Every eCommerce site has a story already running. Products are selling or they’re not. Customers are finding what they need or they’re leaving. Before I’d touch anything, I’d want to know what that story actually is, not what I assume it is. Here’s the order I’d work through, and why I’d work through it that way.
What’s Already Working
Before changing anything, I’d find out what’s already earning its keep. Best-converting pages, top-selling products, the traffic sources actually driving revenue. Not the pages someone thinks are performing well, the ones the data actually says are.
It’s tempting to walk in with a list of fixes on day one. Every new set of eyes sees ten things it wants to change immediately. I’d rather protect what’s working than break it by accident chasing something shinier. If a category page is quietly converting at twice the site average, that’s not a page I touch on a hunch. That’s a page I study, so I understand why it works before I try to replicate it anywhere else.
Current KPIs
Numbers first, opinions second. Conversion rate, average order value, cart abandonment, return customer rate, revenue by channel. I’d want a real baseline before I changed a single thing, so six months from now there’s proof of what moved and why, not a vague sense that things feel better.
This also protects me. If I make ten changes in the first quarter, I want to know which ones actually mattered. Without a baseline, every result becomes a guess. With one, I can say exactly what worked, what didn’t, and what to do next.
What Can Be Used That Already Exists
Most sites have more than they’re using. Product reviews sitting in a plugin that never made it into the layout. Photography from a past shoot that’s not on the site yet. Email lists that haven’t been touched in months. Blog posts written years ago that still rank but haven’t been updated or linked to anywhere current.
Before building anything new, I’d take stock of what’s already paid for and sitting idle. Somewhere on most sites there’s already-purchased software, already-shot photography, or already-written copy that’s never been put to use. Finding it costs nothing but time, and it’s usually the fastest way to make real progress before spending another dollar on something new.
What Can Be Added The Quickest
Alongside the audit, I’d flag the fast wins. Missing meta descriptions on top pages. An FAQ section that could go live in a day because the answers already exist in a support inbox. A size chart or spec sheet that already exists as a PDF and just needs to be built into the page instead of buried in a download link.
Small additions that don’t require a developer or a redesign, just someone taking the time to add them. These are the wins that build momentum. They show up fast, they’re low risk, and they buy trust for the bigger, slower changes that come later. Nobody signs off on a longer roadmap from someone who hasn’t proven they can move fast on the easy stuff first.
Common Complaints And Questions
Customer service tickets and support emails are the most honest audit a site will ever get. Nobody fills out a satisfaction survey when everything goes smoothly. But when something’s confusing, missing, or broken, customers say so, directly and repeatedly.
What people ask again and again is what the site itself isn’t answering. If five customers a week ask the same shipping question, that’s not a support problem, that’s a missing page. If the same product keeps coming back because sizing runs differently than expected, that’s not a returns problem, that’s a product page that needs a better size guide. I’d read through as many of these as I could get access to before making any real decisions, because the patterns show up fast and they tell you exactly where to focus.
What It’s Synced To
I’d check what the site is actually connected to. Google Merchant Center, Bing Shopping, any Copilot or ChatGPT shopping feeds that are starting to matter more every quarter. Feeds break quietly. A product gets disapproved, a price mismatch throws off a whole category, an attribute goes missing, and nobody notices until sales drop and someone finally goes looking for why.
I’ve seen this firsthand: a feed error that sat unnoticed for weeks, quietly pulling products out of shopping results while everyone assumed the drop in sales was seasonal. Checking sync health isn’t glamorous work, but it’s often the difference between a site that’s visible and one that’s technically live but invisible where it counts.
Internal Linking
Are the pages that should be connected actually connected, or is the site relying on nav menus and search to do all the work? A blog post about a product category should link to that category. A product page should link to related products, not just sit there as an island.
Thin, orphaned pages don’t get found, don’t get indexed, and don’t convert, no matter how good the content on them is. I’d map out where the gaps are, category by category, and look for the pages doing all the linking work and the ones getting none of it. This is one of the more overlooked levers on most sites, because it doesn’t require new content, just better connections between what already exists.
Broken Links And Thin Pages
Broken links erode trust fast, both with customers and with search engines. A customer who clicks a dead link once might try again. Twice, and they’re gone. Search engines are less forgiving and faster to notice.
Thin pages with nothing worth indexing are just as costly, quietly dragging down how the rest of the site gets evaluated. Neither takes long to find with the right tools, and neither takes long to fix once found: a redirect here, a rewrite there, a page consolidated into a stronger one. It’s unglamorous maintenance work, but it compounds.
