If you run an operating store, you already have more analytics than you can read. Shopify shows you sessions and sales, Meta shows you ROAS, your email tool shows you open rates. The problem isn't missing data — it's knowing which type of analysis answers the question in front of you, and which type nobody is running at all.
This guide maps the real categories, ties each to the metrics that live inside it, and walks the math on a realistic store so you can see where the numbers connect. We'll also name the type most "ultimate guide" posts skip: profit analytics.
The two ways to slice ecommerce analytics
There are two honest ways to categorize this field, and confusing them is where teams waste quarters.
The first slices by subject area — what part of the business the data describes. The second slices by analytical depth — how far past "what happened" the analysis reaches. A single metric like conversion rate lives in the conversion subject area and can be examined at any depth, from a flat report to a prescriptive recommendation.
You need both maps. One tells you where to look; the other tells you how hard to think about what you find.
The five types by what they measure
Most ranking guides settle on roughly five subject areas. Here they are, with the metrics that belong to each and a worked example so the numbers stay concrete.
Acquisition analytics
This answers where traffic and customers come from. Channel breakdowns, cost per click, cost per acquisition, and customer acquisition cost (CAC) all live here.
Say you spend $2,800 a month on Meta and acquire 200 new customers from it. Your paid CAC is $2,800 ÷ 200 = $14 per new customer. Blend in a $500 email tool and a freelancer and your true blended CAC climbs — which is the number that actually decides whether growth pays.
Audience and behavior analytics
This describes who your visitors are and what they do on-site: device and geography splits, new-versus-returning mix, product views, add-to-cart rate, and on-site search. It's the layer that explains the "why" behind a conversion number before you jump to conclusions.
Conversion and funnel analytics
This is the drop-off story from session to paid order. The headline metric is conversion rate — orders ÷ sessions. The average ecommerce conversion rate sat near two to three percent across 2025 benchmarks, so a store converting below that band has a funnel problem, not a traffic problem.
Cart abandonment lives here too, and it's brutal: Baymard's running average across dozens of studies is about 70 percent of carts abandoned. If you run 1,000 carts a month and finish 300 orders, your abandonment is 70% — right on the documented average, with real money sitting in the gap. For a deeper build-out of these funnel reads, see our guide to ecommerce performance analytics.
Retention and lifetime-value analytics
This measures what happens after the first order: repeat purchase rate, retention and churn, and customer lifetime value (LTV). Say one store has a quarter of customers reordering and another has fewer than one in ten — they face very different growth ceilings, even at identical CAC.
LTV on a margin basis is AOV × purchase frequency × lifespan × margin ratio. At a $31 AOV, 1.6 orders a year, a two-year lifespan, and a 55% margin, that's $31 × 1.6 × 2 × 0.55 = roughly $55 per customer — the ceiling on what you can afford to spend acquiring one.
Profit analytics
This is the type most guides mention last or not at all, and it's the one that decides whether the business survives. It covers gross margin, contribution margin, true per-order profit, and the honest version of return metrics like POAS (profit on ad spend).
Here's why it matters. A 4.0 ROAS looks like a win everywhere else in your stack — but ROAS is a revenue ratio. POAS = ROAS × margin ratio, so that same 4.0 on a 55% gross margin is a 2.2 POAS (profitable), while on a thin 20% product it's 0.8 (you lose money on every sale the ad drives). Revenue analytics will cheerfully hide that loss until your bank balance surfaces it.
The reason profit analytics is rare isn't that sellers don't care — it's that the inputs live in different systems. Revenue is in Shopify, product cost is in Printify or Printful, ad spend is in Meta and Google. Stitching them per order by hand is the work nobody has time for. Our ecommerce business intelligence hub covers how sellers assemble that picture, and combining Shopify's native reports is usually where the attempt starts and stalls.
The four types by analytical depth
The second map is about how far past the raw number you push. Each level answers a harder question than the last, and it's the framing analytics teams borrow from business intelligence more broadly.
Descriptive analytics — what happened
Your baseline reporting: total revenue, orders, conversion rate, sales by product. Every dashboard you already own is descriptive. It's necessary and almost never sufficient, because it tells you the score without telling you why.
Diagnostic analytics — why it happened
This digs into causes. If conversion fell 20% week over week, diagnostic work isolates whether it was a device, a traffic source, a landing page, or a broken checkout step. Shopify's own reports plus an app like Report Pundit get most operators partway here before the manual cross-referencing gets painful.
Predictive analytics — what will happen
This forecasts: expected sales, inventory needs, churn risk, projected LTV. It leans on historical patterns, so it's only as good as the clean, joined data underneath it — which most stores don't have.
Prescriptive analytics — what to do about it
The hardest level. It recommends an action: pause this ad set, raise this product's price, re-order this SKU before the weekend. This is where analytics stops being a spectator sport and starts changing decisions.
Almost every tool on the market stops at descriptive, with a little diagnostic bolted on. The gap between "here's your churn number" and "here's the email to send the at-risk cohort, approve to send" is the entire distance between a report and an operator.
Where POD sellers actually lose the thread
Two failures recur, and both are about type selection, not effort.
First, running only revenue analytics and calling it profit analysis. A store can grow revenue 30% and shrink net profit if the growth comes from low-margin products or rising CAC. Without contribution margin per order, you can't see it coming.
Second, living entirely at the descriptive depth level. Knowing your numbers is not the same as knowing what to change. If your analytics never produces a specific action you took this week, you own reporting, not analytics.
Where Victor fits
PodVector AI is built for the two types most stores never operationalize: profit analytics, and the prescriptive depth level. Victor is an AI employee — not a dashboard and not an analyst — that works across your live data.
Victor connects Shopify, Meta Ads, Google Ads, Printify, Printful, Gelato, and Klaviyo, then computes true per-order profit by joining revenue, product cost, and ad spend that normally sit in separate tools. He delivers reports straight to your Google Drive, and he can draft approval-gated customer-support email that only sends once you say yes.
Every write action Victor takes is approval-gated — he surfaces the recommendation and the seller approves before anything executes. If you want the prescriptive layer without wiring it together yourself, start with PodVector AI. If delivery data is your weak spot, our Shopify delivery reports guide is a good companion read.
FAQs
What are the main types of ecommerce analytics?
By subject area, five: acquisition, audience and behavior, conversion and funnel, retention and lifetime value, and profit. By analytical depth, four: descriptive, diagnostic, predictive, and prescriptive. The two maps overlap — any subject-area metric can be examined at any depth.
What's the difference between descriptive and prescriptive analytics?
Descriptive analytics reports what already happened — your revenue, conversion rate, and orders. Prescriptive analytics recommends a specific next action, like pausing an ad set or re-pricing a product. Most stores have plenty of the first and almost none of the second.
Why is profit analytics treated as a separate type?
Because its inputs live in different systems. Revenue sits in Shopify, product cost in your print provider, and ad spend in Meta and Google — so computing true per-order profit means joining all three. Revenue analytics alone will show a healthy-looking ROAS while a thin-margin product quietly loses money on every sale.
Which type of analytics should a growing POD store prioritize?
Profit analytics at the prescriptive depth. Once you're doing real order volume, the binding constraint isn't more traffic data — it's knowing which products and channels actually make money after costs, and what to change this week because of it.
Is ROAS a profit metric?
No. ROAS is revenue ÷ ad spend, so it ignores your margin entirely. POAS (profit on ad spend) is the profit version: POAS = ROAS × margin ratio. The same ROAS can be healthy on a fat-margin product and a loss on a thin one.