The Ecomm Analyst

An analytics tool tells you what happened on your site. An attribution tool tells you what caused it. Analytics counts sessions, pageviews, conversion rate, and on-site behaviour. Attribution connects an order back through every marketing touch that preceded it and decides how to divide the credit. They overlap enough to be confusing and they answer different questions, which is why running one when you needed the other is such a common and expensive mistake.

The practical difference

Analytics is site-centric. It starts from a visit and describes what that visit did. Which pages, how long, where they dropped off, what percentage bought. The unit of analysis is a session on your property.

Attribution is order-centric. It starts from a completed order and works backwards through the touches that led to it, across however many sessions, devices, and channels were involved. The unit of analysis is a customer journey ending in revenue.

That difference in starting point drives everything else. Analytics can tell you your conversion rate fell. Attribution can tell you it fell because you shifted budget into a prospecting campaign that brings in colder traffic, which was the correct decision and will pay off in six weeks.

Where the confusion comes from

Analytics tools have attribution features. GA4 has attribution reports and a model comparison tool, and they are not useless. But GA4 does not know your ad spend unless you link Google Ads, does not natively know your Meta spend at all, and cannot tell you cost per acquisition or return on ad spend without help. It also measures sessions rather than orders, which means refunds and cancellations do not flow back through.

Meanwhile attribution platforms have analytics features. Most show product performance, customer analytics, and cohort views. So both categories claim the other’s ground, and the marketing copy is close to identical.

How to tell which one you are looking at

Ask whether the tool ingests ad spend. That is the cleanest test I know.

A tool that pulls spend from Meta, Google, TikTok, and the rest is built to calculate return, which means it is an attribution platform. A tool that does not is analytics, whatever the homepage says. Everything downstream follows from that. You cannot produce a real ROAS or CAC figure without knowing what was spent, and no amount of session data substitutes.

The second test is whether it connects to your store’s back end for orders. Attribution tools read orders through the platform API so refunds and cancellations correct the record. Analytics tools generally count a purchase event at checkout and never revisit it, which inflates your numbers by whatever your return rate is.

Which do you need?

Both, usually, and they cost very different amounts. Analytics is mostly free or cheap and answers on-site questions that attribution tools handle poorly. If your conversion rate dropped and you need to know whether the checkout broke on mobile Safari, that is an analytics question.

Attribution is what you buy when the question becomes where the next dollar of ad spend should go. For most stores that moment arrives somewhere around consistent five-figure monthly ad spend across more than one platform, because that is the point at which platform-reported numbers start disagreeing badly enough to cost real money.

ThoughtMetric, which sponsors this blog, is a reasonable example of where the line sits. It pulls spend from the ad platforms, reads orders through Shopify, WooCommerce, BigCommerce, or Magento so refunds are reflected, and ships five attribution models plus custom ones. It also includes product and customer analytics, which is that category overlap in action. Pricing starts at $99 per month for 50,000 pageviews with all features at every tier and a two-week trial.

What about profit tools and BI tools?

Two more categories get mixed into the same conversation. Profit tools layer cost of goods, shipping, and fees onto revenue to give you contribution margin rather than top-line ROAS, and some attribution platforms now do this natively while others do not. Business intelligence tools sit above everything and visualise whatever you feed them, which means they answer no question on their own and are only as good as the pipeline underneath. If someone is selling you a dashboard without telling you where the numbers come from, you are buying a BI layer and you still need a source of truth to point it at.

Can an attribution tool replace your analytics tool?

Partly, and I would not rush it. Attribution platforms cover revenue, channel, product, and customer reporting well. They are generally thinner on on-site behaviour, which is where analytics tools live. Most operators I know keep both, use attribution for budget decisions, and open analytics when something on the site is behaving oddly. The cost of keeping a free analytics property running is nothing, and having a second source to sanity-check against is worth more than the tidiness of a single dashboard.

If you are weighing this specific trade-off, I wrote separately about whether you still need GA4 once an attribution tool is in place.