Marketing Funnel For Each Segment
A first-time buyer doesn’t move through the site the same way a repeat wholesale account does. One needs trust signals, clear pricing, and reassurance. The other needs speed, reorder tools, and account-specific pricing they can find in two clicks.
Treating every visitor the same is where a lot of sites quietly lose money. I’d map the funnel separately for each real customer segment and look for where each one drops off. B2B and B2C buyers especially need different paths, different content, and sometimes entirely different landing pages, even when they’re shopping the exact same catalog.
Checkout Journey
I’d walk it myself, start to finish, on desktop and mobile, using real payment info if I had to, not a test card in a sandbox environment. Not the happy path someone designed on paper: the actual experience a customer has, extra fields and all.
Every unnecessary click in checkout is a chance for someone to change their mind. An account requirement that could be a guest checkout option. A shipping calculator that loads slowly. A form field asking for information that isn’t actually needed to complete the order. Small friction points add up fast, and they’re often invisible to anyone who isn’t deliberately looking for them.
Sales Team Assets Not On The Website
I’d ask the sales team what they’re using in the field that customers never see. Spec sheets, comparison charts, pricing sheets, one-pagers built for a trade show three years ago that are still the best explanation of a product line anywhere in the company.
Half the time that content already exists on someone’s desktop or in a shared drive, it’s just never made it online where a customer could find it on their own. This is usually one of the easiest wins on the whole list: content that already exists, is already accurate, and just needs a home on the site.
Google And Bing Profile Accuracy
I’d confirm the Google Business Profile and Bing Places listings are accurate and fully filled out. Hours, categories, service areas, photos, and any Q&A sections that might have outdated or incorrect answers sitting on them.
These get set up once and forgotten, and for a lot of buyers, especially local and industrial ones, it’s the first impression of the business before they ever land on the website. An outdated phone number or the wrong service area listed can cost a sale before the site even loads.
Review Capture Process
I’d check whether there’s a system for asking customers for reviews, or whether it’s happening by chance, a handful trickling in from customers motivated enough to leave one unprompted.
Reviews are one of the easiest trust signals to build and one of the most commonly ignored. A simple post-purchase email asking for a review, timed to when the product has actually arrived and been used, can make a measurable difference in both conversion and search visibility. If the process doesn’t exist yet, building it is usually cheap and fast compared to almost everything else on this list.
Email Capture
Does it exist. Is it converting. And is anything actually being done with the list once someone’s on it. A capture form that feeds a list nobody emails isn’t a strategy, it’s a placeholder.
I’d look at where capture happens on the site, what’s being offered in exchange for an email address, and what happens after someone signs up. A list that’s collected but never used is a missed relationship with every single person on it. Fixing this doesn’t always mean building an elaborate campaign. Sometimes it just means finally sending the first email.
Customer Bottlenecks
I’d look for the specific places where someone with real intent to buy gives up. That might be a confusing shipping policy that’s never clearly stated, a checkout step that asks for too much too soon, or a product page that doesn’t answer the one question a buyer actually has before they’ll commit to purchasing.
These bottlenecks are rarely obvious from the outside. They show up in analytics as a drop-off point, in support tickets as a recurring question, or in abandoned carts that never come back. Finding them takes looking at the site the way a customer actually experiences it, not the way it was designed to be experienced.
Product Pricing Updates
I’d look at how pricing changes get made when they happen. Is it a five-minute task, someone with access making a quick update, or a five-day process involving three people, a spreadsheet, and an email chain before anything goes live?
That process alone says a lot about how the rest of the operation runs behind the scenes. A slow pricing process usually means a slow process everywhere else too: inventory updates, product launches, promotional changes. Fixing the pricing workflow often ends up fixing more than just pricing.
Product Images
Outdated, missing, or inconsistent images cost trust before a customer reads a single word of copy. A product photographed five different ways across five different listings looks unprofessional even when the product itself is solid.
I’d check for consistency across every listing, not just the ones on the homepage: lighting, background, angle, resolution. I’d also check for missing images entirely, which happens more often than people expect, especially on older product lines or ones added in bulk through a data import. Clean, consistent images are one of the fastest ways to make an entire catalog feel more trustworthy, even before a single word of the copy changes.
None of this is glamorous. It’s the same groundwork every site needs, whether it’s day one or year five. Coffee helps. ☕