Most CRO audits start with the product page. That is the wrong end. If your tracking is double-firing or your consent state is broken, every number you are about to optimise against is fiction — so this checklist starts where a senior engineer starts, and works forward.
A conversion rate optimisation audit is a structured review of everything between a visitor arriving and money landing. Done properly it produces a ranked list of fixes with an expected value against each one. Done badly it produces forty screenshots and a recommendation to make the button bigger.
This is the checklist we run. It is ordered deliberately: measurement first, then speed, then the pages, then checkout. That order matters, because a fix you cannot measure is a fix you cannot defend — and roughly a third of the stores we look at have a tracking fault serious enough that their reported conversion rate is simply wrong.
Want this run on your store instead of doing it yourself? We do a free CRO audit—a senior engineer reviews your storefront, funnel, and analytics and sends written findings within one business day. No app install, no credit card, no call required to get the report. It is normally a $5,000 engagement, and you keep the findings whether or not you ever hire us.
What a CRO audit actually is?
Three things get called a “CRO audit” and they are not the same:
- An automated scan — a tool loads your public pages and checks them against a rule set. Fast, free, genuinely useful as a first look. It cannot see your analytics or your margins.
- A heuristic review — a human walks your funnel against a framework. Better, but it is opinion until something is tested.
- A data-led audit — analytics, session recordings and the funnel read together, so findings are ranked by what they are actually worth.
The checklist below covers all three levels. Items shown in orange are the ones that most often turn out to be the expensive problem.
1. Analytics & tracking integrity
1.1 Purchase events fire exactly once per order — Duplicate purchase events inflate reported revenue and quietly wreck ad bidding. Check by placing a test order and watching the network tab, not by trusting the dashboard.
1.2 Browser and server events deduplicated on event_id — If you run Meta CAPI alongside the pixel without matching IDs, Meta counts the same conversion twice.
1.3 Consent Mode v2 configured and respected — If analytics tags load before consent you have a compliance problem. If they never load after consent you have a data problem.
1.4 Core events carry value and currency — add_to_cart, begin_checkout and purchase all need a value. Without it, value-based bidding has nothing to bid on.
1.5 Cross-domain tracking intact — If checkout or booking sits on another domain, sessions break at the handoff and attribution goes with them.
1.6 Internal and bot traffic filtered — Your own team browsing the store all day drags the conversion rate down and nobody notices.
1.7 Conversion rate measured per template — A site-wide average hides a product page converting at half the rate of the rest.
1.8 Offline or in-store revenue fed back — If a meaningful share of sales close off-site, the ad platforms are optimising against your cheapest half.
2. Speed & Core Web Vitals
2.1 LCP under 2.5s on mobile — Usually the hero image on a product page or the first grid row on a collection. Measured on field data at the 75th percentile, not a lab score from your laptop.
2.2 INP under 200ms — Long JavaScript tasks blocking the main thread. Variant selectors and cart drawers are the usual culprits.
2.3 CLS under 0.1 — Images without dimensions, banners injected after load, fonts swapping and reflowing text.
2.4 Hero image preloaded, sized and in a modern format — A 4 MB campaign image uploaded by marketing will undo a year of optimisation.
2.5 Render-blocking app scripts identified — Review widgets and chat tools that inject their own stylesheets and fonts.
2.6 Dead scripts from uninstalled apps removed — Uninstalling an app rarely removes its theme code. We find leftovers on most stores.
2.7 Every image has width and height — The single cheapest CLS fix there is.
3. Product page
3.1 Price, availability and delivery promise above the fold on mobile — Test on a real phone. “Above the fold” on a desktop mock means nothing.
3.2 Variant selection needs no scrolling to find — If someone has to hunt for the size selector, some of them will not.
3.3 Out-of-stock variants disabled, not hidden — Hiding them makes the range look smaller and confuses returning buyers.
3.4 Gallery loads fast and zoom works on touch — Pinch-zoom that fights the page scroll is a silent conversion killer.
3.5 Reviews visible without a click, count near the price — Star rating and count beside the price is one of the most reliable wins in ecommerce.
3.6 Shipping and returns answered inline — Not a footer link. The question arrives at the moment of hesitation, so the answer has to be there.
