What screen flow is and why it matters

In a mobile app, people do not move along one fixed path. One person opens the catalogue → a product
page → the cart → payment. Another: search → filter → a product page → back → a different product
page → the cart. A third leaves after the catalogue.

Screen flow visualises all those real routes from event data. It is a diagnostic tool: it shows
where people actually go, rather than where the designer assumed they would.

How to build and read a screen flow

Data sources

Every transition between screens is recorded as a screen_view event:

screen_view → { screen_name: "catalog", previous_screen: "home" }
screen_view → { screen_name: "pdp", previous_screen: "catalog" }
screen_view → { screen_name: "cart", previous_screen: "pdp" }

Analytics tools — Amplitude, Mixpanel, Firebase — turn those events into a transition graph: nodes
for screens and edges for transitions, each carrying its share of the traffic.

The key metrics in the analysis

Metric What it shows
Drop-off rate The share of people who left the app from a given screen
Time on screen How long they spend on the screen
Exit paths Where they go from the screen — forward, back or out
Session depth The average number of screens per session

Typical findings in e-commerce apps

  • An unexpected exit from the delivery screen. People open the screen, see the delivery cost and
    close the app — a signal about the price, or about the absence of free delivery.
  • A catalogue → product page → catalogue loop. Someone returns from the product page to the
    catalogue several times — the page may not give enough information to decide.
  • A high exit rate from the cart screen. When more than 60% of sessions that reach the cart end
    in an exit without a purchase, that is the abandoned-cart funnel in mobile.
  • Direct entries through a deep link. People arriving from a push notification or an external
    link often take a much shorter path — worth accounting for in the analysis.

Tip: do not analyse screen flow in isolation from session type. Sessions with a purchase and
sessions without one produce fundamentally different navigation patterns — segment the data before
drawing conclusions.

How to apply it in practice

  1. Define the critical screens — catalogue, product page, cart, payment.
  2. Export the exit paths for each critical screen over a two to four week window.
  3. Find the abnormal drop-offs — compare against the previous period or against competitive
    benchmarks.
  4. Write hypotheses about the causes (experience, content, performance) and test them with an
    A/B test or session recordings.