Skip to content

Public funnel x-ray4 reads · 0 keys · ~5 seconds

Where your SaaS is leaking.The one play that fixes it.

No account. No key. I read your public surface and open Case #0001 in about 10 seconds.

See the $29 case this recovered →
anku.app/queue

Case #0001 · zoju.app

4 reads · 5.0s · 0 keys · [recorded 2026-10-29]

landing page

started with Stuck at the same band? Find out exactly why.

pricing

9 tiers found · starting $0/mo · free tier present

signup form

3 fields before the ask · email · password · confirm password

stack

Google Analytics · Supabase

I can read your promise. I cannot see your product.

Four reads open the case. The next three are what the connections add.

One stage is leaking.

Four public reads name the stage that leaks; three buckets stay unmeasured.

$29/mo

A paying user's card expired and the last payment failed.

  • Failed Cards

    $29/mo · 1 case measured

  • Fake Churn ≤3 actions

    not measured

  • Pre-Cancel Drift

    not measured

  • Job Done / Wrong Fit

    not measured

One play. One approve.

One note, written from the case's own facts, waits for your approval.

To

n•••@example.com

Your Zoju card expired

Hi there — your card on file for Zoju expired, so the last payment didn't go through. Nothing is lost: the plan and the history are untouched.

Review & Approve 1 Case ($29/mo) →

The receipt, not a report.

Verified money on the left, modeled risk on the right. Never blended.

$29
in_test_1Q4x7kR2evt_test_9m2v6q0h

Zoju · consented demo data · PostHog + Stripe

No row without a source event id.

PostHog, Stripe, and Resend on one line.

The failure, the retries, the note and the payment, in order.

18 Oct

Invoice failed · expired_card

Stripe

25 Oct

Smart retry #3 failed

Stripe

28 Oct

Play staged · payment recovery

Anku

29 Oct

Channel connected · ready

Anku

Tells you to stop when software isn't the fix.

When the funnel is not the problem, the case says so and stands down.

When the problem is not the funnel, it says so.

Stop optimizing; go talk to users.

  • 20 minutes of your experience with Zoju?
  • What did you switch to?
  • The weekly check runs next Monday.

Manual until 5 sends pass with 0 complaints.

Nothing sends by itself until five verified sends and zero complaints.

Manual (1/5)

5 verified outcomes and 0 complaints unlock auto-run.
Today every send waits for you.

One payment came back. 281 people were in the window.

[EXAMPLE] · captured 2026-10-29

consented test event

$29

recovered

verified · reconciled

$0

at risk

1

receipt

1

case

The card expired on 18 October. Three retries failed, the last on 25 October. One note went out on 29 October, and the payment landed the same day. Stripe confirmed it.

Case0b8e2f1aInvoicein_test_1Q4x7kR2Eventevt_test_9m2v6q0hCustomercus_test_8f3a2b7cPersonn•••@example.com

Zoju · consented demo data · PostHog + Stripe

Who is leaking.

5 of 281 people in the window

Five left a record long enough to name: no practice test, 20 seconds or more. The other 276 are a count.

  • n•••@example.com6 events · 47s
  • person 1c9d3e4f4 events · 52s
  • n•••@example.com3 events · 31s
  • n•••@example.com8 events · 64s
  • person 4c9d7e8f2 events · 39s

Names are masked at the source. Two of the five have no email on record.

What does “not measured” mean?

It will not guess a cause.

Three of the four cause buckets are not measured. Each one names the rung that would split it.

Failed Cards

measured

Fake Churn ≤3 actions

not measured

Pre-Cancel Drift

not measured

Job Done / Wrong Fit

not measured

What happens when it is not the funnel?

When the problem is not the funnel, it says so.

The people who stay do not stay longer over time. This is not a funnel problem. Software cannot fix this, including me.

The cohort curve stays watched. The weekly check runs next Monday.

What can it do without asking?

Nothing sends without your approval.

Auto-run needs 5 verified outcomes and 0 complaints. Today every send waits for you.

Autonomy: Manual (1/5)

What does it count?

It counts one thing: money Stripe confirmed.

No row without a source event id. Verified money on the left, modeled risk on the right. Never blended.

$29

not measured

I ran this on Zoju first. 21,000 events a week, real churn. I moved activation from 19% to 44%, and the company still died. The wall was never the funnel. That is why the case says no.

Three rungs, three depths.

1 of 3 connected

Each rung buys a deeper truth. One at a time, read-only, revoked in one click.

01

PostHog · read-only

The aha: 61% of the people who reach the first practice test come back. 3% of the people who don't.

02

Stripe · read-only

The dollars: $29/mo at risk, priced from billing. This invoice is a consented test event, not a live key.

03

Auth + DB · read-only

The names: 276 of the 281 people in the window are still counts.

See the connection ladder →

Write scopes are only asked for from inside a play.

One flat price. No surprises.

Free

$0

your case, free · no card

Open a case on your own product from one URL: the funnel read, the blind spots named, the honest close when it isn't a funnel problem. No account. No key. No card.

Open the door →

Pro

$49/mo

the recovery loop · opening when your data joins

The connected case: who is stalled, what it is worth, one play at a time approved by you, receipts verified in Stripe. Plus the optional weekly Monday watch.

Pro is not on sale yet. Request access and we email you when your data can join. No card is taken while Pro is closed; when it opens, cards are entered only on Polar-hosted pages.

Compare plans →

Find the leak in four reads. Then I do the work.

No account. No key. No card.

Anku — Find the leak in four reads