The Monthly SEO Report for an Ecommerce Client: What to Include
A monthly ecommerce SEO report built by page type: category vs product clicks, non-brand demand, indexing of product URLs, seasonality and GA4 revenue.
An ecommerce SEO report should not lead with one sitewide click number, because a shop is several sites in one: category pages, product pages and content pages behave differently. Split non-brand Google clicks by URL pattern, list the product pages that lost the most clicks, check how product URLs are indexed, compare seasonal lines with the same dates last year, and show GA4 organic revenue or key events at landing-page level with the join’s limits stated. Give AI-assistant referrals one small separate line. Fit it on one page and attach the detail behind it.
One sitewide number hides what happens in a shop
The most useful change you can make to an ecommerce SEO report is to stop reporting one organic click number and report it by page type. A shop with 8,000 products and 120 categories can lose a third of its product clicks while category clicks rise, and the total looks flat.
A content site has pages that mostly do the same job. A shop does not. Category pages target broad commercial searches and carry most of the competitive queries. Product pages catch specific searches (a model, a size, a part number) and come and go with the catalogue. Content pages (guides, comparisons, buying advice) feed the other two. Each type fails for different reasons and gets a different fix, so a report that mixes them tells the client very little and tells you even less.
The report below is written for small agencies and freelancers who run monthly reporting for several online shops. It uses Search Console and GA4, which most clients already have, and it fits on a page. The structure is the same for a Shopify store, a WooCommerce shop or a custom build, as long as the URL structure gives you something to match on.
Split every number by page type with URL patterns
Define three groups with regular expressions on the URL, run the same period for each, and report them side by side. Search Console lets you filter by Pages with a custom regex, and the same filter works through the API.
Google’s regex filter documentation says Search Console uses RE2 syntax, matches partially by default and is case-insensitive unless you add (?-i). Anchor the start with ^ so a pattern for products does not also catch a blog post that mentions “/p/” somewhere. These patterns are illustrative; swap in the client’s real structure:
| Page type | Example pattern | What the group tells you |
|---|---|---|
| Category | ^https://(www\.)?example\.com/c/ | Broad commercial demand, ranking strength, internal linking |
| Product | ^https://(www\.)?example\.com/p/ | Catalogue health, indexing, stock and seasonal churn |
| Content | ^https://(www\.)?example\.com/blog/ | Top-of-funnel reach and support for categories |
| Everything else | Doesn’t match the three above | Home page, brand and utility pages |
Check the leftovers. If the three patterns together cover 90% of clicks, the fourth group is the home page and some utility URLs. If they cover 60%, your patterns miss a section, such as /brands/ or /collections/, and the report is quietly wrong.
Shops with messy URLs need a decision at the start. Some platforms put products under /products/ and categories under /collections/, others have product URLs at the root with no folder at all. When the structure does not separate types, ask the client’s developer for an export of category URLs from the sitemap or the platform and filter on that list. It is a one-time job and it makes every later month comparable.
Report for each group: Google clicks, impressions, and the change against the previous period and against the same period last year. Then say what moved in one sentence per group. “Category clicks +9%, product clicks -14%, content flat” is already a more useful opening line than “organic clicks -3%”.
Why the order matters
Put categories first. They usually carry the highest-intent searches and the most revenue per click, and changes there deserve the client’s attention before long-tail product movement does. Products come second. Content comes last unless the client invested in it that month.
Report non-brand clicks, not total clicks
For a shop with a known name, a large share of Google clicks can come from people typing the brand. Those clicks come from the client’s marketing, a newsletter or a TV spot, and your work did not produce them. The number SEO moves is non-brand clicks, so report that by page type and show brand as one reference line.
Define the rule once and keep it stable. The branded vs non-branded guide covers regex rules, misspellings and the caveat about anonymized queries. For a shop, add the product brands the client sells if they are also names people search for: a retailer selling someone else’s brand is not “brand” for the retailer’s own SEO in the same way as the shop’s name. Decide which side of the line those fall on, write it in the report footnote and do not change it mid-year.
Keep an eye on one trap. Many brand queries land on the home page or on a category page named after the brand. If your page-type groups and your brand rule overlap, a category group can look strong because of brand queries pointing at it. Run non-brand on every group, not only on the total.
