Queue Grid Method
Our flagship. Six labs on capturing tap order, drawing drop-off, and running a critique week without theatre.
Kamala Beach studio · since 2021
Read the order people actually tap.
Dashboards count screens. Screen Flow Analytics reads the queue between them — the stubborn backtracks, the polite dead-ends, the one extra confirmation that quietly ends a session.
A note from the desk
Most teams we meet already own a replay tool and a funnel chart. What they lack is a shared language for sequence. Someone says “users drop on step three,” and three people mean three different screens. Systemqueuegrid exists to slow that conversation down. We teach a grid notation so a researcher in Bangkok, a PM in Chiang Mai, and a contractor on a ferry can point at the same tap order without arguing about labels.
The work is unglamorous on purpose. You will spend more time annotating twelve sessions than watching a hundred. You will leave some heatmaps unread. That is the point of Screen Flow Analytics as we practice it here: fewer charts, a clearer queue.
On the calendar
Courses currently taught
Replay Annotation Hours
A shorter desk for teams who already record sessions and still argue about what they saw.
Mobile Transition Lab
Screen Flow Analytics for apps where the back gesture, the sheet, and the push all count as different screens.
How we teach
Four habits, practiced until they stick
-
Name the queue, not the page
Every flow is written as an ordered grid. Titles come second. If two screens swap order, the grid changes even when the screenshot looks familiar.
-
Twelve sessions before a chart
We refuse to open a funnel until twelve annotated replays exist. The number is arbitrary on purpose — it is enough to embarrass a premature slide.
-
One limitation, spoken aloud
Each critique must name what the recording cannot show: offline taps, shared devices, a spouse reading over a shoulder. Screen Flow Analytics is honest about its blind spots.
-
Write the next experiment in the same notebook
A finding that does not produce a testable change to the grid is filed as a hunch, not a result.
18seats in a Phuket cohort
63alumni polled after Kamala desk ’25
4.2studio score from that poll
38recorded critique sessions on file
The Queue Grid Method did not make our Mixpanel cleaner. It made the argument in stand-up shorter, which was the actual problem.— Nalinee K., payments PM, Bangkok
One alumnus asked us to note that the cohort homework still collides with release weeks. We agree. The desk is not a light evening course.
Journal
Recent writing
When the heatmap lies about sequence
A bright cell is not a decision. Notes from a checkout audit where the hottest screen was the one nobody finished.
Checkout screens we keep seeing in Thailand
PromptPay, OTP, and the extra “are you sure” step — a field note, not a national stereotype.
Questions about a team residency or a single seat? Write in English; we reply within two working days, Bangkok time.