The funnel behind a single order: viewed, carted, checked out, bought
Aggregate funnels tell you the drop-off rate. Per-order funnels tell you what the drop-off actually looked like for the people who came back.
Every analytics tool shows a funnel: so many viewed a product, so many added to cart, so many started checkout, so many bought. It is a useful shape and a poor diagnosis, because it averages over people whose behaviour has nothing in common.
The same funnel, per person
When the funnel is attached to a resolved customer, you can see the sequence for the people who eventually bought - and, more usefully, for the ones who did not.
- Someone who added to cart three times over two weeks before buying is not the same as someone who added once and bought immediately, even though both count as one conversion.
- Someone who reached checkout and left is a different problem from someone who never got past the product page.
- Someone who viewed six products and carted none has a discovery problem, not a checkout problem.
Where it changes what you do
A store looking at an aggregate funnel sees "cart-to-checkout drop-off is 60%" and starts redesigning the cart. A store looking at per-person funnels often finds the drop-off is concentrated in a segment - one traffic source, one device, one country - and the cart was never the problem.
The funnel is also what makes recovery specific. A message to someone who reached checkout should not be the same as a message to someone who only browsed, and knowing which they were is the difference.
What to do next
Take last month's abandoned checkouts and split them by whether the person had visited before. First-time abandoners and returning abandoners usually need completely different messages, and most stores send them the same one.
Read next
What they looked at before they bought something else