Frame it: in a single day, 82% is either (a) a real conversion break, (b) a numerator/denominator definition change, or (c) a traffic mix shift that makes the old rate un-comparable.

Diagnostic sequence, top to bottom:

Rule out instrumentation. Did a release ship yesterday? Did the add_to_cart or purchase event schema change? If add_to_cart fires more aggressively (e.g., on hover) without purchase moving, abandonment mechanically spikes. Check event volumes, not just the ratio. This is the single most common cause.
Rule out a funnel break. Walk the checkout funnel: cart → shipping → payment → confirmation. Which step's drop-off moved? If it's concentrated at payment, suspect a payment-provider outage, a new Stripe/Braintree rule, or a CSP change blocking the iframe. If it's at shipping, suspect a new address-validation error.
Traffic mix shift. Did paid traffic spike? Paid has much higher abandonment than organic/returning. A Black-Friday-style campaign sending cold traffic to the cart page will mechanically inflate abandonment with no actual bug. Segment by channel and by new-vs-returning.
Device and browser split. A Safari/iOS update breaking Apple Pay, a Chrome change to SameSite cookies, or a specific extension — these appear as a sharp cliff in one segment and business-as-usual in others.
Geography / currency / fraud rule. New fraud rule rejecting more cards? New region launched with lower baseline conversion? Price-localization bug showing the wrong currency?

What I'd deliver by EOD:

Funnel conversion by step for the last 14 days, yesterday highlighted.
Segmentation by channel, device, geo.
Event-volume sanity check against the most recent release.
Named hypothesis with the single strongest supporting cut.

Red flags in an answer: jumping straight to "users don't want to buy anymore" or proposing UX experiments. A 14-point one-day move is almost always instrumentation or infrastructure, not preference.

Cart abandonment jumped from 68% to 82% overnight. Diagnose.

Draft your answer

Stuck? Peek at a hint

More product metrics questions