We see the same pattern all the time with small and mid-sized businesses: a campaign starts strong, results look solid, then performance gradually levels off.
The typical response is to adjust targeting, raise the budget, or launch a new campaign. In many cases, though, the real issue is somewhere else entirely: the algorithms don’t have enough creative material to keep learning effectively.
A meta-analysis by NCSolutions covering nearly 450 advertising campaigns found that creative assets drive 49% of the incremental sales impact of advertising. That’s more than reach, brand, or targeting.
Yet in most small businesses, creative is still treated as an afterthought:
- a visual recycled from the website
- a single video running for months
- little variation in messaging
- formats never adapted to each placement
Meanwhile, advertising platforms have completely changed how they optimize campaigns.
That gap between what actually drives performance and what gets attention is exactly where today’s algorithms create a real opening, especially for teams without the resources of a fifteen-person creative department.
Algorithms Are Making More and More Decisions
Since the large-scale rollout of Performance Max (Google) and Advantage+ Creative (Meta), the mechanics have shifted. Performance Max runs simultaneously across Search, Shopping, YouTube, Display, Discover, Gmail, and Maps. For each placement and each user signal, it dynamically assembles the best combination of headlines, descriptions, images, and videos from the assets you provide.
The catch is that if your asset library is thin or too repetitive, the algorithm quickly runs in circles. When it has several creative angles, several formats, and several signals to work with, it learns faster and optimizes delivery much more effectively.
Google has confirmed this directly: asset groups rated “Excellent” in Performance Max tend to perform better, largely because they cover more placements and more creative variations.
Meta follows the same logic: Advantage+ can automatically generate and test up to 150 creative variations from the assets you supply.
In short, media performance today depends more and more on the quality and variety of the creative signals you feed the algorithm.
The Real Problem: Ad Fatigue
Here’s a scenario we see constantly with small businesses:
- campaigns perform well at the start
- cost per conversion creeps up gradually
- CTR drops
- frequency rises
- results plateau
And yet:
- targeting is solid
- budget is stable
- the offer works
More often than not, the problem is simply that the audience has been looking at the same ads for too long. Depending on spend levels, ad fatigue can set in after as little as seven to ten days. The typical response is to touch the media side of things, when the problem is actually creative.
A Creative-First Approach with AI
Before, maintaining a steady volume of new creative demanded enormous resources. Today, generative AI tools make it possible to produce and test a volume that previously required a massive production budget:
- hook variations
- format adaptations
- seasonal versions
- animated videos built from still images
- different CTA treatments
- multiple messaging angles
What a Strong Creative Library Actually Looks Like
For a small business, a solid starting point can be fairly simple. Begin with one strong asset (a genuine product photo, an authentic customer testimonial, a short video), then build out new angles from there:
- 3–4 photos showing the product in use
- 1 short video focused on a specific benefit
- 1 authentic customer testimonial
- 2 hooks built around a pain point
- 2 hooks built around a benefit
- 1 promotional visual
- formats adapted for feed, stories, and other placements
The goal isn’t to produce more for the sake of producing more. It’s to give the platforms enough material to identify what actually resonates with each audience segment.
From there, you can use the data those creatives generate to answer a genuinely strategic question: which specific message works for which audience? In other words, campaigns become continuous learning systems.
Our Method
For our small business clients, we’ve built a three-part approach:
1. Input: Build a Varied and Intentional Asset Library
What matters isn’t quantity, it’s coverage of the variables that influence purchase decisions: functional benefit, emotional benefit, social proof, urgency, comparison.
For a small business selling a B2C product, that might mean: a clean product image, a photo showing the product being used, a customer quote in text format, a 15-second benefit-focused video, and two headlines testing opposing angles (cost savings vs. comfort, for example).
2. Signal: Read the Creative Data, Not Just the Overall Numbers
Since 2024–2025, Google has offered asset-level reporting: impressions, clicks, and conversions broken down by headline, image, and description. Meta offers similar insights on Advantage+ variants. That level of detail is what lets you optimize the creative library over time.
3. Iteration: Treat Creative Renewal as a Routine, Not a Crisis Response
Rather than waiting for a performance dip to refresh creative, we recommend a review and renewal cycle every four to twelve weeks, tied to measurable fatigue signals. This isn’t a magic number, and every business and audience is different.
Generative AI makes this cycle viable without sending production costs through the roof.
What This Doesn’t Replace
I want to be honest about an important limitation. Generative tools speed up production, yes. They optimize delivery. But an ad isn’t just an algorithmic asset. It’s a cultural product aimed at people with feelings. It exists in a social, emotional, and temporal context. It lands or it doesn’t based on dynamics that data doesn’t always capture with precision. Reading the room, sensing what will resonate with a specific audience at a specific moment, separating what deserves to exist from what is just noise: that’s the domain of the human creative team.
Left to itself, AI often produces a lot of slop: content that is technically correct and well-structured, but without edge, without that small friction that makes an ad stick in people’s memories.
On top of that, speed of learning will never compensate for a hollow promise.
That’s why at Hamak, our role as a performance agency hasn’t shrunk with AI; it has changed. We spend less time on manual production and more time diagnosing, optimizing, staying ahead of the competition, and pulling out actionable insights for the next round.
Our Recommendation
If you run Meta or Google campaigns with a small team, here are three ways to strengthen your creative output:
- Turn on Advantage+ Creative. It’s free, takes five minutes, and gives you an immediate first round of learning on your variants.
- Block 30 minutes each week to review your asset reports. Not just overall numbers. Which headlines are converting? Which images are generating the most clicks? That data is your creative brief for the following week.
- Treat creative renewal as a budget line, not a one-time expense. If you’re spending several thousand dollars a month on media, set aside a regular budget to keep your asset library fed. Creative drives nearly half your results; budget for it accordingly.
Creative-first isn’t agency-speak to justify production fees. It’s what the data has been saying for years, and what the algorithms are now confirming at speed.
The best media optimization you can make is a creative one. And if you want to know where to start, our team is here to help you figure it out.
Example
49% of your advertising performance comes down to your creative assets. Not your targeting. Not your budget. Your creative assets.
That’s what a meta-analysis of nearly 450 advertising campaigns found. And honestly, it matches what we see in practice.
A campaign launches well. Results are there. Then gradually, it levels off. Cost per conversion creeps up, CTR drops, and the reflex is usually to tweak targeting or adjust the budget.
In many cases, the real issue is elsewhere: the audience has been seeing the same ads for too long, and the algorithms don’t have fresh creative signals to keep learning.
For a lot of marketing teams at small businesses, this has become a genuine challenge: how do you refresh your creative fast enough, without going into full production every week, and without a team of fifteen people to pull it off?
That’s exactly what I explore in my new article. I walk through our three-part method for adopting a creative-first approach.
I cover how to build an intentional creative library, how to set up your workflow to pull out real insights from it, and why tools like Performance Max and Advantage+, which are often underused, can become genuine performance drivers when you know how to feed them.