3.7 Sticky add-to-cart on mobile — On a long page the buy button should never be more than a thumb away.
3.8 Cross-sell placed after the decision, not before — Alternatives shown above the add-to-cart create doubt rather than basket size.
3.9 Product schema valid and complete — Price, availability and rating. Broken schema loses rich results without warning.
4. Collection & discovery
4.1 Filters match how customers shop, not your admin taxonomy — Buyers filter by fit, occasion and problem. Admins organise by supplier and SKU.
4.2 Sort default chosen deliberately — “Featured” usually means “whatever order they were added”.
4.3 Load pattern does not break pagination or SEO — Infinite scroll with no crawlable pagination hides most of your catalogue.
4.4 Sold-out products pushed down — A grid that opens with three unavailable items reads as a dead store.
4.5 Collections have unique intro copy — Both a conversion and an indexation issue.
4.6 Search handles misspellings and synonyms — Search users convert several times better than browsers, right up until they get zero results.
5. Cart & checkout
5.1 Shipping cost revealed before the final step — Unexpected cost at checkout is consistently the top cited reason for abandonment.
5.2 Free shipping threshold shown with progress — “$12 away from free delivery” is one of the highest-yield lines on a store.
5.3 Express wallets prominent — Shop Pay, Apple Pay and PayPal remove the form entirely for returning buyers.
5.4 Discount field not over-prominent — A big empty promo box sends people to a coupon site, and some never come back.
5.5 Guest checkout available — Forced account creation is still one of the largest single drop-offs in ecommerce.
5.6 Cart drawer versus cart page tested — Both work. Which works for you is an empirical question, not a preference.
5.7 Checkout built on extensibility, not legacy scripts — Script-based checkout customisation is end-of-life. If yours still depends on it, that is a deadline, not a preference.
5.8 Abandoned checkout flow live and actually firing — Confirm it sends. We regularly find flows that stopped months ago after a tag was renamed.
6. Mobile
6.1 Tap targets at least 44px — Variant swatches and quantity steppers are the usual offenders.
6.2 Sticky elements not fighting each other — Announcement bar, sticky header, sticky add-to-cart and a chat bubble can leave a phone-sized hole to read through.
6.3 Correct input types on forms — type=”email” and inputmode=”numeric” change the keyboard and measurably reduce errors.
6.4 Zoom not disabled — Blocking pinch-zoom is an accessibility failure and a conversion one.
6.5 Tested on a mid-range Android on 4G — Not an iPhone on office wifi. This single change reframes most speed arguments.
6.6 Entry popups not blocking content — An interstitial on first load is both a ranking risk and an instant bounce.
7. Trust & post-purchase
7.1 Reviews are real and photographic — Five seeded five-star reviews read worse than none.
7.2 Contact details and a real address findable — For higher-value carts this is a genuine purchase blocker.
7.3 Returns policy in plain language, linked from the product page — The confidence to buy usually comes from the confidence to send it back.
7.4 Transactional emails branded — Order and shipping emails have the highest open rates you will ever get. Most stores waste them.
7.5 Post-purchase flow live — Internal link: /retainers/lifecycle-marketing/
7.6 Trust badges that mean something — A real certification helps. A generic seal from 2011 does not.
What changes by vertical
The 50 checks apply to every store. These are the additional ones that matter most in specific categories.
Fashion, apparel & streetwear
Fit is the objection. A size guide with real garment measurements, model height and size worn on every image, and a returns policy that removes the risk of guessing wrong. Drops need queue handling and stock that updates live — a sold-out size discovered at checkout is the worst possible moment.
Beauty & skincare
Shade and skin-type matching, full ingredient lists (a real objection, not a formality), and subscription framed as convenience rather than commitment. Before-and-after imagery needs care to stay within advertising rules.
Fine jewelry & high-AOV
A single visit almost never closes it. Saved shortlists that persist across devices, certification and appraisal surfaced at the point of hesitation, financing shown before the price is read, and an appointment or consultation as a valid conversion. [LINK: /shopify-plus-for-jewelry-brands/]
Supplements & consumables
Replenishment is the whole business. Subscription discount visible on the product page, cadence matched to actual consumption, and compliance-safe claims. Repeat rate matters more than first-order conversion.
Furniture & home
Dimensions, delivery lead times and room-scale imagery. Delivery cost and timeline must appear on the product page — for large items it is the single biggest unanswered question.