List the product pages that lost the most clicks
Pick the ten product or category pages with the biggest absolute click loss against the previous period and the same period last year, and add one line of explanation for each. Rank by absolute loss, because a product that went from 6 clicks to 2 is a 67% drop and a rounding error.
For each page ask in order:
- Is the page still live and indexable? A discontinued product that now returns 404 or redirects to a category explains a drop without any SEO problem. Check status, canonical and
noindex. - Did impressions fall or did clicks fall? Falling impressions mean fewer searches or lower ranking. Stable impressions with falling clicks usually point to a snippet change, a competitor’s richer result, or a different query mix.
- Is it a product, or a product line? If a dozen similar products all dropped together, look at the shared template, the parent category and internal links before touching individual pages.
- Was it in stock? The client knows. Ask. Out-of-stock pages for seasonal lines are the most common non-SEO explanation, and the question costs nothing.
Do not build a story for every row. Mark small changes as noise and say so. The client reads “these 40 products each lost fewer than 10 clicks, so no action” as competence.
The traffic drop checklist has the full diagnostic order if one page group collapses.
Check that product URLs are indexed, by reason
For a catalogue site, an indexing line is part of the report, because the risk goes beyond lost rankings: thousands of product URLs may never have entered Google’s index, or may have been removed from it.
Search Console’s Page indexing report gives not-indexed counts by reason. Two features matter for shops. You can filter the report by sitemap: Google’s Page indexing documentation lists All known pages, All submitted pages, Unsubmitted pages only, and a specific sitemap URL. Filtering to the product sitemap gives you product-only counts. The same page says the list of example URLs per reason is limited to 1,000 items and that you cannot search or filter by URL in the report.
That second limit changes how you work. On a catalogue of 20,000 products, the report tells you how many URLs sit in each status but cannot answer “is this specific product indexed?”. For that you inspect URLs.
The monthly indexing block for a shop can be short:
- Submitted product URLs in the sitemap, and how many are indexed
- Not-indexed product URLs by reason, with the change from last month
- Which of the reasons are expected (redirected old products, canonical alternatives, intentional
noindexon filtered pages) and which need action - A sample of 20 to 50 important product URLs inspected individually, with the result
Choose the sample from products that earn money, because a random draw wastes inspections. Use best sellers, new arrivals this month and products you changed. The Search Console not-indexed guide explains what each reason means. If a reason such as “Crawled - currently not indexed” grows on products that matter, that is the finding of the month.
Remember the quota when you plan. The URL Inspection API allows 2,000 requests per day per property, as covered in the 500-URL inspection guide, so a sample of a few dozen is trivial, and a full sweep of a big catalogue has to be spread over days and shared with any other tools using the same property.
Hypothetical triage so the client sees what the line looks like. These numbers are illustrative:
| Not-indexed reason (product sitemap) | URLs | Action |
|---|---|---|
| Page with redirect | 410 | None, old products |
| Alternate page with proper canonical tag | 260 | None, variants point to the main product |
| Crawled - currently not indexed | 85 | Review the 12 best sellers among them |
| Discovered - currently not indexed | 40 | Check internal links and sitemap hygiene |
| Not found (404) | 25 | Redirect or remove from the sitemap |
The report line is “12 best sellers are in Crawled - currently not indexed, and here is what changes this month”. “820 pages not indexed” tells the client nothing.
Separate seasonality from real loss with year over year
A shop’s demand curve is part of the product, so compare seasonal lines with the same dates last year before you write the word “decline”. A month-over-month comparison in January against December records Christmas ending and says nothing about SEO.
Use two comparisons in every report: previous period for momentum, same period last year for the shape of the year. Search Console keeps about 16 months of performance data, enough for one year-over-year comparison, and the stored history page explains what that means if you need to keep older data yourself.
For each page-type group, and for the two or three biggest seasonal categories, read the pattern:
- Impressions down about as much as they fell last year, positions stable: demand, not a loss.
- Impressions up year over year but clicks flat: visibility is improving without traffic, look at titles and the result’s appearance.
- Impressions and positions both worse than last year’s shape: a real loss, go to the page-level work above.
The method is spelled out in the seasonality vs ranking loss guide. Two shop-specific cautions. Moving events such as Black Friday, Easter or school-year starts shift between years, so align the windows on the event, not on calendar dates. And the catalogue changes: a category with twice as many products as last year will not compare cleanly with last year’s category, so note the change in the report.
