How do I find my biggest revenue leak?

Start by asking which of visitors, conversion rate, average order and repeat purchases has moved against itself. Not against a published benchmark, because those describe other people's catalogues, traffic mix and prices.

Put six months of each number side by side and look for the one that is flat or falling while the others hold. That single comparison usually narrows the search to one part of the business in about ten minutes.

Read the combinations carefully, because they mean different things. Traffic up and revenue flat means the extra visitors are not the same kind of visitors, which is a targeting or a landing page issue. Conversion steady and revenue down means fewer people are arriving. Orders steady and revenue down means average order fell, which usually traces to discounting, product mix or a shipping threshold change. Everything steady except returning customers means the retention work has stopped compounding.

Then check for the boring explanations before the interesting ones. Did a best seller go out of stock. Did a price change. Did the season turn. Did tracking break. Any of those will look exactly like a marketing problem in the summary numbers.

Write the conclusion as a sentence: revenue is short because conversion on mobile fell in the last eight weeks. That sentence is the difference between an investigation and a fishing trip.

Once you know which number is weak, break the journey into steps and find where people leave. The steps are landing on the site, viewing a product, adding to cart, starting checkout, and completing the order.

Calculate the ratio between each pair of steps and compare each ratio with the same ratio three and six months ago. What you are looking for is not the biggest drop, because the biggest drop is always at the top. You are looking for the ratio that has changed most against its own history.

Each step points somewhere specific. Landing to product view is a navigation, search and category page problem. Product view to add to cart is the product page: photography, information, price, availability, delivery time. Add to cart to checkout start is where shipping cost, thresholds and cart friction appear. Checkout start to order is payment options, form length, account requirements and trust.

Then convert the leak into money, because that is what decides whether it is worth fixing. Take the number of people who reach the step in a month, the number lost, and what an order is worth. A step losing a few hundred people a month on a fifty dollar order is a different priority from one losing thirty.

And check the steps that do not appear in a standard funnel: people who search on the site and get no results, people who hit an out of stock variant, and people who reach a page that errors on mobile.

Averages hide leaks. Before drawing conclusions, split the weak step by the four dimensions that matter, because a store wide average often looks unremarkable while one segment is badly broken.

Device is the first and the most productive. Compare mobile and desktop for the same step. A large gap between them is normal in most categories, but a gap that has widened recently points at something that broke on phones: a popup covering the buy button, a slow image, a form that will not accept an address, a payment button that fails on one browser.

Source is the second. Paid, organic, email and direct convert differently, so a shift in the traffic mix moves the blended number without anything getting worse. Look at each source against its own history rather than against each other.

Country and currency is the third. A conversion rate that collapsed for one country usually means shipping cost, delivery time, a payment method people there expect, or a price that now looks wrong after an exchange rate move.

Product is the fourth. Sort products by sessions against orders and look at the worst ratios among the pages getting real traffic. In most stores a handful of products absorb a large share of visits and convert far below the rest, and the reason is visible as soon as you open the page.

New against returning is worth a look too. If returning customers convert normally and new ones do not, the problem is trust, clarity or expectation setting rather than the mechanics of the site.

Analytics tells you where people leave. It does not tell you why. For that, use the store the way a stranger would, on the phone your buyers use, on a normal connection, in a private window.

Arrive from a search rather than by typing the address. Find a product the way a first time visitor would. Read only what is on the screen and ask the questions a buyer asks: what is this made of, will it fit, when will it arrive, what does delivery cost, what if it is wrong. Note every point where you had to guess or scroll to find an answer. Then add it to the cart and go all the way to payment, with a real address and a real card if you can.

Add two more sources of why. Session recordings for the step you identified, watched in a batch of twenty rather than one at a time so you see patterns instead of anecdotes. And customer contact: support tickets, chat transcripts, product reviews and post purchase survey answers. The same three questions usually appear again and again, and each one is a fix.

Then rank what you found rather than doing all of it. Estimate the money each fix is worth, how confident you are that it will work, and how much effort it takes. Do the high value, high confidence, low effort items first, in order, one at a time so you can tell which change did what.

And measure before and after on the specific step, not on total revenue, because total revenue moves for a dozen reasons every week.