The CRO audit checklist we run on every store

Above the fold: the five-second test

A first-time visitor gives you a few seconds before deciding whether to scroll or leave. Within the first screen — on a phone, not your 27-inch monitor — check:

• Can a stranger say what you sell? Not your tagline — the actual product category. Clever headlines lose to clear ones on cold traffic almost every time this gets tested.
• Is there one visually primary action? If "Shop now," a newsletter modal, and a promo banner all shout at once, the real test is removing things, not adding them.
• Does the hero image show the product, or an abstraction? Lifestyle mood shots test worse than product-in-use shots for stores that haven't already earned brand recognition.
• Is the price visible early on product pages, or hidden below reviews and accordions? Hiding price reads as having something to hide.

Trust: reasons to believe you exist

An unfamiliar store starts from zero credibility, and every trust element is a deposit against that. The checks:

• Reviews near the decision point — a star rating beside the product title and buy box, not only a widget at the bottom of the page.
• Shipping cost and delivery time before checkout. Unexpected shipping cost is the single most-cited reason for cart abandonment in nearly every survey ever run on the subject.
• Returns and guarantee visible where the doubt happens — next to add-to-cart — not paraphrased in a policy page nobody opens.
• Signs of a live business: a reachable contact page, consistent product photography, no placeholder text. Shoppers pattern-match on these fast.

Friction: walk the path to purchase

Add a product to the cart and buy it on your own store, slowly, counting every decision you are forced to make. Look for:

• Steps that exist for you, not the shopper — forced account creation being the classic. Guest checkout is table stakes.
• Fields nobody needs. Every optional form field is a tax; a second phone number field is a tax with no revenue.
• Variant selection that can silently fail — size pickers that don't show what's out of stock until after the click.
• Dead ends after adding to cart. Does the shopper get a clear "in your cart, here's checkout" signal, or does the page just... sit there?
• Discount-code fields that scream at full price. An empty promo box sends full-price buyers off-site hunting for coupons; test collapsing it behind a link.

Mobile: where most of the traffic actually is

For most DTC stores, mobile is 65–80% of sessions and converts at half the desktop rate. Some of that gap is intent; a lot of it is ergonomics. Check on a real phone:

• Tap targets: can a thumb hit the size selector, the quantity stepper, the close button on that popup? 44px is the floor.
• Sticky add-to-cart: once the buy button scrolls away on a long product page, is there still a way to buy without scrolling back?
• Keyboards: does the email field open the email keyboard, the card field the number pad? Wrong keyboards are silent checkout killers.
• Popups: a desktop-sized modal on a phone screen with a 20px close target is a bounce machine, and on top of that Google penalizes intrusive interstitials.
• Horizontal overflow: scroll every template page — one overflowing element makes the whole site feel broken.

Speed: the tax on everything else

Nothing above matters if the page hasn't painted yet. We pull real performance measurements (PageSpeed) into every audit rather than eyeballing it, because perceived speed and measured speed diverge. What to look at:

• Largest Contentful Paint under 2.5s on mobile. The usual store culprits: a 4MB hero image, four font families, and an app or tag manager loading twelve scripts before content.
• Cumulative Layout Shift near zero. A buy button that jumps 40px as late assets load produces mis-taps at the worst possible moment.
• Third-party weight. Every review widget, chat bubble, and pixel was added for a reason; audit whether the reason still exists. Stores routinely carry apps uninstalled from the admin but still shipping script.

Then stop auditing and start testing

An audit's output should not be a 40-page PDF. It should be a short, ranked list of hypotheses — each naming the evidence, the change, and the expected effect — with the top items turned into A/B tests. This is exactly the shape Looplift's audit forces: every run ends in three ready-to-launch experiment proposals, because in the end an audit finding you never test is just an opinion with a screenshot.

Run the free audit and compare its findings against your own pass of this checklist. Where they agree, you have your first test. Where they disagree, you have your second.