Drop-Off Rates Tell You Where. Funnel Forensics Tells You Why.
If your form or funnel builder stops at step drop-off rates, you are missing most of the story. Funnel Forensics tracks the events and behavior behind every step, so you know what made people leave and what to fix first.
Same step, same visitors. One number says where. The rest says why.
A Drop-Off Rate Is a Symptom, Not a Diagnosis
Every builder can tell you that 38% of visitors left on step 4. That number cannot tell you whether the question confused them, the phone field rejected their number, the page was slow on mobile, or the button looked disabled. So you guess, change something, and wait two weeks to see if you guessed right. Funnel Forensics puts the evidence next to the number.
- Which field people last touched before they left
- Which error messages your visitors were actually shown, and how often
- How mobile compares with desktop on the same step
- Each figure labelled with its source: every visitor, or the recorded sample
Counted for every visitor, on every step, without watching a single recording.
Event and Behavior Tracking on Every Step
Open any step and see what happened inside it. Which answers get picked and which get changed, which fields end the session, which validation messages fire, and how phone verification turned out. These tables count every visitor and work on every plan. Turn on Lead Playback and you also get the behavior only a recording can show.
- Answer clicks per option, and answers people changed away from
- Exit fields and the validation messages shown on each step
- Phone verification outcomes: sent, confirmed, abandoned
- From recordings: time per field, hesitation, rage clicks, dead clicks and taps on a disabled button
Every finding carries its count, its source and a next step.
Findings With a Count, a Cause and a Next Step
A fixed set of conversion checks runs on every funnel automatically: first-step drop-off, the worst step, mobile against desktop, time on page, confusion signals and landing page scroll. Your own edits are part of the evidence too, compared against that step for the 7 days before and after. Press Run forensics and the AI reads the last 30 days of sessions, analytics and edits, then writes up to five findings you can act on from the same card.
- Each finding shows its count, its source and whether it is new, worsening or improving
- Stage the suggested change, or set it up as an A/B test, from the same card
- When a save caused the shift, compare the two versions side by side
- Up to three example recordings per finding, so you can see it for yourself
The same evidence, wherever you work: the funnel page, your AI agent, or a PDF for a client.
Ask Your AI Agent, or Hand a Client the Report
The evidence is not locked inside a dashboard. Connect Claude, Codex, Cursor or any MCP client and your agent can read the forensics for a funnel, pull the full report in one call and tell you what to fix. For clients and teammates, the report lays out every step on one page for the period you choose, ready to print or save as a PDF.
- MCP access to findings, conversion signals and every step table
- Agents can take a fresh snapshot, with your confirmation
- A one-page report for any date range, print or save as PDF
- The same findings feed the built-in Copilot
Why drop-off rates alone are not enough
Two funnels can lose the same 38% on the same step for completely different reasons. One has a phone field that rejects valid numbers. The other asks a question people cannot answer yet. The fix for one does nothing for the other. You can only tell them apart by looking at what visitors did, not just where they stopped.
Where, and why
Drop-off per step, with the events and behavior behind it
Every visitor
Step tables count all traffic, not just a recorded sample
Your edits included
See what a save did to the step it changed
No recordings to watch
Hundreds of sessions reduced to the patterns that matter
Explore More Features
Stop Guessing Why Your Funnel Leaks
Start a free trial, send traffic to a funnel, and see the evidence behind every step: what people clicked, where they stalled, and what turned them away.
Start Free Trial