Framework · Creative strategy

Stop Spaghetti Testing. Build a Diverse Creative Ecosystem.

Spaghetti testing is throwing unrelated ad ideas at the wall to see what sticks, then scrambling again when they fatigue. A diverse creative ecosystem is the alternative: a range of ads that each play a different role, so performance compounds instead of spiking and crashing.

What is spaghetti testing?

Spaghetti testing is the practice of throwing a pile of unrelated ad ideas at the wall to see what sticks, then repeating the scramble every time performance drops. The ideas come from wherever: an ad someone saw on Twitter, a competitor’s angle, something off a podcast. You launch the best of them, a few work for a while, they fatigue, and you’re right back in the brainstorm. Nothing compounds.

If you run paid social long enough, you’ve lived this. I named it because I needed a word for the pattern I kept watching smart teams fall into, including my own.

And it’s exhausting. Every dip in performance feels like an emergency. CPA climbs, the client or the CEO gets nervous, and the pressure lands on whoever owns the ads. Even a winning ad doesn’t feel like a win for long, because you know it will fatigue and there’s nothing lined up behind it. You’re always one tired ad away from the next fire drill.

The trap is that we always think a new idea is the solution. The next angle, the next format, the next campaign setting will be the one that saves the quarter. It almost never is, and even when it is, it doesn’t last.

A new idea is never the fix. It’s just the next lap of the cycle.

How the spaghetti testing cycle works

It runs on a loop. Step five leads straight back to step one.

  1. Performance drops

    CPA climbs, the client gets nervous, and the pressure lands on the team.

  2. The team brainstorms

    Ideas from ads people saw online, the competition, a podcast. Can we test a traffic campaign? What if the ads were funnier? Should we only target hot states?

  3. Launch the best guesses

    A batch of unrelated ideas goes live to see what sticks.

  4. A few stick

    Performance improves. Everyone feels better, for a while.

  5. The winner fatigues

    Performance slides again, and you're back at step one, throwing more spaghetti.

The cycle feels like work because you’re busy. But you end the quarter roughly where you started.

Why does spaghetti testing feel productive but fail?

Because the wins are accidental. When an ad works and it didn’t come from anywhere, it was just a random idea that happened to land, you have no idea why it worked. And if you don’t know why, you can’t replicate it. You’re stuck waiting for the next lucky winner to show up.

So the real cost isn’t the losing tests. It’s that none of your tests build on each other. Each one starts from zero. Your account stays volatile, your results stay unpredictable, and the stress never ends.

Spaghetti testing vs. building an ecosystem

Spaghetti testing produces spikes: results jump when something lands, then slide back to where they started. When tests come from research and build on each other, the gains compound instead.

Before

Results rise and fall with each launch but return to the same baseline: no net gain. No net gain Ideate & launchIdeate & launchIdeate & launchIdeate & launch
New ideas go out, one hits, then fades, and the next batch starts from scratch. Nothing builds on what came before.

After

Starting from research and strategy, each round of review and iteration builds on the last, so results compound. Compounding returns Research, strategy & launch Review & iterateReview & iterateReview & iterate
Research sets the direction, and every review feeds the next round. Winners keep running while new tests fill the gaps.

The real problem isn’t ideas or optimization

When performance drops, the instinct is to blame the settings: we need a different campaign structure, a different objective, a new optimization. Or we decide we just need a better idea.

But most of the time, the problem isn’t optimization and it isn’t ideas. The problem is that the account was never built diverse enough in the first place. It’s overinvested in one ad, or a handful of ads, so the moment they fatigue the whole account wobbles.

You can’t optimize your way out of a structural weakness, and no single new idea fixes an account that’s overinvested in one winner. What the account actually needs is an ecosystem.

What is a diverse creative ecosystem?

A diverse creative ecosystem is an ad account built from a range of creative that each plays a different role: speaking to different people, at different stages of buying, for different reasons. Instead of betting the account on one or two winners, you spread risk and reward across many ads, and every test builds on the last.

When you have that, performance gets steadier, more predictable, and far easier to scale as you increase spend.

Think of your ads like an investment portfolio

A good portfolio has different assets doing different jobs. Some are there to grow your money. Some are there to protect you from risk. And you’re spread across many holdings, so you’re never overexposed to any single one.

Your ad account works the same way. You want ads that play different roles, and you never want to be over-indexed on any one idea, or even a handful of ideas. A diverse creative ecosystem is simply a portfolio approach to creative: predictable, consistent, and built to compound.

Never over-indexed on any one idea, or any handful of ideas.

