When you build a SaaS, you celebrate signups.
More users usually means growth… right?
That’s what I thought too.
But after running my own SaaS products, I discovered something most founders don't talk about:
Not all signups are real users.
Some are bots.
Some are trial abusers.
Some are competitors scraping data.
Some are just spam.
And they silently destroy your product metrics.
The problem I noticed
At first everything looked good:
Signups increasing
Trials increasing
Traffic growing
But revenue wasn't growing at the same rate.
That was the first red flag.
When I investigated deeper I found:
Disposable emails creating multiple trials
Bots creating accounts automatically
Fake users testing system limits
Spam accounts polluting dashboards
The biggest damage wasn't infrastructure cost.
It was bad decisions caused by bad data.
When fake users enter your funnel:
Conversion rates become inaccurate
CAC calculations become wrong
Product decisions become misleading
Growth experiments become unreliable
You think your product has a conversion problem.
But sometimes you have a user quality problem.
Why existing solutions didn't work
I tried:
CAPTCHA
Email verification
Domain blocking
Manual review
Basic fraud rules
They helped, but didn't solve the real problem.
Fraud today is not simple spam anymore.
Modern abuse includes:
Residential proxy IPs
Temporary email rotation
Human-like automation
AI-generated behavior patterns
Static rules fail because attackers constantly adapt.
The idea behind OnSefy AI
Instead of blocking users based on one rule, I wanted to measure risk signals.
The idea was simple:
Every fake user shows small warning signals. Together they form a risk pattern.
So I started building OnSefy AI around:
Risk scoring instead of binary blocking
Behavior intelligence
Infrastructure detection
Pattern recognition
Trust signals
The goal was not to stop signups.
The goal was to improve signup quality.
Early lessons while building
Some things I learned quickly:
Lesson 1 — False positives are worse than missed fraud
Blocking a real customer hurts more than allowing one fake signup.
Lesson 2 — Fraud prevention is a growth feature
Better users = better conversion data.
Lesson 3 — Most founders notice this too late
Usually after scaling problems appear.
Lesson 4 — Prevention is cheaper than cleanup
Stopping fake users early saves support, infra, and analytics headaches.
What we are building next
Current focus areas:
Real-time fake signup detection
Trial abuse prevention
SaaS risk scoring API
Next roadmap:
Account takeover detection
Promo abuse detection
Payment fraud signals
Trust scoring system
Why I'm building in public
Building in public helps with:
Accountability
Feedback
Learning faster
Meeting other founders solving similar problems
And honestly, many founders are facing this problem but few talk about it openly.
A question for other founders
I am curious:
At what stage did you first notice fake or low-quality users affecting your SaaS?
Was it early or after scaling?
Would love to learn how others are solving this.
