.d53b12e5.png&w=3840&q=75 "Trivas.ai")
by Trivas.aiLog into Shopify, you see one number. Log into Meta, you see another. Google Ads shows a third. Amazon Seller Central shows a fourth, and it showed up two days late. None of these numbers agree, and nobody tells you which one is "real."
That's the actual starting point for most people searching for an ecommerce analytics for beginners guide 2025 edition. Not a lack of data. Too much of it, none of it reconciled.
Here's the structural problem: every platform defines "conversion" to suit itself. Meta counts a conversion if someone saw an ad and bought within its attribution window, even on a different device. Google Ads has its own model. GA4 has yet another. Amazon doesn't share raw attribution data at all, it gives you aggregated reports and expects you to trust them.
Stack three or four of these next to each other and the totals won't match. That's not a bug. It's just how each platform is built to make itself look good.
2025 makes this worse, not better. iOS privacy changes choked off a lot of the device-level tracking that ad platforms used to rely on, so their own numbers got fuzzier. GA4 replaced Universal Analytics and brought a genuinely different mental model (event-based instead of session-based), which means a lot of beginners are configuring it wrong without realizing it. And Amazon's attribution model still disagrees with whatever Meta or Google tells you about the same sale.
So let's set expectations. This guide isn't about becoming a data analyst. It's about knowing which 8 to 10 numbers to check every week, where to get them, and what they actually tell you. That's it. That's the whole skill.
Most beginners think "analytics" means a dashboard full of charts. It's actually three layers, and you need all three or you're flying blind.
Store-level data: orders, average order value, return rate. This is what Shopify or your Amazon seller dashboard shows you natively.
Marketing-level data: ad spend, ROAS, customer acquisition cost, broken out by channel. This lives in your ad platforms, not your store.
Customer-level data: repeat purchase rate, lifetime value. This is the layer almost nobody looks at until they've already overspent on acquisition.
There's also a difference worth being blunt about: reporting tells you what happened. Analytics tells you why, and what to do next. A report says "revenue was up 12% last week." Analytics says "revenue was up 12% because a TikTok creator's post spiked traffic, but conversion rate dropped, so most of that traffic didn't buy."
Picture a $500K/year DTC brand. Founder checks Shopify every morning, watches the revenue number, feels good or bad depending on the day. But they've never once connected that revenue to which channel actually drove it. Maybe it's all coming from branded search and their paid social spend is dead weight. They wouldn't know, because they're reading a report, not doing analytics.
Skip the vanity metrics. Start with these eight:
The mistake almost everyone makes early on: judging ad performance against revenue instead of margin. If your gross margin is 35% and your blended CAC is eating 40% of revenue per new customer, you're losing money on every sale, even if ROAS looks fine on paper. Scaling that spend doesn't fix it. It just loses money faster.
Here's a Monday morning checklist, five numbers, in this order: revenue by channel, orders, blended CAC, gross margin dollars, repeat purchase rate. Pull those before you open a single ad platform.
One more distinction worth nailing down early: platform-reported ROAS versus blended ROAS. Meta will tell you your campaign did 4x ROAS. Google will tell you its campaigns did 3.5x. Add those claims together and the number will almost always be bigger than your actual total revenue, because both platforms are claiming credit for overlapping sales. Blended ROAS (total revenue divided by total ad spend across all channels) is the number that can't lie to you that way. It's less flattering. It's also real.
Shopify Analytics gives you orders, sessions, and conversion rate out of the box, and it's genuinely solid for store-level data. Its ceiling is clear though: it has no idea what your ad spend is doing. It'll tell you a sale happened. It won't tell you which channel earned it.
Amazon Seller Central splits its reporting into Business Reports (your sales and traffic numbers) and Brand Analytics (search terms, basket data, if you're brand registered). Useful, but slow. Amazon data commonly lags 24 to 48 hours, so if you're checking it expecting same-day accuracy, you're going to misread trends constantly.
GA4 handles funnels differently than the old Universal Analytics did, built around events instead of sessions. Most beginners misconfigure purchase or add_to_cart events in their first few weeks without noticing, which means the funnel report they're staring at is quietly wrong. If a conversion rate looks off from what your store actually shows, check your event setup before you trust the chart.
The thread connecting all three: none of them talk to each other. Shopify doesn't know what Meta spent. Amazon doesn't know what GA4 recorded. You're the integration layer, manually, with CSV exports, unless you build something that isn't manual.
Four steps, in order, and don't skip to step three because it looks more interesting.
Step 1: Centralize your store data. Connect Shopify and/or Amazon to one place instead of logging into each separately every day. Even a basic combined view beats toggling between five tabs.
Step 2: Verify GA4 before trusting it. Install GA4, then go check that purchase and add_to_cart events are actually firing correctly, with the right values attached. Do this before you build a single funnel report off it. A beautiful dashboard built on broken events is worse than no dashboard, because it looks trustworthy.
Step 3: Pick one blended view. Revenue, spend, ROAS, CAC, in one place, so you check one dashboard instead of five logins every morning.
Step 4: Set a weekly ritual. Twenty minutes, same time each week, same five numbers from the checklist above. Ad hoc checking (opening an app when you're anxious about sales) is how people miss trends and overreact to noise.
Trusting platform-reported ROAS as gospel. Meta's dashboard wants to look good. It's not lying exactly, it's just counting conversions by its own rules, which overlap with every other platform's claims. Blended ROAS across total spend is the only honest number.
Ignoring gross margin. Revenue growth with shrinking margin is not growth. It's a slower-moving problem that looks fine until it doesn't.
Mixing new and returning customers into one CAC number. If your "CAC" includes orders from existing customers who didn't need an ad to convert, your acquisition efficiency looks better than it is. Separate new customer orders out, every time.
Manually pulling CSVs from three platforms every week. This is the one that costs the most invisible hours. Nobody budgets time for it, it just eats a chunk of every Monday, forever, until someone automates it.
There's a pretty predictable tipping point. Once you're pulling from three or more platforms weekly, spreadsheet work stops being a quick task and starts being a real weekly tax, often 3 or more hours, every single week, just to get numbers into one view before you can even start analyzing them.
A unified analytics layer does the join differently: Shopify, Amazon, and ad platform data land in a warehouse (Trivas runs on Redshift) and get reconciled there, instead of someone copying and pasting CSVs into a spreadsheet and hoping the columns line up. That's the difference between BI reporting built for this specifically and a spreadsheet held together with formulas.
Once basic metric tracking is routine, an AI-assisted insight layer (Trivas's Wingman, for instance) becomes useful on top of that foundation. It's not a replacement for understanding what CAC or blended ROAS means. It's a layer for once you already get it and want the pattern-spotting done faster.
Three steps, really: pick your core metrics, connect your platforms into one view, set a weekly ritual instead of checking apps at random. That's the whole foundation.
If you want a more structured starting point, our getting started resource walks through the setup in more depth, and our metrics glossary is worth bookmarking the first time a term trips you up.
Tool choice matters less than people think at this stage. Consistency is what actually moves the needle early on. Automate the tedious parts once the habit is solid, not before.
explore more insights
3 min read
3 min read
3 min read