Funnel Analysis and Drop-off

By the end of this lesson, you should be able to: read a funnel for the step that moved rather than the step that's lowest, decompose a change in orders into volume and each conversion step with no leftover term, and find a leak that no single segmentation dimension can locate.

The funnel, and the step everyone points at

Eleven weeks of Basket, every event in the window.

Open to order, end to end: 18.48%.

Now, the question every funnel review opens with: which step should we fix?

Almost everyone answers search to add-to-cart. It's the worst step on the chart, losing 44.4% of the people who reach it, and it's where the eye lands.

That's the wrong answer, and understanding why is most of this lesson.

Fix the step that moved, not the step that's lowest

A funnel's step conversions are mostly a description of the product's shape, not a list of problems. Browsing is how grocery shopping works. Plenty of people search, see the price of coffee, and close the app. That step has probably lost 44% of users since the day Basket launched, and no amount of design work will change it much, because it isn't broken.

So don't read the level. Read the change. Split the window at 2 May and compare:

The worst step on the chart didn't move at all. Two other steps moved, and the largest by far is the one that looked healthiest: checkout to order was the best conversion in the funnel and it lost fourteen and a half points.