The Funnel Signal Method
A close-reading studio for Support Funnel App Analytics: in-app help, deflection, ticket birth, and the quiet exits nobody graphs.
Read the syllabusMost product teams stare at volume charts and call it Support Funnel App Analytics. We teach the slower skill: naming each hand-off inside the app, writing down who disappears, and arguing about the sentence — not the sparkline.
A close-reading studio for Support Funnel App Analytics: in-app help, deflection, ticket birth, and the quiet exits nobody graphs.
Read the syllabusEvent naming, dead-end articles, and the difference between a tap that looks like success and a person who actually found an answer.
Enquire for the next benchModule 2 forced us to print the in-app help tree on A3 and mark every article that ended in “contact us”. We had seventeen of those. The Funnel Signal Method did not make the queue prettier; it made the tree embarrassing enough to rewrite.
I came for dashboards and left with a vocabulary. The Signal Bench office hours were useful, though the live sessions assume you can spare a Wednesday morning — hard if you still take overflow chats yourself.
Identity Nodecore started because too many Thai and APAC product desks inherited a “funnel” view that began at ticket create. Everything upstream — search in the help center, bot loops, the forgotten in-app coachmark — was treated as folklore. We write it down.
The studio in Pathum Thani is small on purpose. Cohorts stay readable. Faculty still grade notebooks by hand. If you want a tool certification, you are in the wrong reading room.
A funnel that starts at “ticket opened” is already a confession that the app failed quietly.
You leave with a one-page Atlas: help search, article, bot, form, human, escalate. Each edge has an owner and a failure sentence, not a vanity conversion rate.
We teach how to isolate exits that look like “resolved” in the product analytics export but are really people giving up after the third unrelated article.
The last week is a written brief, not a slide dump. Leadership gets three claims, the evidence, and the one instrumentation gap you still cannot close.
Why completion rate on a help article is a weak proxy for whether the funnel actually held.
Continue readingA queue is not a backlog metaphor. It is a stage with entry criteria you probably never wrote down.
Continue reading