The Ecomm Analyst

Usually yes, but for a much narrower job than you are probably using it for now. Keep GA4 as free on-site behavior analytics and as a second opinion when a number looks wrong. Stop treating it as the source of truth for revenue by channel, because that is the one thing it is worst at for an e-commerce store and the one thing most operators lean on it for.

The two tools answer different questions. An attribution platform is trying to tell you which marketing produced which orders. GA4 is trying to tell you what people did on your website. Those overlap enough to feel redundant and differ enough that dropping one leaves a hole.

What GA4 still does that your attribution tool does not

On-site behavior is the big one. Which landing pages convert, where people abandon, how the collection pages perform, what internal search terms come up empty, how a new PDP layout changed scroll and add-to-cart behavior. Attribution platforms are built around the order and the touchpoints before it. They are generally not built to tell you that your size guide page has a 70 percent exit rate.

Organic and direct traffic analysis is another. If you are doing SEO work, GA4 plus Search Console is still the standard view, and attribution tools tend to treat organic as a channel bucket rather than something you can dig into at the query and page level.

Then there is the Google Ads relationship. If you are importing conversions from GA4 into Google Ads for bidding, that plumbing is doing real work in how the algorithm optimizes, and ripping it out has consequences beyond reporting. Same for audience building. Decide that deliberately rather than by neglect.

Finally, it is free. The cost of keeping GA4 running is not licensing, it is attention, and attention is only spent if people keep opening it expecting revenue answers.

Where GA4 stops being useful for revenue

GA4 measures sessions, and its default reporting attributes conversions on a last non-direct basis within its own model. That means a customer who found you through a creator six weeks ago, subscribed to email, and converted from a campaign send gets recorded as email. Not wrong exactly, but not a picture you can allocate budget against.

It also does not know what you spent. Without ad spend joined to orders you cannot compute a real ROAS or a blended number inside GA4, and pulling spend in means building the pipeline yourself. It does not know your COGS either, so profit is out of reach.

And it will not match Shopify. It never has. Tracking prevention, consent mode, ad blockers, and the gap between a session-scoped tag and an order record all push GA4 revenue below what your store actually processed. If you have not seen this yet, you will, and I wrote up the mechanics in more detail alongside the GA4 alternatives worth considering.

Does running both create conflicting numbers?

Yes, and that is fine as long as somebody decides in advance which number is official. The failure mode is not disagreement, it is disagreement without a rule. A weekly meeting where the paid media manager quotes the attribution platform and the ecommerce manager quotes GA4 will burn more hours than either tool saves.

Write down the rule. Revenue and channel performance come from one system. Site behavior comes from GA4. Shopify is the final word on what you actually sold. When two systems disagree by a wide margin, that is a signal to investigate tracking rather than to average them.

What about the tracking overhead of running both?

This is a fair concern and worth checking rather than assuming. Every additional script on a storefront costs something, and stores that have accumulated six or seven tags over the years do feel it. The relevant question is how each tool loads.

ThoughtMetric, which sponsors this blog, uses a single async pixel plus server-side tagging specifically so it does not slow sites down, and is priced at $99 per month with all features at every tier and a 14-day free trial. Whatever you use, the thing to verify is that it loads asynchronously and does not block rendering. Running GA4 plus one well-behaved attribution pixel is not the reason your site is slow. Running GA4, two attribution tools you never removed, a heatmap tool, and four abandoned pixels might be.

When you can actually drop GA4

There are a few cases. If nobody has opened it in three months, it is not serving a purpose and the honest move is to remove the tag rather than let it sit. If you are not running Google Ads and not doing SEO work, two of its strongest use cases disappear. And if you have a privacy-first analytics tool covering site behavior, the overlap gets thin enough to justify cutting.

For most stores between $1M and $20M, though, the answer is to keep it and demote it. It is a free behavioral analytics tool that happens to have a revenue report you should stop looking at.