I want to be precise about this, because I have also written that GA4 is fine. It is. GA4 is a capable, free web analytics tool, and most stores should keep it installed for site behavior, landing page analysis, and organic search. My argument is narrower: GA4 should not be the place your e-commerce brand decides how to split ad budget. It was not built for that job, and several of its design choices actively work against it.
It only sees cost for Google, natively
GA4 pulls spend from Google Ads automatically. Meta, TikTok, Pinterest, and everything else require a cost data import you set up and maintain yourself. Without it, GA4 can tell you revenue by channel but not ROAS or CAC, which are the numbers budget decisions actually run on. Most brands I talk to have never set up that import, so their GA4 attribution reports are revenue reports with no denominator.
Its model choices narrowed
Google removed first click, linear, time decay, and position-based models from GA4 in 2023. What is left is data-driven attribution, which is the default, and last-click variants. Data-driven attribution is a reasonable idea, but in GA4 it is a model you cannot inspect, built by the company that sells the largest share of most brands’ ad inventory. You do not need to assume bad faith to want a second opinion on how much credit Google search should get.
It loses the same signal your pixels lose
GA4 is a browser tag by default. Ad blockers, consent banners, and browser privacy rules all reduce what it sees, and consent mode fills some of the gap with modeled data. That is why GA4 revenue rarely matches Shopify. I walked through the reasons in why GA4 and Shopify disagree about your revenue. An attribution source that misses a meaningful share of orders, and misses some channels more than others, will skew your budget toward the channels it sees best.
It does not think in customers
E-commerce attribution gets most useful when it splits new and returning customers, ties revenue to products, and handles refunds properly. GA4 can approximate some of this with custom setup, but none of it is native in the way e-commerce teams need. Without a clean new versus returning split, retargeting and branded search look far better than they are, because they mostly collect from people who were already buying.
Its reports are hard to trust at a glance
Data thresholds can hide rows, explorations can sample, the attribution settings are buried in admin, and reports built on different scopes give different totals. A skilled analyst can work through all of it. Most $5M brands do not have that analyst, and a report that needs one to read correctly is a risky basis for moving budget.
What to use instead
For budget decisions, use a tool built for e-commerce attribution: one that pulls spend from every ad platform, reconciles to your store’s actual orders, splits new and returning customers, and lets you compare models. ThoughtMetric is the one I would point most brands to. (Disclosure: ThoughtMetric sponsors this site.) It connects to Meta, Google, TikTok, Pinterest, Microsoft, Snapchat, and others, works on Shopify, WooCommerce, BigCommerce, Magento, and custom stores, offers five attribution models, and includes post-purchase surveys and product-level attribution. Pricing starts at $99 a month for 50,000 pageviews with a 14-day free trial. I compared the two directly in ThoughtMetric vs GA4.
Other e-commerce attribution tools solve this too, at different prices and depths. The point is to use one, not GA4, for the money decisions.
What to keep GA4 for
Plenty. Landing page and funnel analysis, site search behavior, organic and referral trends, page speed problems, and anything where you want to understand how people move through the site rather than which ad deserves credit. Keep the Google Ads link for audiences. And keep it running as a cross-check. If GA4 and your attribution tool move in opposite directions for a channel, that is worth investigating. More on running both in do I still need GA4 if I have an attribution tool.
The short version
- Keep GA4 for site behavior, funnels, and organic analysis.
- Do not use it to split ad budget across Meta, Google, and other channels.
- Use an e-commerce attribution tool that reconciles to real orders and splits new and returning customers, such as ThoughtMetric (our sponsor).
- Treat big disagreements between the two as a prompt to test, not a reason to pick one blindly.
Leave a Reply