Signals
Metric shifts Tamda detected in the last 7 days
Suggested next experiments
Sticky mobile checkout bar with a collapsed promo field
purchase_completed dropped 13.6% last week and has never been the target of an experiment in this workspace.
Mobile web keeps showing up across your signals
mobile_web is the top driver in 3 of the last 5 signals, across purchase_completed and add_to_cart. That is a systemic mobile issue, not a one-off.
purchase_completed
The drop is almost entirely mobile web. On small screens the promo-code field sits directly above the Pay button and pushes it below the fold, so shoppers who reach payment on a phone lose sight of the primary action and abandon. iOS app and desktop, where the layout differs, held flat.
Driven by mobile_web (platform), 82% of the change
Sticky mobile checkout bar with a collapsed promo field
Page: mobile web checkout. Control: current checkout with the promo-code field expanded above the Pay button. Treatment: collapse the promo field behind a "Have a code?" link and pin a sticky bottom bar showing the order total and a full-width Pay button that stays visible as the page scrolls. Primary metric: purchase_completed. Guardrails: checkout_started, add_to_cart, average order value (so a hidden promo field does not quietly suppress discount use or shrink baskets).
add_to_cart
The decline is concentrated in paid-social traffic, which currently lands on the homepage rather than the specific product from the ad. The extra hop to find the advertised item adds friction before anyone can add to cart. Organic and email traffic, which lands deeper, were flat.
Driven by paid_social (utm_source), 79% of the change
Route paid-social clicks to the advertised product page
Page: paid-social landing. Control: all paid-social ads deep-link to the homepage. Treatment: deep-link each ad to the product page for the item it features, preserving UTM parameters. Primary metric: add_to_cart. Guardrails: bounce rate and purchase_completed, to confirm the deeper entry point does not raise exits or shift the funnel without lifting purchases.
signup_completed
A modest, encouraging lift, concentrated in first-time visitors from organic search. It lines up with last week’s change that moved the account prompt below the first purchase step, so new shoppers see value before being asked to sign up. Worth confirming it holds, then codifying.
Driven by first_time (visitor_type), 88% of the change
Confirm the deferred sign-up prompt holds, then make it default
Page: post-first-purchase sign-up prompt. Control: the previous flow asking for an account before checkout. Treatment: keep the deferred prompt that asks after the first purchase completes. Primary metric: signup_completed. Guardrail: purchase_completed, to confirm deferring the account step did not cost any sales.