B2B & wholesale
Different funnel entirely: account application, net terms, quick reorder from history, and price visibility rules by customer group. Conversion often means an approved account, not a cart.
If you sell in a category that is not listed, the method is the same — find the one objection your buyer has that other categories do not, and answer it on the page where it occurs.
Free CRO audit tools, compared honestly
There are three ways to get a store audited, and they suit different moments.
Option | What you get | Cost | What it cannot do |
Our free CRO audit | A senior engineer reviews your storefront, funnel and analytics and writes findings ranked by revenue impact. | Free. One business day. No app install, no card, no call needed to receive it. | It is not instant — a human is reading your store, so it lands the next working day. |
Instant online scanners (e.g. CRO.Clinic) | Loads your public URL and checks it against a rule set. Report in seconds. | Free, often with a cap on scans. | No access to your analytics, so it cannot see broken tracking, real conversion rates, or which fix is worth most. |
Shopify App Store audit apps (e.g. Sellify, Smart CRO Agent) | In-admin scanning and monitoring across home, collection and product pages. | Free tiers, usually with paid plans above them. Requires installing an app with store access. | Still rule-based. And it is another app in your theme — see check 2.5. |
Paid agency audits | Human review, usually with a call and a slide deck. | Commonly four figures, paid before you see anything. | You have committed budget before you know whether their thinking is any good. |
Use a scanner first if you want something in the next thirty seconds — they are genuinely useful for catching the obvious. But understand the ceiling: an automated tool can only see what is in the HTML. It cannot tell you your purchase event fires twice, that your abandoned-cart flow stopped in March, or that your mobile product page converts at a third of desktop. Those are the findings that move revenue, and they need someone reading your data.
Get the human version, free. We run this full checklist against your store and send written findings within one business day, ranked by what each fix is worth, not by severity score. Normally a $5,000 engagement. You keep the findings either way, and there is no call required to receive them.
How to prioritise what you find
A 50-item list is useless without an order. Rank by expected value, which needs three inputs:
- Traffic through the page. A product page fix on 2% of sessions is worth a fraction of the same fix in checkout.
- Size of the leak. Measured, not guessed — which is why section 1 comes first.
- Cost to fix. A theme setting and a checkout rebuild are not the same bet.
Frequently Asked Questions
How do I get a free Shopify CRO audit?
Two routes. Instant automated scanners like CRO. Clinic, or App Store tools like Sellify and Smart CRO Agent, give you a rule-based report in seconds useful for the obvious issues. Or request our free CRO audit [LINK: /request-a-free-cro-audit/], where a senior engineer reviews your storefront, funnel and analytics and sends written findings within one business day. No app install, no credit card, and no call needed to receive the report.
What should a CRO audit checklist include?
Seven areas, in this order: analytics and tracking integrity, speed and Core Web Vitals, product page, collection and discovery, cart and checkout, mobile, and trust and post-purchase. Tracking comes first because every other measurement depends on it — if purchase events are firing twice, your reported conversion rate is wrong and so is every conclusion drawn from it.
Are free CRO audit tools any good?
For a first look, yes. They reliably catch missing schema, oversized images, absent alt text and obvious layout problems. Their limit is structural: they only see your public HTML. They cannot access your analytics, so they cannot find a broken purchase event, a flow that stopped firing, or tell you which of forty findings is worth the most money. Use one to triage, then get a human to read the data.
How long does a CRO audit take?
An automated scan is seconds. A written human audit is one business day for our free version. A full data-led audit with session recordings, funnel analysis and a prioritised roadmap is typically one to two weeks, because it involves reading real behaviour rather than inspecting pages.
What conversion rate should a Shopify store expect?
There is no useful universal number, and anyone quoting one is selling something. It varies enormously by category, price point and traffic source — a $30 consumable and a $14,000 ring have nothing in common. The only benchmark worth optimising against is your own, segmented by device and template. A site-wide average almost always hides a page performing far worse than the rest.
Should I fix everything the audit finds?
No. Fix what is broken, test what is arguable, and ignore what is small. Broken tracking, dead scripts, and oversized images get fixed there is nothing to debate. Layout, copy, and merchandising decisions get tested if you have the traffic for a test to reach significance. Everything else waits for a month when it is the biggest thing on the list.
