Skip to content
You finish all six stages, you have a single consolidated backlog of problems/opportunities with evidenceand every item is scored and ranked using ICE (Impact, Confidence, Ease). Finally, you convert the ranked backlog into a 90-day execution calendar with an implementation cadence of roughly two prioritized items per week (which typically yields about 24–26 items over ~12–13 weeks).
Do the deep audit on your highest-traffic landing pageAnd also walking the full store user journey across core pages (homepage, navigation menu, collection page, product page, cart drawer, checkout flow, and any other high-traffic pages like About). (Conversion Alchemy OS PDF, page 7.)
Shopify Analytics to confirm which pages get the most sessions and where drop-offs happen. Shopify defines “online store sessions” as the total number of sessions and “online store conversion rate” as the percentage of total sessions that resulted in an order, which helps you quantify where to focus first. Heatmaps and session replays to see real on-page behavior. Click/tap heatmaps show where users click or tap, which helps you identify engagement patterns, misclicks, and friction points. Scrollmaps to see how far people scroll and what percentage reaches each depth. Heatmap survey tool to capture objections, motivations, and customer language. Short, targeted post-purchase surveys are a standard way to learn what almost stopped buyers and what convinced them. PageSpeed Insights to audit speed – because slow mobile experiences create friction and are measurable. LCP (Largest Contentful Paint) is a Core Web Vital that measures when the largest piece of content in the viewport renders; web.dev recommends targeting LCP of 2.5 seconds or less for good UX (measured at the 75th percentile).
Pull heatmaps for the pages in scope and review them on the same device segment you sell on (usually mobile-first for Shopify). Start with click/tap heatmaps because they answer the simplest question: where do people try to interact? Then review session replays for the top problem areas you saw in the heatmap (for example, people tapping non-clickable images, rage clicking, or repeatedly opening/closing elements).
Hotjar or Microsoft Clarity are alternative tools for seeing session replays.Tie engagement back to revenue impact. Your internal ICE screenshot references tools like Heatmap.com and highlights “revenue per session (RPS)” thinking: if a section is upstream of revenue, improving it has higher impact. RPS is revenue generated per engaged session (revenue ÷ sessions). What to write down as findings. For each major section (hero, product imagery, offer/price block, add-to-cart zone, cart, checkout), write a plain-English observation plus what you think it means. Example: “Users are tapping the product image expecting it to zoom; it doesn’t; likely increases friction and reduces add-to-cart.”
Review scrollmaps for the same pages and segments. A scroll map is a type of website heatmap that shows how far users scroll down a page on desktop, tablet, or mobile, typically using a hot-to-cold color gradient and percentages at depth. Translate the scroll data into layout decisions. If only a minority of users reach the section that explains your key differentiator, move that differentiator higher. Or decide on whether to replace that section with something else entirely that can be a more powerful sales asset.
Use either (a) a post-purchase survey, (b) an on-site survey for non-buyers (exit intent or timed).How to synthesize. Don’t overcomplicate it: copy/paste answers into themes. Make one list for motivations (desired outcomes), one for objections (fear/concerns), one for proof that convinced them, and one for customer phrasing you should reuse in ads and on-page copy.You can also feed into an AI tool to get it to do the pattern-analysis for you
Break up the page into each sectionDuplicate each section - you will be “editing” the duplicateReorder the page based on your Scrollmap analysis — are there any sections that need to be higher up?Then apply the High-traffic zones, and the 10% better rule and optimize the remaining sectionsFocus on most important sections (hero section, order form... etc)Based on Heatmap data, you can also see what sections get most engagementDo above the fold firstThen do below the foldApply the same approach if you are doing your full site user journey audit — just start from your homepage, and use the to copy to figma each page in your user journeyHomepageNavigation menuCollection pageProduct pageCart drawer(Screenshot the checkout flow yourself)Baymard’s checkout research repeatedly finds that many sites show too many fields and that reducing visible checkout friction (for example by collapsing optional fields) can improve usability and reduce abandonment risk. Screenshot and map out the carousel imagery in your product order form (either on your landing page or on one your product pages)Think critically on how you can improve the order and the quality of contentSee this figma example:
If you’re auditing your entire funnel, make sure you actually go through the steps on your mobile as well, as the experience may be slightly different to the designs you’ve got from the HTML to figma tool.
