If you run an ecommerce brand, or you are the one who has to report on it, you probably don’t have a data problem. You have a trust problem. GA4 says one thing, Shopify says another, Meta and Google Ads each say something else, and somebody on the team rebuilds the real report by hand in a spreadsheet every week. So copy ours instead. In the video below we build the ecommerce reporting dashboard we hand to retail clients, panel by panel, and every one of them is something you can rebuild on your own store this week.
One caveat first, because it is the backbone of the whole build. A dashboard is only ever as honest as the data layer underneath it, so we start one level lower than most people ever look. This post follows the video chapter by chapter, with timestamps.
Table of contents
Build the Ecommerce Reporting Dashboard From the Data Layer Up
Everything below runs in Google Data Studio (the tool spent a few years named Looker Studio, and it is Data Studio again). But the order matters more than the tool. We audit the data first, then build, then make it look nice. Never the reverse.
Why the Data Layer Matters (0:39)
A dashboard is only ever as honest as the data layer underneath it. A clean chart sitting on broken tracking does not fix the problem, it hides the problem behind a nice looking number. That is the backbone of the whole build, and it is why the first stop is not a chart type. It is the least glamorous screen on your site.
Inspecting the Data Layer (1:23)
On a Shopify store there are at least two places to see what your tracking is actually saying. The Pixel Helper shows you what a pixel reports, and the on-page data layer shows you what the site itself is publishing. In the video we open both on a product page and catch the exact kind of mismatch that poisons reports downstream: the Pixel Helper carries the custom product ID we set up, while the items array flowing into Google Analytics carries the Shopify ID. Same product, two identities. Neither view is wrong, but if you build channel and product reports without knowing which ID lives where, your numbers will never reconcile and you will not know why. Sales channels like Google and Meta add their own reporting standards on top, so the mismatches multiply. And in our experience across retail stacks, the default Shopify data layer and web events are not adequate on their own; the layer breaks more often than teams expect, and correcting it rarely gets treated as a priority until the reports stop agreeing. Understanding this layer first is the baseline for every report you build on it.
Choosing Your Data Sources (3:30)
Two sources, two jobs. Shopify is the source system for the money, because that is how cash actually moves in and out of your accounts. GA4 with its data-driven attribution model (DDA) is the source for channel performance, because the model decides whether the email, the referral, or the direct visit gets full or partial credit for a sale, and that is what tells you how to stretch your marketing dollars. We covered why we trust a data-driven model over old last-click crediting in our piece on attribution models. Know which question each source answers, and half the “whose number is right” arguments disappear before they start.
Building the Base Dashboard (3:51)
The build starts at datastudio.google.com with two sources wired in: Shopify sales through funnel.io, and Google Ads added as a second resource. Then one decision before any chart: who is this ecommerce reporting dashboard for? If the report gets screenshotted into slide decks, as ours usually do, choose a freeform layout at a 16 by 9 canvas so it drops straight into a presentation. If your team opens it ad hoc on whatever device is handy, responsive is the better call (that is the route we took in our Shopify executive dashboard build). Tdhe audience decides the layout, not your taste.
From there, the base panel is four scorecards and two charts: gross sales, returns, margin, and sale quantity, each with a comparison date range set to the previous period, because a flat number tells a stakeholder nothing. Up or down against last period is the signal. Then a time series of sales by day, and a product category chart limited to the top five with everything else bucketed into other, so the shape reads at a glance. Title every panel; an unlabeled chart is a question waiting to interrupt your day. The finished branded template (8:51) shows the same layout dressed for a client: percent of returns to gross sales, no needless decimals (nobody sells half a shoe), and consistent formatting on every page so the reader always knows where to look. All figures in the video are a demo store, and we say so on camera.
Channel Reporting (10:28)
Revenue gets bucketed the way the brand actually spends: paid, organic, email and SMS together (one platform, one spend, split it later if needed), and direct. The channel table below it separates channels you are actively paying for from channels still converting on residue; in the demo, ads stopped on a Wednesday and sales kept trickling in for days with no cost attached. That distinction alone clears up a lot of confused ROAS math.
Then the panel that ends the loudest argument in ecommerce reporting. Every ad platform grades its own homework, so the table shows return on ad spend from GA4’s data-driven attribution next to what the platform claims for itself. On the demo store, Facebook self-reports a 234 percent ROAS, while its cross-channel contribution under DDA lands closer to 108 percent. Roughly half. That gap is normal, and seeing both numbers side by side is the point: the self-reported figure tells you how the platform is optimizing week over week, and the DDA figure tells you what the channel actually contributes once every other touch gets its share. Shopping journeys are convoluted now. Someone researching an engagement ring might come back five days later, after the ads did their work, and buy direct. One reconciled view keeps that from turning into a vendor fight.
