Three weeks ago, a founder told me his ads weren't working anymore.
He'd spent €18K in the last 90 days. His conversion rate was 1.2%. He was about to fire his agency.
The creative wasn't tired. The audience wasn't fatigued. The targeting wasn't wrong.
The ads were fine. The funnel was broken in step three.
This is the diagnostic I run every time a founder tells me "ads aren't converting like they used to". In nine cases out of ten the problem is not the ads.
Why founders misdiagnose ads
The instinct is rational. Ads are the surface. They're what you see. They're the line item on the credit card statement. When revenue stops scaling, the most visible thing to blame is the most expensive thing.
But ads are an amplifier, not a generator. They amplify whatever your funnel is doing. If the funnel converts at 2%, ads will pull 2% out of the traffic you send. If the funnel converts at 0.6%, ads can be perfect and the result will still feel broken.
Most founders never look at the funnel because:
- The funnel is invisible until you build the dashboards.
- The funnel feels like "what the developers should fix", not what marketing owns.
- The funnel is harder to A/B test, slower to iterate, less satisfying to talk about.
So they swap agencies, refresh creative, retarget warmer audiences. Three to six months pass. The numbers don't improve. The next conversation is the discovery call with me.
Here are the five funnel steps where money is actually leaking, in the order I check them.
1. Awareness alignment
The trap: the ad creative promises one thing, the landing page delivers another.
You see this constantly. The ad says "scale your team without burning out". The landing page hero says "AI-powered task management". Same product. Two different stories. A cold visitor takes seven seconds to decide if they're in the right place. If the page doesn't echo the ad, they leave.
What to check:
- Headline of the LP and headline of the ad are the same sentence or near-identical. Not "similar tone". Same concept, same words where possible.
- The first proof point on the LP addresses the same pain the ad implied. Not a generic benefit list.
- The first CTA matches the action the ad was setting up. If the ad was "see how it works", the LP CTA should not be "start free trial".
This is the cheapest fix in the diagnostic. Most founders can do it in a week. Conversion rate often moves 30-60% just from aligning these three points.
2. Landing page conversion math
The trap: thinking the LP is "fine" without computing the rate.
Most founders never actually benchmark their LP conversion rate. They look at "leads this month" and feel okay if the number isn't catastrophic. But "okay" hides huge variance.
The benchmarks:
- Cold paid traffic to demo or audit or lead: 1-3% is normal, 3-5% is good, 5%+ is excellent.
- Cold paid traffic to free trial signup: 2-5% is normal, 5-8% is good.
- Cold paid traffic to direct purchase (ecom): 1-2% is normal for first session, second-touch can do 3-5%.
If you're below the "normal" floor, the LP has structural issues that no ad creative will compensate for. Above the floor, you can optimize ads with confidence.
How to diagnose at the page level:
- Scroll depth heatmap. 50% of visitors should reach the first CTA section. If 80% bounce above the fold, the hero is broken.
- Time on page. Under 20 seconds on a sales LP means the message didn't land. Under 10 seconds means the page is wrong for the audience entirely.
- Form abandonment rate. If you have a form, what percent of users who start it complete it? Below 60% means friction is too high.
When ESA BIC Brindisi launched their incubator program, the funnel started at the LP level. Brand-new website built ground-up, single CTA, single message. Result: €9.30 per booking on first-cohort cold-traffic acquisition. The cost wasn't low because the ads were brilliant. It was low because the funnel converted what arrived.
3. Activation event clarity
The trap: optimizing ads on "leads" when the real metric is "qualified users".
Founders set their ad pixel to optimize for "form submit" or "signup". The platform happily delivers more form submits. Most of them are unqualified, and the founder doesn't realize because the optimization layer is celebrating noise.
The bridge that needs to be defined:
- What is the action that signals a user is real and valuable? Not a form. Not a signup. Something downstream — a logged-in session, a first action in the product, a meaningful page view, an email reply.
- Can your pixel optimize on that event? If yes, switch the optimization layer to it. If no, you're optimizing on noise.
For SaaS, this is "activated trial user". For ecom, it's "first purchase" (or sometimes "second purchase within 30 days" — the better predictor for retention). For B2B service, it's "qualified discovery call" (post audit-funnel filter).
The shift is profound. Same ad spend, suddenly delivering 2-3x more qualified users, because the platform's algorithm has a more honest objective. No new creative needed. No new audience.
4. Tracking integrity
The trap: assuming the events are firing.
This is the silent killer. The founder says tracking is "installed". The marketing team is "optimizing". The CFO is asking why ROAS is declining. Six months in, someone opens the GA4 events panel and three of the five funnel steps don't fire correctly. The whole optimization was steering by broken instruments.
What to verify:
- Open the Real-time events tab of your analytics. Walk through your funnel manually. Each step should appear within seconds of you triggering it. If it doesn't, the event is broken or misconfigured.
- Cross-check your ad platform conversion totals against your analytics. They will never match perfectly (different attribution windows, different methodologies). But if they're off by more than 30-40%, one of them is lying.
- Audit your UTM convention. Are all paid links using the same UTM source, medium, and campaign structure? If not, your channel breakdown is fiction.
Ads on broken tracking is burn rate dressed as growth. If you've been spending €5K+ a month and never audited tracking end-to-end, that's the diagnostic step that pays back instantly.
5. Offer-market alignment
The trap: doing all four steps above and still not seeing scale.
If ads are aligned, the LP converts above benchmark, the activation event is honest, and tracking is clean — and growth still doesn't move — the problem is the offer itself.
Not the product. The offer.
- Pricing. Is the price right for the buyer's reference point? Software priced like consulting will struggle. Consulting priced like software will undercut its own value.
- Promise. Is what you say the buyer will achieve specific enough to be credible? "Boost your growth" is not a promise. "Reduce CAC by 30% in 90 days" is a promise — and it's testable.
- Proof. Do you have one concrete, verifiable proof point that the promise has been delivered before? Not testimonials. Not logos. A specific number applied to a specific situation.
If the offer is misaligned, no funnel fix will scale you. This is the moment to stop optimizing channels and renovate the proposition.
Most founders never reach step 5 in the diagnostic because steps 1-4 absorb all their attention. That's fine. Step 5 is where the strategic conversation lives. It usually requires positioning work, customer interviews, a real audit. It's not something you fix on a Monday.
What founders ask me
"How do I know which step is broken?" Run the steps in order. The first one that fails honestly is the one to fix first. Almost always 1, 2, or 4 — in that frequency.
"Can I just A/B test my way out of this?" You can A/B test creative endlessly without finding the answer. A/B testing presumes the funnel is working and just needs optimization. When the funnel is leaking, testing inside it is noise.
"What if my agency tells me the ads are fine?" They might be right. Agencies are responsible for ads, not the funnel. A good agency will tell you "ads are fine, the funnel is the bottleneck". A defensive agency will keep pointing at creative. The behavior signals which kind you have.
"How long does this diagnostic take?" A week to walk through it carefully. A month to fix the top two issues. A quarter to compound the results.
Where to start
If you're spending more than €5K a month on ads and the math has been getting worse, take the audit. Five minutes, auto-qualifies fit. If we're a fit, we run this diagnostic together. If we're not, you get the Founder Playbook and you move on with something useful.
Ads optimization fixes the symptom. Funnel diagnostic fixes the cause. The first is a 30-day cycle. The second is a 90-day investment.
Which step in your funnel hasn't been audited in the last six months?