The Founder's Guide to Ecommerce Analytics: What to Track and Why
by Trivas.ai
|
9 min read
Sep 08, 2026
Most founders don't set out to build bad reporting habits. They just start with whatever their ad platform tells them, because it's right there in the dashboard the day they launch their first campaign. Meta says 4x ROAS, so Meta gets more budget. Simple enough, until it isn't.
The problem is that platform-reported numbers are grading their own homework. Meta counts a sale if someone saw an ad and bought within a wide attribution window, even if that same customer also clicked a Google ad, came in from an email, or would've bought anyway. Google does the same thing, on its own terms. Stack those two "true" stories on top of each other and you get double-counted revenue that doesn't exist anywhere except in ad manager dashboards.
Here's the number that tends to wake founders up: a brand looking at 4x ROAS in Meta Ads Manager might be sitting at 2.1x once you reconcile that spend against actual Shopify orders. Same spend, same period, wildly different story depending on whose report you trust.
That gap isn't a rounding error. It's the difference between a channel that's actually driving growth and one that's just skimming credit off demand you already had. Bad data here doesn't just mean a confusing dashboard, it means real budget going to campaigns that look great in isolation but aren't adding incremental revenue at all.
This guide is a founder guide to ecommerce analytics in the practical sense: which metrics actually matter, where founders get tripped up, and how to decide whether to duct-tape a spreadsheet together or invest in something built for this. No fluff, no vanity metrics.
The Core Metrics Every DTC Founder Should Track
You don't need forty metrics. You need the right five to ten, checked consistently.
Revenue metrics. Track gross revenue, but don't stop there, net revenue (after returns and discounts) is what actually hits your bank account. Add blended ROAS across every paid channel combined, not each platform's self-reported number. Blended ROAS is the one that tells you the truth about total ad efficiency.
Profitability metrics. Gross margin is not the same as contribution margin, and conflating the two is how founders convince themselves a product line is healthy when it's actually break-even. Contribution margin per order subtracts shipping, payment processing fees, and ad spend from revenue. It's a less flattering number. It's also the real one.
Customer metrics. Split revenue between new and returning customers. Track CAC by channel, not blended CAC, because a channel with a high CAC but strong repeat purchase behavior can still be your best one. Pair that with a 90-day LTV:CAC ratio to catch problems before they compound over a full year.
Operational metrics. Inventory turnover and stockout rate belong on this list even though they feel like an operations problem, not a marketing one. A "declining" ROAS is sometimes just a stockout in disguise. If your best-selling SKU goes out of stock for two weeks, every channel's efficiency numbers will look worse, and none of them are lying, they're just reacting to a supply problem nobody flagged on the dashboard.
Most founders at this stage genuinely track five to ten metrics. Beyond that, nobody opens the dashboard consistently, and a report nobody reads is worse than no report at all. If you want a reference for how each of these should actually be defined and calculated, the data dictionary is worth bookmarking, since half of analytics disagreements come down to two tools defining "ROAS" differently.
Where Ecommerce Data Actually Lives (and Why It's Scattered)
Every DTC brand ends up with data spread across the same handful of places: Shopify or WooCommerce for orders, Amazon Seller or Vendor Central for marketplace sales, separate ads managers for Meta, Google, and TikTok, and GA4 for on-site behavior.
Each of these systems is honest about its own slice and blind to everyone else's. Amazon Ads has no idea that a shopper searched your brand name on Amazon because they saw a Meta ad three days earlier. It just sees a branded search click and calls it a win. Meanwhile Meta takes credit for the same customer because its pixel fired on a page view. Nobody's lying, exactly, they just can't see past their own walls.
That's the root cause of scattered, contradictory reporting. It's not that founders are bad at spreadsheets, it's that the underlying data was never designed to be reconciled against anything else.
The fix most mature brands land on is a data warehouse layer, something like Amazon Redshift, that pulls every source into one place and calculates metrics the same way regardless of which platform the raw data came from. Once orders, ad spend, and marketplace fees live in the same warehouse, "ROAS" means one thing, everywhere, instead of five slightly different things depending on which tab you're looking at.
Spreadsheet-stitching works fine at low volume, honestly. If you're doing a few dozen orders a week and pulling three CSVs, a founder or a single marketing hire can keep that current. But it breaks down past a few hundred orders a month, because manual exports introduce lag, someone forgets to refresh a tab, and by the time you catch a problem it's been live for a week.
