Leaks, not absences
A revenue leak is different from a growth problem. Growth is about getting more people to the store. A leak is revenue you have already earned the right to — a visitor who intended to buy, a return customer, a paid click — that drains away before it becomes an order. Leaks are cheaper to fix than growth is to buy, because the demand already exists. You are not creating it; you are recovering it.
The hard part is that leaks are quiet. Nothing errors out. The store looks fine. Conversion is simply a few points lower than it should be, spread across a dozen small failures that no single dashboard surfaces. Finding them is a diagnostic exercise, and like any diagnosis it works best with a method instead of hunches.
Start where the money is closest to leaving
Work backwards from the moment of purchase, because leaks near checkout cost the most — those visitors were ready.
Checkout. On Shopify Plus, checkout extensibility lets you customize without breaking the funnel, but it also lets you add friction. Look for unexpected fields, shipping rules that surprise people at the last step, payment methods that fail silently, and discount logic that doesn't apply the way customers expect. A single mismatched shipping threshold can cost more than an entire marketing channel.
The cart and add-to-cart. Test add-to-cart on real devices, not just your desktop. A slow or flickering add-to-cart, a drawer that doesn't open reliably, or a cart that loses line-item properties is a direct leak. If you have upsell or bundle apps stacked here, each one is a suspect.
The PDP. The product page is where intent is won or lost. Common leaks: slow image loading, a price or variant selector that shifts as the page loads, missing or thin structured data, and trust signals buried below the fold. If your PDP renders attributes from metafields through Liquid, check that every product actually has the data — empty fields render as broken trust.
The leaks that don't show up on the page
Some of the worst leaks are invisible to a shopper and only appear when you look at the data layer.
Broken or duplicated tracking. If three pixels and an app all claim the same add-to-cart event, your numbers contradict each other and every optimization decision after that is built on sand. Worse, browser-side tracking is increasingly blocked, so you may be losing visibility on a meaningful share of real purchases. Moving to first-party tracking with server events usually recovers both accuracy and attributed revenue.
Inventory and sync gaps. Orders that don't sync cleanly to fulfillment, inventory that updates late, and overselling that triggers refunds are all revenue leaks wearing an operations costume. If a webhook is missing or failing its HMAC check and silently dropping order events, you can lose orders without ever seeing an error.
Performance as a tax. Every extra script — often from apps you forgot you installed — adds milliseconds. Past a threshold, those milliseconds convert directly into abandoned sessions, and the effect is heaviest on mobile where most traffic lives.
A repeatable way to hunt
You don't need to guess which of these is hurting you. You measure.
- Instrument the funnel honestly. One trusted source of truth for each key event: viewed product, added to cart, reached checkout, purchased. When those four numbers are clean, the biggest drop-off points itself.
- Segment by device and source. Leaks are rarely uniform. A funnel that's healthy on desktop can be badly broken on mobile Safari, or fine on direct traffic but leaking on paid.
- Reproduce the failure. Once a drop-off shows up, walk the exact path on a real device until you feel the friction yourself. Most leaks are obvious once you stop testing the happy path.
- Quantify before you fix. Estimate the revenue attached to each leak so you fix the expensive ones first. A 0.5-point checkout fix usually beats a flashy PDP redesign.
This sequence — instrument, segment, reproduce, quantify — turns a vague "conversion feels low" into a ranked list of specific, fixable problems.
What a finished audit produces
The goal isn't a list of opinions; it's a short list of leaks, each with an estimated cost and a concrete fix. Some fixes are configuration. Some are removing an app that's doing more harm than good. A few require building a small custom app or a clean event layer so the data stops lying to you.
Most stores that already sell are leaking somewhere between a fraction of a point and several points of conversion across these areas combined. At real volume, that is not a rounding error — it is often the single highest-return work available, ahead of any new traffic spend.
If you want this done systematically rather than by guesswork, that is exactly the scope of a Shopify audit. You can also see how we approach the underlying instrumentation and fixes in our services.