Funnel Evolution — Rise 2 & sitewide

The diverging steps

Each funnel step as a share of Rise 2 visitors, window by window. The red and purple lines are mirror images from May 22 on — that crossover is the rerouting. The green line staying flat is what exonerates the product pages. The solid black line (white in dark mode) is blended CAC (right $ axis) and the thin amber line is Rise 2 CVR — the correlation on one canvas: the routing peak window is $91 CAC / 2.14% CVR, the trough $125 / 1.54%. Bars along the bottom: A/B tests launched in each window; amber shading marks windows with Rise-2-specific tests (Rise 2 / Super LP / Switcher) live.

Lines = % of all Rise 2 visitors whose second page was a PDP / another content page, and add-to-cart rate per PDP visitor. Bars (right scale, count printed above each) = tests launched per window from the Intelligems span list. Amber band = a Rise 2 URL test was live at some point in the window. Blended CAC = Daily Spend Tracker spend ÷ Shopify new customers, computed from complete days only — the Jun 5–11 window has just 3 complete days (asterisked in the tooltip). Rise 2 CVR is drawn on its own auto-scaled hidden axis so the 1.5–2.2% line isn’t flattened by the 0–28% step axis; the tooltip shows real values with units.

Canonical page only — was the collapse real?

Step table — all seven windows

Sessions by path — the two roads out of Rise 2

Every engaged Rise 2 visitor takes one of two roads: straight to a product page, or to another content page first. The left pair of columns is the mix shift — that’s what moved. The right pair is each road’s add-to-cart rate — that’s what didn’t.

Share of all Rise 2 visitors routed down each path, per two-week window. Amber band = Rise 2 URL tests live, same convention as Chart A.

Canonical page vs redirect variants — the clean split

Shane’s theory: the redirect-variant pages themselves convert worse, and they’re dragging the blended Rise 2 numbers down. Here’s the blended bucket split in two — the canonical page on the left, everything the redirects created on the right.

Three segments, four metrics — the honest split

What June looks like with April’s routing

Where the second click goes

Share of Rise 2 visitors by second page — non-bounce paths sum to the engaged share. The bounce band is shown muted so the mix shift is honest: bounce got smaller while the purple content band ate the red PDP band. People stayed on the site; they just stopped being sent to product pages.

“Other engaged path” = 100% − bounce − PDP − content (collections, cart, account, search… — Intelligems does not split these further in this view). All bands are raw shares of total Rise 2 visitors.

Downstream health

The bottom of the funnel held all quarter — which is exactly what you’d expect if the damage is upstream rerouting rather than purchase intent. One amber flag: checkout→buy dipped in the final week.

Sitewide, same story