Ads and creative angles. Use Meta Ad Library, Foreplay, or TikTok to see what competitors are running right now.TikTok. TikTok’s support documentation describes its Commercial Content Library and that ads generally appear within about 24 hours of campaign initiation, and you don’t need an account to access it. intended to help users learn more about ads seen on Search, YouTube, and Display (useful for competitive scanning of messaging themes). Landing pages and offers. Click through from competitor ads into their landing pages and product pages and apply your Stage 4 audit lens (mobile-first, hero clarity, CTA placement, offer framing, trust).Email and SMS capture. If you want to understand competitor flows, opt into their email/SMS list and map their welcome flow, abandoned cart flow, and post-purchase flow. This directly supports your internal competitor checklist around emails and upsells/cross-sells Unboxing. Recommends buying 1-3 top competitor products (including upsells) and comparing their unboxing experience to yours so you can make yours “10% better”.
Schedule out your optimizations using our Google Sheet template To edit, go to File → Make a copyGo to the Audit implementation tab
Write out your optimizations there, and input the ICE score so you can know what you should work on firstRecommend implementing a maximum of 2 optimizations per week (so you don’t add too many variables and so it’s easier to manage everything)Do not skip QA
Doing a "Deep Audit"
Introduction
What this deep audit is and what “done” looks like
This deep audit is a structured quarterly research process that combines multiple data sources (behavior data, qualitative feedback, UX QA, and competitive pattern research) so you can see what customers are doing, why they’re doing it, and what to fix first.
Your internal “Quarterly deep audits” doc frames the intent as combining methods (instead of relying on just one like heatmaps or surveys) to build a clearer map of customer behavior and the biggest opportunities to lift conversions.
What done looks like this:
Preparation and scope
Before you start the six stages, set the scope so you don’t “collect everything” and do nothing with it.
Scope the pages you will audit.
Set up your tool stack and make sure it’s collecting real data.
The six-stage deep audit SOP:
1️⃣ Heatmap audit
(What people engage with, and what drives revenue per session)
Goal: Identify which sections users actually engage with (or ignore), and which page elements appear to drive or block downstream revenue.
LIVE tutorials of how to do Heatmap + Scrollmap audits:
How to do it.
2️⃣ Scrollmap audit
(what people actually see, and the best order of information)
Goal: Learn how far users scroll, where the biggest drop-offs occur, and whether your best selling content is placed too low.
How to do it:
3️⃣ Customer survey
(why they buy, why they hesitate, and the words they use)
Goal: Capture “voice of customer” so you can sell with the customer’s language, not internal brand language. This stage should reveal motivations, objections, alternatives, and missing information.
How to do it (keep it short):
4️⃣ + 5️⃣ UX-UI + bug-test audit
(the Shopify CRO audit stage, executed mobile-first)
Goal: Identify friction, bugs, clarity gaps, and presentation issues that quietly kill conversion rate (especially on mobile).
Part 2: (UX-UI optimization):
6️⃣ Competitor + “proven pattern” research
(what’s working elsewhere, and how to be 10% better)
Goal: Discover patterns that already work in your market so you can de-risk what you build and avoid reinventing the wheel.
How to do it without guesswork:
Turning findings into a prioritized backlog using ICE
At this point you will have many notes.
The job now is to convert notes into actions that can be implemented or tested.
We basically want to answer:
“If we [make this change], then a 10% uplift would look like [this $$], because of [this evidence/reasoning].”
Use the ICE framework here:
Calculate the ICE number for each optimization
(If you have a lot of ideas, then just pick the top 30 to run through the ICE framework)
Decide “implement vs split-test” for each item.
Reserve split-testing for bigger changes and launch a max of 4-split test per month
If you’re doubtful about launching a test but you dont want to split test for whatever reason, then don’t implement that test. Skip it and move onto another one.
Scheduling and tracking your optimizations
Want to print your doc?
This is not the way.
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.