Your Shopify store is not losing traffic. It is losing revenue between the clicks.

A low conversion rate describes an outcome across all visitors, including those who were never going to buy. A revenue leak is narrower and more actionable: it describes visitors who demonstrated intent — a cart, a question, a checkout start — and were lost without any response. Conversion problems are fixed by changing the offer; leaks are fixed by adding a response.

You can locate them, but not size them accurately. Shopify gives you stage counts, which is enough to find the widest drop. What it does not give you is the reason: whether the customer hesitated on price, waited for an answer, or hit a payment failure. Locating a leak takes an afternoon; understanding it requires behavioural context alongside the counts.

For passive email-only recovery, 3–7% is typical. For timely multi-channel recovery with a message that references the actual cart and reaches the customer within the first hour, 10–20% is achievable. Anything advertised above 30% is either measuring assisted conversions generously or counting customers who would have returned anyway.

In the first hour, yes. Recovery probability decays steeply with time because the purchase decision moves out of working memory and, often, into a competitor's cart. Content determines whether a timely message converts; timing determines whether the message matters at all. Optimize timing first, then content.

Read rates and response latency. An email recovery sequence competes with a promotional inbox and is often opened hours later, if at all. A message on a channel the customer uses for personal conversation is typically read within minutes and can carry a real exchange — a question answered, a size confirmed — rather than a one-way reminder.

Because those tools do not share what they know. The chat widget does not tell the email flow that the customer asked about delivery time; the email flow does not tell the widget that a discount was already offered. Each tool addresses one symptom without a shared view of intent, so the customer receives fragments instead of a coherent response.

Take the lost sessions at that stage over 90 days, multiply by average order value, apply a conservative recovery rate of 10–15%, then multiply by four for an annual figure. Use the conservative rate deliberately: a defensible number that survives scrutiny is more useful internally than an optimistic one that collapses in the first review.

The one with the highest recoverable value that can be evidenced within 30 days. The biggest leak is often structural and slow to close, and the easiest is often trivial in value. Ranking by recoverable value divided by time to evidence keeps momentum while still moving real revenue.

They close some and create others. A discount can recover a price-hesitant cart, but applied broadly it trains customers to abandon deliberately and erodes margin on revenue you would have earned at full price. Discounts should be a targeted response to a diagnosed price objection, not the default recovery mechanism.

Detection produces findings within days, because the data already exists. Recovered revenue typically becomes measurable within two to four weeks, once enough recovery attempts have accumulated to separate signal from normal weekly variance. Anything claiming meaningful recovery in the first week is measuring noise.