BEH / Purchase journey
The checkout was not one funnel
Mapping the wider purchase ecosystem turned fragmented journeys, product choices, and renewal paths into a defensible experiment backlog.
Signal trail
- 1
Product demand
Concentrated
- 2
Journey paths
Fragmented
- 3
Renewal cohorts
Actionable
The brief
The question behind the work
A confidential consumer subscription business offered several plan structures, optional extras, and multiple paths into checkout. Product and lifecycle teams needed a clearer view of which choices mattered and where the evidence was too weak to act.
“Which parts of the product offer and purchase journey deserve attention first—and which apparent opportunities disappear once the underlying data is challenged?”
Evidence available
- Buyer, non-buyer, and renewal journeys in GA4
- Order, product, and license records at several levels of detail
- Plan mix, add-on demand, fulfillment choices, and renewal settings
- Landing pages, referrers, and acquisition-source signals
How I worked
From raw evidence to a defensible answer
- 01
Fix the unit of analysis
Separate events, order items, orders, and licenses so that one commercial action is not counted several different ways.
- 02
Map demand and choice
Compare core plans and optional products on a consistent window, then look for concentration, long tails, and choices that may need clearer framing.
- 03
Reconstruct the journey
Segment purchasers, non-purchasers, and renewers to see which pages, properties, and sources appear before a completed order.
- 04
Build comparable renewal cohorts
Separate renewal configuration from realized renewal and only compare customers who had a full opportunity to renew.
- 05
Turn patterns into tests
Size the opportunity directionally, document uncertainty, and define experiments with a primary metric and guardrails.
What the evidence said
The findings that survived validation
Demand concentrated around one low-friction core plan, with meaningful but smaller demand for flexible and group-oriented alternatives.
Educational extras complemented the core offer more clearly than a broad catalogue of generic upsells.
Customers entered high-intent journeys from several owned surfaces, making a single checkout funnel an incomplete model.
Renewal settings varied by product and market, but opt-in status could not be treated as realized retention.
Decisions
Signal → interpretation → move
A steep product long tail
Interpretation
More choice did not automatically create more useful choice; some tiers needed clearer differentiation.
Business move
Test plan framing and simplify low-signal choices before changing the underlying offer.
Several checkout entry points
Interpretation
Commercial intent started outside the main marketing journey, while attribution remained incomplete.
Business move
Treat checkout as a cross-property ecosystem and repair identity and source measurement first.
Different renewal patterns
Interpretation
Configuration, eligibility, and actual renewal answered different questions.
Business move
Build mature cohorts and test explicit, benefit-led renewal communication with holdouts.
Delivery & limits
What the client could use next
Delivered
- Product-demand and choice diagnostic
- Purchase-entry and attribution map
- Renewal cohort definition
- Data-quality and KPI note
- Prioritized experiment backlog
Where the evidence stops
- Behavioral analytics was used directionally where attribution or consent made exact reconciliation impossible.
- Observed demand did not explain why customers chose a plan or add-on.
- No experiment implementation, revenue lift, or retention uplift was documented, so this is presented as diagnostic work.
Your decision
Have a question your current dashboard cannot answer?
Let’s find the right grain, evidence, and test before another metric becomes the strategy.
Start a conversation