Funnel Visualization
Funnel visualization turns step-by-step conversion data into charts — Sankey, step-funnel, retention curves — that make drop-offs and optimization opportunities obvious.
Funnel Visualization
Charting how visitors move through conversion steps so drop-offs, leaks, and branching paths become visible at a glance.
Funnel visualization is the practice of rendering step-by-step conversion data as a chart — most commonly a step-funnel bar chart, a Sankey diagram, or a cohort retention curve. Each format makes a different pattern visible: step-funnels expose the largest absolute drop, Sankey diagrams reveal branching and side-exits, retention curves show how cohorts decay over time.
The chart you pick frames what your team will optimize. A Shopify checkout shown as a four-bar step-funnel will surface the shipping-address page; the same data as a Sankey will surface that 18% of carts re-route to the login page first. Both are true, only one is actionable for the test you want to run.
Funnel visualization sits inside the broader practice of funnel analytics — the measurement layer — and feeds funnel optimization, the experimentation layer that acts on what the chart reveals. The visualization is the hinge between the two: without it, drop-off numbers stay as rows in a table that nobody scans.
Three chart types cover ~95% of e-commerce use cases. Step-funnel charts compare absolute and relative conversion between sequential steps. Sankey diagrams show how traffic splits and re-merges across non-linear paths. Retention curves plot the share of a cohort still active at day 1, 7, 30, and 90 — essential for subscription and replenishment-driven stores.
step_conversion_rate = users_entering_step_n+1 / users_entering_step_n
users_entering_step_n
Step n entrants
Unique users (or sessions) who reached the start of step n.
users_entering_step_n+1
Next-step entrants
Unique users who successfully advanced to step n+1.
step_conversion_rate
Step conversion rate
The share of users who continue from one funnel step to the next; the height of each bar relative to the previous one in a step-funnel chart.
A Shopify apparel store tracks a four-step checkout funnel over one week.
Cart viewed: 12000
Checkout started: 7200
Shipping submitted: 5400
Order placed: 3960
→ Step rates: 60% → 75% → 73%. Overall cart-to-order: 33%.
The cart-to-checkout step (60%) is the weakest link in absolute terms — a step-funnel chart will surface this immediately. Shipping-to-payment (73%) is the next priority despite being later in the funnel.
Picking the chart type is a decision about what question you're answering. Use a step-funnel when the path is linear and you need to rank drop-offs by size. Switch to a Sankey when users can skip, loop back, or arrive at the same step from multiple sources — typical for stores running pop-up offers, login walls, or guest-vs-account checkout splits.
Choosing the right funnel chart by use case
| Chart type | Best for | Funnel shape | Typical e-commerce use |
|---|---|---|---|
| Step-funnel (bar) | Ranking drop-offs by size | Linear, 3-6 steps | Checkout flow, signup, product → cart → purchase |
| Sankey diagram | Branching paths and side-exits | Non-linear, many paths | PDP → cart vs PDP → search vs PDP → exit |
| Retention curve | Cohort decay over time | Time-based, single cohort | Subscription re-order, replenishment, app re-open |
| Heatmap funnel | Funnel × segment matrix | Two-dimensional | Conversion by traffic source × device |
| Horizontal flow | Long funnels on dashboards | Linear, 6+ steps | B2B-style lead funnels, multi-page configurators |
A common failure mode: defaulting to a step-funnel for every analysis. If 30% of your checkout traffic enters at step 2 (express checkout, Shop Pay) and the rest at step 1, a step-funnel will quietly average them and hide the real drop-off pattern. A Sankey or a segmented step-funnel surfaces it. The visualization choice is itself an analytics decision worth a few minutes.
Funnel visualization FAQ
A step-funnel chart assumes a single linear path and shows the share of users continuing at each step. A Sankey diagram allows branching, looping, and side-exits — it shows where users go, not just whether they continued. Use step-funnel for checkout flows, Sankey for browsing or multi-entry-point funnels.
When the conversion event repeats over time rather than ending at a single step. Subscription re-orders, replenishment purchases, and account re-activation are retention questions, not funnel questions. A retention curve plots the cohort's surviving share at days 1, 7, 30, 90 and exposes decay patterns a funnel chart can't show.
Three to six steps is the practical range for a step-funnel chart. Fewer than three rarely justifies the chart; more than six compresses the bars to the point where small but important drop-offs become invisible. For longer flows, split into two linked funnels or switch to a horizontal flow layout.
GA4's Funnel Exploration is a step-funnel by default and offers a 'Path Exploration' report that is Sankey-like but limited in depth and segmentation. For full Sankey visualizations across non-linear DTC checkout paths, most teams export GA4 data into a dedicated tool or use a CRO platform with native Sankey support.
Use a segmented step-funnel — the same funnel rendered as multiple parallel bars per step, one per segment. This makes it obvious that, for example, mobile drops 12 points harder than desktop at the shipping step. A heatmap funnel (steps × segment matrix) works for 5+ segments where stacked bars get noisy.
Defaulting to a step-funnel for non-linear flows. If users can enter mid-funnel (express checkout, deep links, returning carts), a step-funnel will under-count the first steps and make the drop-off look smaller than it is. Either segment by entry point or switch to a Sankey.
Show both, but rank by absolute lost users. A 5-point drop at a step with 50k visitors is a bigger opportunity than a 20-point drop at a step with 800 visitors. Most platforms display the percentage by default — annotate the chart with absolute counts so prioritization conversations stay grounded.
The visualization is the diagnostic step. It tells you which step to test, what segment is worst affected, and roughly how much volume is at stake. Funnel optimization then turns that into hypotheses and A/B tests. A bad visualization choice produces good-looking but low-impact tests.
You need at least a few thousand sessions per step for the visualization to be stable week-over-week. If you're under that, either widen the time window (4-8 weeks rolling) or import historical GA4 data so you have a baseline from day one rather than waiting for fresh data to accumulate.
Match the window to the user's likely time-to-convert plus a buffer. For impulse-purchase apparel, a 1-7 day window is fine. For considered electronics, use 14-30 days. Too short and you mislabel non-purchasers as drop-offs; too long and seasonality blurs the steps.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.