YOUR SAAS CONVERSION RATE IS DROPPING. THE PROBLEM ISN'T
IN YOUR FUNNEL.
|
Summary
In this post
Latest insights. Straight to your inbox
01 – THE FIX THE FUNNEL INSTINCT
WHY THIS PATTERN MATTERS MORE THAN YOU THINK
When a SaaS conversion rate starts dropping, the instinct is to fix the funnel. Most of the time, that instinct is wrong.
Three months ago, I was watching Aimfox’s free-to-paid conversions soften. Not collapse — just slow. We were still growing. But the growth was running a few percentage points below where it had been in November and December.
The obvious move was to look at the funnel. Review the onboarding sequence. Tighten the in-app messaging. Run the numbers on every conversion step and find where users were dropping off. That is what you do when conversions slow down.
Except the funnel wasn’t broken. The problem was upstream — and it took a specific diagnostic to see it.
The SaaS instinct when conversion softens is to optimize the conversion layer. That is a reasonable instinct. Most of the time it is also the wrong one.
I have worked with 300+ teams on positioning and growth. The pattern I see repeatedly is this: teams spend months improving onboarding, email sequences, and in-app flows, while the actual problem — what kind of user is entering the funnel in the first place — goes untouched. The funnel gets more polished. The numbers stay flat. Nobody asks the upstream question.
Work with Andrej
Is your growth stalling?
A quick discovery call to find out if positioning is the problem.
if your onboarding hasn't changed but your conversion rate is dropping, stop optimizing the onboarding.
02 – WHAT’S ACTUALLY GOING ON
THE REAL REASON THE NUMBERS WERE DROPPING
When I ran the analysis on Aimfox’s free trial cohort, I was expecting to find the standard drop-off pattern. What I found instead was a bimodal audience.
Roughly half the users were engaging exactly the way you want free trial users to engage. Email open rates between 30 and 45 percent. Strong in-app behavior. Responses to triggered messages. These users were converting at the rate we expected.
The other half were completely dark. Zero email opens across the entire welcome sequence. No in-app activity. No responses to anything. As if they had never signed up.
My first assumption was bots. It turned out to be something more specific — and more damaging.
A competitor had been injecting dummy traffic into our Google Ads campaigns. Not at a rate that would trigger an obvious alarm. The invalid traffic share climbed gradually: from roughly 20 percent of signups, to 25, to 30, eventually reaching around 50 percent. Over months. Slow enough that at any given moment, the numbers looked like normal variance.
But the damage was not just wasted spend. The deeper problem was what this did to Google’s Smart Bidding algorithm. The algorithm trains on conversion signals in near real-time. When fraudulent traffic converts, it feeds those signals into the lookalike model. The algorithm starts recruiting audiences that look like the fraudulent users. The refunds Google issues restore the budget; they do not restore the audience data. According to Search Engine Land, the average invalid click rate across Google Ads is now 11.4–12.3 percent. That is what gets detected. The damage to audience modeling happens before detection.
03 – the reason
LET'S TALK REAL AD NUMBERS
The reason this is hard to catch is not that the data is hidden. It is that the data looks fine at the level most teams monitor.
We were growing. CAC looked within range. The conversion drop was gradual enough to absorb into the noise. And when conversions soften slowly, the rational response is to improve the thing you can control. That is a reasonable, defensible decision. It just happens to be aimed at the wrong layer.
There is also a structural reason this kind of audience degradation happens at the $5M–$10M ARR stage specifically. At that stage, most teams have found one or two channels that work, and they scale them. That concentration is not a mistake — it is how you build momentum in the early growth phase. But it creates a single point of failure.
Gaurav Agarwal, COO of ClickUp, has documented this principle publicly. His channel portfolio framework recommends balancing high-risk/high-reward channels against more predictable ones — the logic being that every growth motion eventually decelerates, and companies that do not build a second motion before the first one fades will feel the loss.
At Aimfox, we had not revisited our channel mix after the initial scale. We knew the theory. We still got caught.
After analyzing 1M ads, we found 6.34% of ads were making a profit
04 – The pattern
DIAGNOSTIC SIGNALS TO CHECK TODAY
01
Segment your free trial cohort by email engagement.
02
Cross-reference the dark segment against traffic source.
03
Look at your in-app behavior by cohort.
04
Check your invalid click rate directly.
05
Look at the trend, not the snapshot.
05 – The fix
WHAT TO DO ABOUT IT
Audit audience quality before auditing conversion mechanics. The first question is “where are users dropping off? If you can’t find the drop of, the question is “what kind of user is entering this funnel?”
Build invalid traffic monitoring into your regular reporting. The invalid click column in Google Ads should be a standing metric in your weekly or bi-weekly review.
Add a second channel before you need one. If 20 percent of your growth is coming from a second channel with clean audience signals, you notice the degradation in Channel One earlier. I wrote about the AI-educated buyer pattern reshaping SaaS funnels — the same principle applies here.
When you do have an audience problem, use The Conversion Gap process — not a standard funnel audit. A funnel audit assumes the problem is in the conversion mechanics. The Conversion Gap diagnostic maps the full funnel including acquisition source.
Work with Andrej
Is your growth stalling?
A quick discovery call to find out if positioning is the problem.
06 – THE NEW FUNNEL
WHAT AIMFOX LOOKS LIKE NOW
We stopped optimizing onboarding. We started auditing the campaigns.
Within a few weeks, we restructured the most contaminated campaigns, tightened ICP targeting in the keyword groups most exposed to competitor interference, and added negative exclusions to reduce the surface area for dummy traffic.
The engaged segment — the half that was converting well even during the contamination — told us the funnel was not the problem. The fix was not to improve the funnel to the average of both segments. It was to remove the contaminated segment and protect the signal.
We are still in the middle of this. Growth is recovering. But the more durable lesson is about channel concentration: a gradual attack is designed specifically to stay below the threshold of the metrics you are already watching.
Your conversion rate tells you what is happening. Your audience data tells you why. Most teams only look at one.