Common Analytics Mistakes Founders Make Early On
Mistake 1: trusting last-click attribution. Last-click hands nearly all the credit to retargeting and branded search, the bottom-of-funnel channels that catch people who were already going to buy. It systematically undercredits the prospecting campaigns that actually created the demand in the first place. Optimize purely on last-click and you'll starve your top of funnel while pouring more into retargeting that's already saturated.
Mistake 2: checking dashboards every day. Daily numbers are mostly noise, especially for smaller brands where a single big order can swing the day's ROAS by 30%. Set a weekly or biweekly review cadence instead. You'll make calmer decisions and stop chasing single-day blips that mean nothing.
Mistake 3: blending Amazon and Shopify data together. A strong Prime Day week on Amazon can paper over a DTC site conversion rate that's been sliding for a month. Keep the two separate in your reporting. They're different businesses with different customers, different margins, and different problems, and merging them into one number hides exactly the thing you need to see.
Mistake 4: skipping cohort-based LTV analysis. Blended average LTV looks fine even when a specific acquisition cohort is quietly terrible. Run an influencer campaign that brings in one-time discount shoppers, and you won't see the damage in your overall numbers for months, usually right around the time those customers should've been repurchasing and don't.
Build vs Buy: Should You DIY Your Reporting Stack?
The DIY path
What it looks like: Google Sheets pulling manual exports from Shopify, ad platforms, and Amazon, reconciled by hand
Who runs it: usually the founder directly, or a single marketing hire wearing an analytics hat
Time cost: roughly 2 to 3 hours a week once order volume is meaningful, more if something breaks or a platform changes its export format
The buy path
What it looks like: a dashboard tool that pulls Amazon, Shopify, and ad platform data into one place automatically
Who runs it: anyone on the team, no manual pulls required
Time cost: usually under 30 minutes a week for the same review that used to take hours
There's a real rule of thumb for when to switch: once you're running paid ads on three or more channels, or selling across two or more marketplaces, manual stitching stops scaling. The number of reconciliation points grows faster than your headcount does, and that's usually the moment a founder realizes they've turned into a part-time data analyst instead of running the business. This is exactly the kind of decision point founders and CEOs run into earlier than they expect, often right around the point they add a second sales channel.
What to Look For When Evaluating an Analytics Platform
Data ownership. Can you actually export your raw, underlying data, or are you locked into whatever dashboard views the vendor decided to build? If the answer is "you can only see what we show you," that's a real constraint down the line.
Attribution methodology. Ask directly: does this tool calculate blended or incremental numbers, or does it just repackage each platform's self-reported ROAS in a nicer layout? A lot of tools in this category are prettier spreadsheets, not better math.
Marketplace coverage. If you sell on Amazon, check whether the tool actually handles Amazon-specific reconciliation: FBA versus FBM, referral fees, returns, storage costs. A tool that only knows how to pull ad spend and revenue but ignores fee structure will overstate Amazon profitability every time.
Speed to insight. How long does setup actually take, and can a non-technical founder read the resulting dashboard without a data analyst sitting next to them translating it? If you need a specialist to interpret your own reporting, the tool hasn't solved the actual problem, it's just moved it. This is the core of what BI and reporting tooling should actually deliver: numbers you can read cold, without a translator.
Getting Started: A Practical First Step
The short version: pick five to ten metrics that map to revenue, profitability, customer quality, and inventory health. Get your data sources centralized instead of stitched together by hand. Set a weekly review cadence and stick to it. And revisit your tooling honestly once manual reporting starts eating more than two hours a week, because that's the point where the DIY path starts costing you more than it saves.
If you're setting up your first unified dashboard, the getting started guide walks through the setup sequence without assuming you're a data analyst.
And if you want to see what centralized reporting actually looks like in practice, Trivas pulls Amazon, Shopify, and ad platform data into one dashboard built on Redshift, so the number you see for ROAS or contribution margin is the same number no matter which report you open. Worth a look if you're past the spreadsheet stage and want to see what a founder guide to ecommerce analytics looks like once it's actually built into your tools instead of stuck in a doc. Feel free to subscribe if you'd rather get the next one of these delivered instead of found.
Content author and contributor at Trivas.ai, sharing insights on e-commerce analytics, business intelligence, and data-driven strategies to help businesses grow.
Continue Reading
explore more insights
ROAS vs ROI: What's the Actual Difference (and Why Ecommerce Brands Mix Them Up)
3 min read
Transform Your Business Intelligence Today
3 min read
How to Get Proactive Insights Without a Data Analyst
.d53b12e5.png&w=3840&q=75 "Trivas.ai")