Show organic revenue from GA4 and say what the join can’t do
The client wants to know whether SEO sells things. Report GA4 revenue or purchases from organic search sessions, by landing page group when you can, and state next to the figure what it can and cannot support.
What you can say: organic search sessions that ended in a purchase event, with the revenue GA4 recorded on them, for the same two periods as the clicks. If the shop sends a purchase event with a value, GA4 can report it. If it does not, tell the client that revenue is not measured and ask their developer to set it up, rather than substituting sessions and hoping. The same applies to cheaper key events such as add to cart or begin checkout when a shop’s volume is low.
What you cannot say: that a particular keyword produced a particular order. Search Console reports queries and pages for clicks on Google. GA4 reports sessions, landing pages and events from its own tag. Google’s Search Console and Analytics integration joins them at landing-page level, not at query-to-order level. The honest sentence for the report is: “Category pages A, B and C gained clicks, and organic revenue from those landing pages rose, but individual orders can’t be attributed to individual searches.”
Other limits to name in a footnote:
- GA4 sessions usually differ from Search Console clicks, often lower, for reasons covered in GSC clicks vs GA4 sessions. Report the trend and skip the exact match.
- GA4 attributes by session source, so a visitor who found the shop through Google, left and came back through an email may credit the later channel.
- Revenue in GA4 and the shop platform’s own order total will not agree exactly (refunds, consent choices, blocked tags). Use GA4 for trend and the platform for accounting.
Give AI-assistant traffic one small line
Product discovery through AI assistants is small for most shops, and it is still worth a single separate line so the client has a baseline before it grows. Do not mix it into organic search.
GA4 has a default AI Assistant channel for recognized assistants such as ChatGPT and Gemini, but Google does not publish the full list, and visits from AI Overviews and AI Mode stay in Google organic search. That means the line is a floor that undercounts the total. The setup, its gaps and how to read landing pages are in the guide to tracking ChatGPT and Perplexity traffic in GA4.
Report sessions, the top three landing pages, and any key events or purchases. If the numbers are tiny, say “small, tracking as a baseline”. The client gets a trend over time without a claim you cannot support.
A one-page layout that fits the whole report
The monthly ecommerce report should fit on one page, with the detail attached behind it for whoever wants it. The page holds the verdict, the page-type table, the movers, the indexing line, revenue and AI traffic.
A layout that works:
- Verdict (two sentences). What changed, and whether it needs a decision. For example: “Category clicks grew and product clicks fell, mostly in two seasonal lines that behave like last year. 12 best sellers are not indexed and that gets fixed this month.”
- Page-type table. Google clicks and impressions, plus non-brand clicks as a separate column, with change vs last period and vs last year for categories, products, content. Brand clicks as one reference row.
- Top 10 losers and top 5 gainers. Page, absolute change, one line of explanation.
- Indexing. Submitted vs indexed for the product sitemap, not-indexed reasons with the month’s change, sample result.
- Revenue and key events. GA4 organic revenue or purchases for both periods, landing-page group, footnote on limits.
- AI-assistant traffic. One line.
- Next month. Three items at most, each tied to a number above.
Choose the comparison window with care. The last 28 complete days against the 28 before is cleaner than calendar months, which differ in length. End the window at least three days back, because Google says collected Search Console data usually takes 2-3 days to appear. The note on cadence and how to talk about a bad month is in how to show SEO results to clients, and the same structure works for a shop.
GSC Magic’s Agency plan can e-mail a monthly report on the 1st to the people you invite into a client group (you switch it on per group, and each person uses one of the 5 seats, the owner included). The report covers part of this on its own: month-over-month clicks and impressions, top queries and pages, and AI-assistant traffic. It does not split by page type or replace the indexing check, so treat it as the delivery layer and keep the one-pager for the shop’s categories and products.
Prompts to run with your client’s data connected
If you have the client’s Search Console and GA4 connected to Claude or ChatGPT through GSC Magic, each block of the report is one prompt. These use tool names that exist in the tool list; replace the domain and patterns with the client’s.
For the page-type table (total clicks and impressions per group):
For example.com, run gsc_query_analytics for the last 28 complete days ending three days ago, filtered to pages matching the regex ^https://(www.)?example.com/c/, then ^https://(www.)?example.com/p/, then ^https://(www.)?example.com/blog/. Repeat the same three for the 28 days before, and for the same 28 days one year earlier. Give clicks and impressions per group in one table. Then add a fourth row for pages matching none of the three patterns.
