Been doing more CRO-focused work lately and I’ve realized everyone seems to have their own starting point when they sit down to audit a store. Some people go straight to analytics, others just start clicking through like a customer would.
Here’s roughly where I start right now:
Check the data first — where are people actually dropping off? Product page? Cart? Checkout? No point “fixing” a product page if the leak is at checkout.
Go through the store on mobile as a real buyer — most traffic is mobile, and problems that are invisible on desktop jump out fast on a phone.
Time the full path — how many taps from landing to “order placed,” and where friction or hesitation creeps in.
Look at the obvious trust gaps — missing reviews, no shipping/returns info near the buy button, slow load.
I try to resist jumping straight to “change the button color” stuff and start with where the money’s actually leaking instead.
Once I’ve spotted the gaps, the fixing part used to mean fighting theme code for every little block. Lately I’ve been using a free app called Conversion Rate Optimizer to add the common stuff — trust badges, star ratings, a free-shipping progress bar, urgency/stock indicators, upsells — straight from the theme editor without touching code. What I actually like most is the built-in analytics that shows impressions, clicks, and attributed revenue per block, so I can see which changes are actually pulling their weight instead of guessing. Made it much easier to turn “here’s the problem” into “here’s the fix, and here’s proof it worked.”
Curious how everyone else approaches the audit part though:
Do you start with data (analytics/heatmaps) or with gut (just browsing the store)?
Is there one page or metric you always check first?
And how do you decide what to fix first vs. what’s just a nitpick?
Would love to hear everyone’s starting point — always looking to tighten up my own process
Same first move as websensepro: data before browsing, so you’re not fixing things nobody’s actually stuck on. I’d start with Shopify’s own Analytics > Behaviour report (it shows the conversion funnel from session to checkout to purchase, so you immediately see whether people are dropping before adding to cart, at shipping, or at payment). Only once you know which step is leaking do the mobile walkthrough, so your hands-on testing is targeted instead of guessing.
The one thing worth adding to your checklist: also look one layer past the funnel, at who’s actually completing (new vs. returning customers, average order value by segment, abandoned checkout volume over time). That’s less “where do they drop” and more “who drops and does it repeat,” which changes whether the fix is trust signals vs. something in your returning-customer experience. Shopify’s built-in reports get you partway there; if you want that segmented over time I’ve used Mipler for the order-side half of that view.
Here is the screenshot from our test store with fake data:
Really well put That “data tells you where, using the store tells you why” distinction is spot on — the numbers point you to the leak, but walking the journey on mobile is where you actually understand it.
And I like your prioritization framework a lot: revenue impact × how many users affected × ease of fix. That’s a clean way to avoid the trap of polishing tiny visual stuff while a real checkout friction point quietly bleeds sales. Reducing friction and building trust over cosmetic tweaks — totally agree that’s where the bigger wins live Appreciate the thoughtful breakdown
Love this addition The Behaviour report as the starting point makes total sense — see exactly which funnel step is leaking first, then do the mobile walkthrough so your hands-on testing is targeted instead of guessing. That ordering is smarter than how I’ve been doing it.
And the “who drops, not just where” layer is a great one. Splitting new vs. returning, AOV by segment, abandoned checkout over time completely changes the diagnosis — trust signals for new visitors is a totally different fix than something broken in the returning-customer experience. Appreciate the Mipler tip and the screenshot too, that segmented-over-time view is exactly the gap in Shopify’s native reports Thanks for going deep on this
For me, it’s the conversion rate on organic vs paid. If organic converts at 3% and paid converts at 0.4%, the store is fine and the ads are the problem. Without that check, you might spend a week rewriting product pages that were never broken.
After that, what I go looking for is whether the page answers the two or three questions the product forces on people. Every category has them.
Apparel is fit, supplements are safety, furniture is dimensions and whether the thing gets through a doorway. In my experience, that’s where the money actually leaks, and those questions are specific enough that a generic audit checklist walks straight past them.
The point above about the two or three questions a product forces on people is the right one, and I’d add the boring half nobody does: you don’t have to guess what those questions are. You already have the list.
Before touching the page, pull the last 50 or so pre-purchase messages the store received - support inbox, chat transcripts, Instagram DMs, the contact form. Not post-order complaints. The ones asked before someone bought, or instead of buying. Then sort by how often the same thing repeats.
What comes out is a ranked list of the objections your product page failed to answer, in your customers’ own words. It beats a heuristic checklist because it is evidence of hesitation from people who were far enough down the funnel to bother typing.
Two things that make it actionable:
Cross-check it against the drop-off data everyone above started with. If the leak is on the product page and your top three repeat questions are sizing, materials and delivery time, you have found the fix without guessing which one to change first.
Then answer them on the page, not just in the reply. This is the part that gets skipped. Most merchants are already answering the same question one-to-one ten times a week, and every answer dies in the inbox - the next shopper with the same doubt never sees it and just leaves.
One addition to the mobile walkthrough: do a pass with the keyboard only, and another with the phone’s screen reader turned on. Same friction audit, but it surfaces unlabelled buttons and focus traps that a sighted tap-through never hits - and those are conversion bugs before they are anything else.
Disclosure: we build AC - Questions & Answers, which is one way to collect and publish those questions on the product page. The inbox-mining step above is worth doing whichever way you end up publishing them.
The Behaviour report as the starting point makes total sense — see exactly which funnel step is leaking first, then do the mobile walkthrough so your hands-on testing is targeted instead of guessing. That ordering is smarter than how I’ve been doing it.
Thanks for going deep on this