.d53b12e5.png&w=3840&q=75 "Trivas.ai")
by Trivas.aiMost ecommerce dashboards hand you one number: total sales. That number feels good right up until you need to make a decision with it, because it hides everything that actually matters. A product doing $40k a month might be thriving on Shopify and quietly bleeding cash on Amazon, and you'd never know from the blended total.
Product performance by channel is the view that actually tells you where to put money next. Without it, you're optimizing based on vibes.
Here's a concrete version of the problem. Say you sell a skincare serum. On Shopify, it runs a 3x ROAS on Meta ads, margins are healthy, and repeat purchase rate is strong. Same SKU, listed on Amazon. Once you factor in referral fees, FBA costs, and PPC spend to stay visible against three competitors on the same search term, that "winning" product is barely breaking even, maybe losing money.
Blended revenue would show you one healthy number and completely mask that split. You'd keep scaling Amazon ad spend on a product that's quietly funding someone else's ad platform instead of your bank account.
The rest of this article is about how to actually build the view that catches this: what counts as a channel, which metrics matter, how to set channel-specific goals, and how to track it over time without losing a week to spreadsheets every month.
"Channel" gets used loosely, so let's define it properly. There are four real categories:
A channel can be where a product is sold, where it's marketed, or both. Amazon is a sales channel and an ad channel at the same time, since Amazon Ads sits right inside it. Shopify is a sales channel that depends entirely on outside channels to drive traffic to it. Keeping these roles straight matters, because "Meta performance" for a product really means ad performance feeding into whichever storefront the link points to.
Most brands we talk to are running somewhere between three and six of these at once. Almost none of them have a side-by-side view. They have Seller Central open in one tab, Shopify analytics in another, and a Meta Ads Manager export downloaded as a CSV, with someone stitching it together by hand every Monday.
Revenue by channel is the metric everyone starts with and it's also the one that lies to you most often. A $50 product selling well on Amazon can look identical in revenue terms to the same product selling well on Shopify, but the margin left over after fees and ad spend can differ by 15-20 points. Revenue tells you volume. It doesn't tell you if you're making money.
The actual list worth tracking per SKU, per channel:
Two more that get skipped constantly: sell-through rate and stockout frequency, broken out by channel. A product can be a genuine top performer on Amazon but show flat performance in your monthly report simply because it was out of stock for nine of thirty days. That's not a demand problem, it's an inventory visibility problem, and blended reporting won't separate the two.
Worth adding a cohort layer too: new versus repeat customer behavior per product. Marketplace buyers behave very differently from owned-store buyers, repeat rate on Amazon is usually lower since Amazon owns the relationship, not you. A product that looks weak on repeat purchases overall might actually be strong on Shopify and dragged down by Amazon's structural ceiling on loyalty.
People ask "what's the best ad channel" like there's one answer. There isn't. It depends on price point, how long someone needs to think before buying, and where they already go to search for this kind of thing.
The pattern that shows up most often: low-price, low-consideration products do well on TikTok and Meta, where the purchase is impulsive and the creative does the convincing. Higher-ticket or considered purchases tend to convert better on Google Search, where the buyer is already looking, and on marketplaces, where they're comparing specs and reviews before deciding.
This means a single-channel strategy applied across an entire catalog is usually wrong for at least half the catalog. A $20 impulse item and a $200 considered purchase need different channel bets entirely, and brands that test per product line instead of assuming one playbook fits everything tend to find this out a lot faster than the ones that don't.
A reasonable sequence for a new product launch looks like this:
The part brands skip is setting channel-specific success criteria upfront. A blended "get a 3x ROAS" target doesn't mean much when Meta and Amazon Ads have completely different cost structures. Set a CAC target for paid social. Set an ACOS target for Amazon Ads. Judge each channel against its own number, not a shared average that hides which one's actually underperforming.
Build in a 30/60/90 day checkpoint to decide where more budget goes, based on actual product-level margin per channel, not top-line sales. A channel can post the highest revenue and still be the one you pull spend from, if its margin is worse than the others.
A single snapshot of last-30-days numbers will mislead you. Seasonality shifts things. Promo periods spike one channel and flatten another. Amazon changes its ranking algorithm, Meta changes its auction dynamics, and your "best channel" from last quarter can quietly become your worst without any change on your end.
A rolling weekly or monthly trend view per product per channel catches this. One month of Amazon dominance doesn't mean much on its own. Six months of a consistent trend does.
The common failure mode is familiar to anyone running more than two sales channels: Amazon data pulled from Seller Central, Shopify data from its own dashboard, ad data pulled separately from Meta and Google, then reconciled by hand in a spreadsheet. That process works fine at 20 SKUs. At 200, it breaks, and by the time someone notices a trend in the data, the trend is three weeks old.
This is the piece most teams are missing, not because they don't understand the problem but because stitching the data together by hand doesn't scale.
Trivas pulls Amazon data, Shopify data, Meta and Google Ads data, and GA4 funnel data into one Redshift-backed warehouse, so product performance by channel lives in a single place instead of four separate logins. The BI reporting layer is built around this exact comparison: SKU by channel, side by side, with margin already factored in rather than bolted on after the fact.
The Wingman AI layer sits on top of that and flags which SKU-channel combinations are underperforming or trending, rather than requiring someone to scan rows of a spreadsheet every Monday morning looking for the anomaly. It's the same process described above, just without the manual reconciliation step that breaks down once your catalog gets past a couple dozen products.
Before anything else, get three things in place: consistent margin data per channel (not just revenue), a recurring review cadence you actually stick to, and product-level tracking instead of order-level tracking, since order-level data hides exactly the SKU detail you need.
If you're piecing this together from Seller Central exports and Shopify dashboards right now, it's worth seeing what a consolidated version looks like. Have a look at how Trivas structures this across Amazon, Shopify, and ad accounts, and if you want more detail on setup for a specific platform, it's worth a look at our guides and reports for the walkthroughs.
explore more insights
3 min read
3 min read
3 min read