Product Reporting (14:00)
This panel depends entirely on the data layer work from the start of the video: clean product IDs, full price versus sale, variations handled deliberately. Get that right and the panel earns its keep. Items purchased, items returned, return rate, and units per transaction across the top; a top-seller highlight with a clickable link straight to the product on the site; and a top ten table with category, subcategory, quantity sold with its change from the prior period, returns, and net sales.
The detail that makes merchandisers actually open it: product images in the table. We publish the image URL into the data layer so it flows into Google Analytics and is available across every report, then in Data Studio change the dimension’s data type from URL to image, and the photo renders right in the row. It is one of the most underutilized features in the tool, and it turns a wall of SKU codes into something a buyer can scan.
Site Funnel Reporting (16:48)
This is where you catch revenue leaking, period over period. Three rates read as early indicators: purchase rate (a drastic drop usually means something on the site broke, not that shoppers changed their minds), engagement rate (new marketing channels often start low and rebound as the algorithms find your buyers), and add-to-cart rate (a leading signal, because carts feed your abandoned-cart campaigns, so keeping it high makes those campaigns work harder).
Then two funnels. The site funnel counts homepage, collection, and product page views, and the shape is diagnostic: on the demo store, more people land on product and collection pages than the homepage, so those pages get the optimization attention, and if the shape ever inverts, that is a marketing or SEO shift worth digging into. The checkout funnel is the critical path, add to cart, begin checkout, purchase, read at a glance for changes in the drop-off. If begin-checkout starts hugging the purchase line, something upstream is discouraging carts, out-of-stocks, a broken quick-checkout button, whatever it is. Establish the baseline first, then the glance is all it takes.
Keeping It Simple for Executives (20:14)
The biggest mistake in executive reporting is the complicated dashboard, numbers everywhere, and an executive who asks ten questions every time it opens. The goal is one place the whole organization can hit and be on the same baseline, so the conversations stop being “is this right, where did this come from, why is this different from what I’m looking at.” It also buys back your time: screenshot it into a presentation, or send a link with a date range and talk about the same screen together, instead of rebuilding the numbers by hand.
And the order of operations we hold to: we made this report look nice after we made sure the data was correct. Data first, cross-checked in a set scenario, then the template, then the polish, with every page in a consistent format so the reader’s brain knows where the same data lives on every slide. A beautiful report is correct data, understandable placement, and then the branding. In that order.
The Part You Cannot Eyeball From a Chart (22:16)
Start by copying this dashboard. Every panel in the video is something you can rebuild on your own store this week, and it will already beat the spreadsheet somebody has been updating by hand. That is the copyable part, and we mean it.
But we started one level down at the data layer for a reason. The dashboard only tells the truth if what feeds it is correct, and on most stores it is not, and nothing on screen warns you. The events fire wrong, the match rates drop, the product feed is malformed, and the consent step leaks. You cannot eyeball any of that from a chart.
So if you want to know whether the data under your reporting can be trusted, that is the thing we audit. We check the stack, work with your dev team to fix what is off, and scope the framework so your reporting holds up to the people you answer to.
FAQs
A: Yes, that is the point. Every panel in the video is built live in Google Data Studio and none of it needs custom code: four scorecards with previous-period comparisons, a channel table reconciled against data-driven attribution, a product table with images rendered from the image URL, and site and checkout funnels. What you cannot copy is the state of your own data layer, which is what decides whether the panels tell you the truth once they are built.
A: Because each platform grades its own homework. Facebook credits itself for sales that other touches helped create, while GA4’s data-driven attribution spreads credit across every channel that contributed. On our demo store, Facebook self-reports 234 percent ROAS while its cross-channel contribution under DDA lands near 108 percent, and a gap of roughly that size is common. Use the self-reported number to judge week-over-week optimization inside the platform, and the DDA number to decide where marketing dollars actually go.
A: The structured data your site publishes on each page, product IDs, prices, and events, that pixels, GA4, and sales channels read from. It is the source every report inherits from, so a mismatch there (for example, a custom product ID in one system and the Shopify ID in another) shows up as numbers that never reconcile downstream. Tools like the Pixel Helper and the on-page data layer let you inspect what is actually firing.
A: Both, for different questions. Shopify is the source of truth for the money, because it reflects how cash actually moves through your accounts. GA4 with data-driven attribution is the source of truth for channel performance, because it decides how much credit each marketing touch earns. Trouble starts when a team uses one source to answer the other source’s question.
A: Start where the reports start. Open the Pixel Helper and the on-page data layer on a product page and compare what each says: do the product IDs match across Shopify, GA4, and your sales channels, and do the events fire when they should? Mismatched IDs, missing events, dropped match rates, a malformed product feed, and consent leaks are the usual suspects, and none of them are visible from a chart. That stack-level check is what a data audit covers.