For non-brand by group:
Use gsc_query_analytics with dimensions query and page, a page filter includingRegex for each pattern (^https://(www.)?example.com/c/, then /p/, then /blog/) and a query filter excludingRegex acme, for the same 28-day window, and sum the clicks per group (visible queries only).
For the losers:
Run report_traffic_drops for example.com by page, comparing the last 28 complete days with the previous 28. List the ten pages with the largest absolute click loss. Mark any page that lost fewer than 10 clicks as noise. Then inspect the top ten with gsc_inspect_urls_batch and report Google’s last crawl result for each (pageFetchState, coverageState, canonical).
For indexing:
Run report_index_health with sitemap_url set to the product sitemap of example.com and max_urls 200 (a sample of the first 200 URLs, not the Page indexing totals), and group the sampled URLs by coverage state. Then use gsc_inspect_urls_batch on these 40 best-selling product URLs and tell me which are not indexed, with the reason Google reports.
Both the sitemap sample and the batch spend URL Inspection quota, and the Page indexing totals still come from the Search Console report itself.
For revenue:
Use ga_list_key_events on the GA4 property for example.com and tell me whether a purchase event exists. Then use ga_run_report to give organic search sessions, purchases and revenue by landing page for both windows. If no purchase event is configured, say so instead of substituting another metric.
Always read what the assistant returns against the report’s definitions. Ask it to say which window, which filter and which property it used, and check that the product regex matched what you expected before you trust the number. Build the one-pager by hand once for each client, then rerun the prompts monthly. That is faster than trying to automate before you know the structure, and a fixed weekly routine keeps several clients on the same rhythm.
What GSC Magic doesn’t cover for a shop
Be clear with the client about the gaps. GSC Magic reads Search Console, Bing Webmaster Tools and GA4 (GA4 is read-only; write actions are off on Free and opt-in on paid plans). It has no crawler, no Google Merchant Center or Shopping feed data, no rank tracker and no Google Ads data. If the client’s revenue depends on Shopping listings or paid campaigns, those need their own reports.
The Free plan covers one site and 10 tool calls a day. The page-type table alone takes about 12 gsc_query_analytics calls (three patterns, three windows, plus the leftovers), so a full monthly run needs the 7-day Pro trial or a paid plan. The Agency plan is $35 a month for 25 sites and 5 seats, with client groups so each client sees only their own properties. The page-type split, the indexing line and the revenue block need no data source beyond Search Console and GA4.
Frequently asked questions
What should an ecommerce SEO report include?
Include non-brand Google clicks split by page type (category, product, content), the pages that lost the most clicks, the indexing status of product URLs, a year-over-year comparison for seasonal lines, organic revenue or key events from GA4 with its limits stated, and a short next-steps list. One page is enough, with the detail attached.
How do I split Search Console data by category and product pages?
Use a Pages filter with a regular expression that matches the URL pattern, such as ^https://(www\.)?example\.com/p/ for products and ^https://(www\.)?example\.com/c/ for categories, and run the same window for each. Search Console uses RE2 syntax and matches partially unless you anchor the pattern. Check what share of clicks your patterns cover, so you know you did not miss a section.
Can I tie Search Console queries to ecommerce revenue?
Not at query level. Search Console shows which pages got clicks, and GA4 shows which landing pages produced purchases, so the two can be read together by landing page. A specific keyword cannot be linked to a specific order from these two sources, and the report should say so.
How do I tell seasonal demand from a real loss on a product page?
Compare the same dates last year for the page group. If impressions fell the way they did last year and positions held, treat it as seasonal. If impressions held or rose and clicks fell, or positions slipped against last year’s pattern, investigate the page, starting with status, canonical and stock.
How do I check whether product pages are indexed?
Filter the Page indexing report to the product sitemap to see not-indexed counts by reason, then inspect a sample of important product URLs with URL Inspection. The report lists at most 1,000 example URLs per reason and cannot be searched by URL, so inspection is how you check specific products.
Give your agent your search data.
GSC Magic connects Claude, ChatGPT and Cursor to Google Search Console, Bing and GA4. Free plan, 7-day Pro trial, no card.