Spaghetti testing vs. a diverse creative ecosystem

Spaghetti testingA diverse creative ecosystem
Where ideas come from Whatever someone saw this week: a competitor, a podcast, a viral ad Customer research: the problems, motivations and words real buyers use
Where the spend goes Most of the budget rides on one ad, or a few versions of the same idea Spread across many ads, each doing a different job
What a test tells you Whether this one ad worked Why it worked, and what to make next
When an ad fatigues The whole account wobbles and the scramble starts again It barely moves the needle; other ads pick up the slack
When you raise spend More pressure on the same few ads, so fatigue arrives faster More budget finds more of the right buyers
Over a quarter Spikes and crashes, ending roughly where you started Steadier results that climb as the library grows

What makes an ecosystem actually diverse

Diversity isn’t just casting different people in the same ad. Taking one script and having five different people read it is not a diverse ecosystem.

Real creative diversity has nine dimensions, and we plan them in priority order. The first four change what the ad says, and they do the most work, because a new message reaches people an old one never could. The next three change how the ad is made. The last two keep the account in step with the calendar and get it beyond your own brand page.

Each of these is a lever, and most accounts are only pulling one. And because the creative comes from customer research, when an ad wins you know why, and you can make ten more like it instead of guessing again.

The 9 types of creative diversity, in priority order

  1. What the ad says

    The message. These four matter most.

    1. Persona Who the ad speaks to. Different buyers have different reasons to buy.
    2. Angle The reason to buy it argues: a pain point, a desire, an objection or a comparison.
    3. Concept The idea or story that carries the angle: a demo, a confession, a side-by-side.
    4. Role The job the ad does: open a new audience, answer an objection, or close a warm buyer.
  2. How it’s made

    The execution. Same message, new ways in.

    1. Format Video, static, carousel or text-led, at the lengths and ratios each placement favours.
    2. Talent Who’s on screen: customers, creators, the founder, an expert.
    3. Style The look and feel: phone-shot and raw, polished studio, motion graphics.
  3. When and from where

    Timing and distribution.

    1. Seasonality Creative that fits the moment: the season, holidays, launches and sale windows.
    2. Account Partnership ads that run from creators’ handles as well as your brand page.

The goal: an ad for every customer

Your audience is far more diverse than your creative usually is. A lot of different people could buy from you, for a lot of different reasons, and the platform’s algorithm is remarkably good at matching the right person to the right ad. Your job is to give it enough variety that it always has the right ad to show.

That’s the whole philosophy in one line:

An ad for every customer.

Case study · Perfectly Snug · Ecommerce

Case study: from $1M a year to a $1M month

  • Perfectly Snug was launching a new product category at a premium price, in a seasonal business, against a well-funded category leader. One great ad was never going to carry it.
  • Every month we added new concepts and customer motivations instead of replacing old winners, so the library kept growing: air vs. water for colder months, hot summer nights for peak season, founder-led ads, and urgency when the warehouse neared capacity.
  • The business grew from about $1M in annual revenue in 2024 to $3M in 2025, then hit its first $1M month in April 2026.
  • In 2026, CPA was 24% lower than in the same months of 2025, while spend increased by 122%.

Creative was one part of a broader business effort; these figures aren't attributable to ads alone.

Frequently asked questions

What is spaghetti testing in advertising?

Spaghetti testing is launching a batch of unrelated ad ideas to see which ones stick, with no underlying research or strategy connecting them. It produces occasional wins you can't explain or repeat, so performance spikes and crashes instead of compounding.

Why does spaghetti testing stop working?

Because its wins are accidental. When you don't know why an ad worked, you can't replicate it, so every test starts from zero and the account stays volatile and dependent on luck.

How do I know if I'm spaghetti testing?

The signs are a new brainstorm every time performance dips, test ideas that come from whatever someone saw that week, winners nobody can explain, and an account that leans on one or two ads at a time. If you end each quarter roughly where you started, despite constant testing, you're probably in the cycle.

Is A/B testing the same as spaghetti testing?

No. Structured testing isolates one variable to learn something specific. Spaghetti testing throws unrelated ideas at the wall at once, so even when something wins you can't tell what caused it.

What is a diverse creative ecosystem?

It's an ad account built from a range of creative that each plays a different role (different audiences, buying stages and reasons to buy) so performance compounds and scales instead of spiking and crashing on one or two winners.

Why does diversifying ad creative improve performance?

Because it removes single points of failure. When no single ad carries the account, fatigue in one ad barely moves the needle, and the algorithm always has the right ad to match to the right buyer, making results steadier and easier to scale.

How many ads do I need for a creative ecosystem?

There's no magic number. The point is coverage, not volume. You want an ad for every meaningful customer, buying stage and angle, with no single ad responsible for too much of your spend.

Is creative diversity the same as making variations of an ad?

No. Variations change the surface of one idea: a new hook, new music, a different person reading the same script. Diversity changes what the ad says and who it's for, such as a new persona, angle or concept, so it reaches buyers the original never would.

Resources

Services that use this framework

Ready to stop throwing spaghetti?

Request a Call