A WooCommerce store owner walks into a stakeholder meeting and puts up the slide: 38% year-over-year revenue growth. The room nods. Everyone feels good. The problem? That number is almost certainly not real.

Here’s what the slide doesn’t show: Q1–Q3 2023 was measured in Universal Analytics. Q1–Q3 2024 was measured in GA4. These two platforms count users, sessions, and conversions differently. Comparing them is like measuring your store’s square footage in feet one year and metres the next, then reporting it as growth.

Universal Analytics shut down permanently on July 1, 2023 (Google, 2023). As of early 2026, any GA4 year-over-year comparison that stretches before mid-2023 crosses a methodology boundary. That’s not a data problem you can fix by clicking differently in GA4. The underlying comparison is structurally invalid.

How to Build a Real Historical Baseline#

This isn’t alarmism—it’s arithmetic. There are three separate reasons your GA4 YoY numbers are unreliable. Each one is enough to make the comparison suspect on its own. Together, they make it fiction.

Three Layers of the Problem#

Layer 1: UA and GA4 Are Incompatible Methodologies

Universal Analytics counted sessions and goals. GA4 counts events and conversions. The same purchase produces different numbers in each system—not because something went wrong, but because they define “conversion” differently at a fundamental level.

When you pull a year-over-year comparison in GA4 that spans the July 2023 migration, you’re comparing against a period that GA4 itself did not measure. GA4 may show data for that period through data import or modeling, but it’s not the same data UA collected. It’s GA4’s interpretation of what UA would have seen.

Any YoY comparison in GA4 that includes dates before July 2023 is comparing apples to the idea of oranges.

Layer 2: GA4 Silently Deletes Your Historical Data

GA4’s default data retention setting is 14 months (Google Analytics Help, 2025). After that, user-level and event-level data is deleted. Aggregate data may remain, but the granular records needed for accurate comparisons are gone.

Here’s what makes this dangerous: GA4 doesn’t warn you. When you run a YoY report that reaches back beyond your retention window, it returns results—just silently incomplete ones. There’s no banner that says “this period is missing 60% of its data.” The chart just looks normal.

If you haven’t changed your GA4 retention settings from the default 14-month limit, your 2026 vs 2024 comparison is already running on partial data.

The fix for Layer 2 is straightforward: go to GA4 → Admin → Data Settings → Data Retention and change it to 14 months → 50 months. It won’t recover data already deleted—but it stops the clock from running on future data.

You may be interested in: WooCommerce Says 50 Orders, GA4 Says 32, Facebook Claims 18

Layer 3: GA4 Modeling Changes the Past

GA4 uses behavioral modeling to fill in data gaps created by consent rejection and browser restrictions. This modeling continuously learns and updates—which means it retroactively changes historical numbers.

The same Q3 2024 data you looked at in October 2024 shows different figures when you look at it again in March 2026. Not because anything changed in Q3 2024. Because GA4’s model got smarter and re-attributed conversions differently.

GA4’s data-driven attribution is dynamic and continuously learning—meaning the same period can show different attribution numbers when reviewed later (Google Analytics Help, 2025). This isn’t a bug. It’s working as designed. The problem is that “working as designed” makes historical comparison unreliable by definition.

You’re not misremembering last year’s numbers. GA4 genuinely changed them.

What This Means for Your 2026 YoY Reports#

Let’s be specific about where this leaves you as of early 2026:

• 2026 vs 2025: Both periods are GA4—potentially comparable if your retention is set to max and you’re within window. Still subject to modeling changes.
• 2026 vs 2024: Retention risk is real for early 2024. Check your settings before drawing conclusions.
• 2026 vs 2023: The comparison crosses the UA shutdown. H1 2023 is a different methodology. This number is not valid.
• 2026 vs 2022 or earlier: UA data is inaccessible. GA4 has no record of this period at all.

Most WooCommerce store owners doing YoY analysis right now are in the 2026 vs 2025 or 2026 vs 2024 scenario. Both carry real risk. And none of them solve the deeper problem: GA4 underreports WooCommerce revenue by 15–50% due to ad blockers and browser restrictions (Seresa / Industry analysis, 2025). Even within a clean GA4 comparison period, you’re measuring a changing percentage of the same underlying reality—not the reality itself.

You may be interested in: You Open 4 Dashboards Every Morning and None of Them Agree

The Only Authentic YoY Baseline: Your WooCommerce Order Database#

Here’s the thing: you’ve had the real data the whole time.

WooCommerce records every completed transaction server-side through the payment processor, independent of browser behavior, ad blockers, or consent status (WooCommerce Documentation, 2025). Your WooCommerce orders table doesn’t have a retention limit. It doesn’t use behavioral modeling. It doesn’t change retroactively. Every order: date, total, products, customer—going back to the day your store opened.

For WooCommerce YoY revenue, the authentic comparison is simple:

• Q1 2025 WooCommerce completed orders → compared to Q1 2026 WooCommerce completed orders
• Same definition of “completed” both years (WooCommerce “Processing” + “Completed” statuses)
• Same revenue field—order total before refunds, or net of refunds, your choice—but consistent

This is the number that’s real. It’s the number your payment processor agrees with. It’s the number your accountant uses.

The native WooCommerce analytics tools most stores ignore give you access to this data right now, without any third-party tools. For a reliable YoY revenue check, you don’t need GA4 at all.

BigQuery: The Infrastructure for Future YoY Comparisons#

The WooCommerce order database solves the historical revenue baseline problem. But it doesn’t tell you about traffic sources, customer journey, attribution, or channel performance over time. For that, you need a forward-looking infrastructure—one without GA4’s retention limits or methodology changes.

That infrastructure is BigQuery.

When you stream events to BigQuery, they stay there permanently. No 14-month deletion window. No retroactive modeling. No methodology rewrites. The event that fires when a customer completes a purchase in 2026 will read exactly the same in 2028 when you’re running a three-year comparison.

This is the Data Trees principle (coined by Seresa): plant the data infrastructure now, harvest real comparisons later. The stores that start streaming to BigQuery in 2026 will have their first methodology-consistent YoY comparison by 2027. The stores relying on GA4 will still be comparing approximations against approximations—and wondering why the numbers keep changing.

Transmute Engine™ is a first-party Node.js server that runs on your subdomain (e.g., data.yourstore.com). The inPIPE WordPress plugin captures WooCommerce events and sends them via API to your Transmute Engine server, which routes them simultaneously to BigQuery, GA4, Facebook CAPI, and other destinations—creating a permanent, unmodeled event archive that grows more valuable every year you run it.

Key Takeaways#

• GA4 YoY comparisons crossing July 2023 are structurally invalid—UA and GA4 use incompatible methodologies for sessions and conversions.
• GA4 deletes user-level data after 14 months by default—older YoY queries return silently incomplete results without any warning in the interface.
• GA4 behavioral modeling changes historical numbers retroactively—the same period shows different figures when you look back at it months later.
• Your WooCommerce order database is the only authentic historical baseline—server-side, unaffected by ad blockers, consent rejections, or GA4 changes.
• Streaming to BigQuery now builds the real YoY infrastructure—in 12 months you’ll have your first clean, methodology-consistent year-over-year comparison.

Your WooCommerce revenue history is sitting in your database right now—complete, accurate, unaffected by any platform migration. Seresa helps you build the forward infrastructure so your 2027 YoY comparison is just as reliable.