every ratio, a square and a landscape logo, a vertical video, enough varied headlines? Green means the slots are filled. Then ignore it for the real decision and judge assets on per-asset performance data, conversions and conversion value, not on the score.
Chasing the green label is how education accounts end up with twelve interchangeable headlines that all say “advance your career” and nothing that says what the program actually does. Completeness without distinctiveness. The checklist is satisfied and the campaign still can’t find its people.
Reading the limited reporting: when it’s the creative, not the bid
Performance Max hides a lot. You can’t see per-surface spend without a script, and the asset reporting is thinner than anyone wants. But there’s enough in there to tell you whether your problem is creative or something else, and most of the time, when an established school with a solid offer stalls in Performance Max, it’s the creative.
A few things to read for:
Asset performance labels. Google rates each asset Low, Good, or Best. A pile of “Low” assets and one lonely “Best” tells you the algorithm has nothing to work with, it’s serving the one thing that works and burning impressions on the rest. That’s a creative supply problem, not a bid problem. Swap the Low assets for genuinely different angles, not reworded versions of the loser.
Channel skew. If you can see your network distribution and the campaign is dumping most of its spend on Display when you expected Search and YouTube, your creative is pulling the algorithm somewhere you didn’t intend. Weak or off-message assets get cheap impressions on cheap inventory, and the campaign chases the cheap clicks.
Impressions without conversions. When you’re getting plenty of reach but enrollments aren’t moving, the bid is clearly finding people. The message isn’t landing once it does. That’s the offer and the creative, not the auction.
The tell is consistent. When volume is there and cost-per-enrollment is climbing, more budget makes it worse, because you’re amplifying the same message that isn’t converting. The lever is the asset, and the offer behind it.
None of this works on broken tracking
Here’s the catch that sits underneath everything above. A creative test on broken tracking is a coin flip with extra steps. If Performance Max is optimizing toward a conversion that fires unreliably, or toward a cheap form fill that never becomes an enrolled student, then your “best” assets are just the ones that produced the most of the wrong thing.
For education this is the trap. Performance Max for lead gen finds the cheapest form fills, and the cheapest form fills are usually your worst leads, unless you feed it a real downstream signal. Import the quality data: the enrollment call that happened, the application completed, the deal that closed, with values attached. Then the algorithm optimizes toward students, not toward whoever fills a form fastest. I’ve written more on conversion tracking for online course businesses if you want the full setup.
And I’ll be honest about how often this is the actual problem. I’ve never met one client whose conversion tracking was perfect. Not one, and I’m serious. It’s not impossible, it’s just hard for non-technical teams to get right. So before you blame your creative, confirm the campaign is even measuring the thing you care about.
Where this sits in the bigger picture
Performance Max is additive and it’s an amplifier. It works best running next to a tight Search campaign, on top of solid tracking and a clean account structure, with creative that carries a strong offer to the right awareness stage. That’s the order: tracking and structure first, then the creative and video that amplify a working offer. Get those right and Performance Max scales a school that already converts. Get them wrong and it scales the waste.
The reason creative is the lever now is simple. Google took the bidding, the targeting, and the placements. The asset is the last thing that’s genuinely yours, which is why we build the foundation first and then produce the creative and video that carry the offer across surfaces. That’s the through-line in why creative is the new targeting for education, and it’s worth reading next.
If you want the surrounding strategy, mastering Google Search ads for online schools covers the Search side that Performance Max should complement, and the top trending eLearning ad campaigns show the kinds of creative angles that earn the engagement Performance Max learns from.
Start with the standalone test on your current assets this week. Pull up your asset group, look at one headline and one image at a time, and ask whether each says something true and complete to a stranger mid-scroll. The ones that fail are quietly costing you. That’s the cheapest fix in the whole campaign.
