Migrating from GA4
Migrating analytics platforms is not trivial. You lose historical data, your dashboards break, and your team loses institutional knowledge of GA4’s quirks. Before you migrate, ask yourself: am I migrating for the right reason?
Before You Migrate: The Hard Questions
Section titled “Before You Migrate: The Hard Questions”“GA4 is confusing.” Not a good reason. GA4 is complex, but alternatives are too. You’ll just have different confusion.
“GA4 sends data to Google.” Valid if you have data residency requirements or philosophical objections. Not valid if you just have a vague privacy concern but still run Google Ads.
“I need Google Ads integration to work differently.” Also not a reason to leave GA4. GA4’s Ads integration is GA4’s best feature.
“I need data I don’t have.” Now we’re talking. If GA4 doesn’t give you ecommerce detail, or if you need cookieless tracking, or if you want 5+ years of data without a 14-month rolling window, then migration might be justified.
The Reality of Migration
Section titled “The Reality of Migration”You will lose:
- Google Ads integration (audiences and conversions)
- Historical GA4 data (you can export, but you can’t backfill the new tool)
- Team expertise in GA4
- Audience building for ad platforms (without a CDP)
You will gain:
- A tool that might fit your use case better (if you picked right)
- Simplified dashboards (if you picked a lightweight tool)
- Data ownership (if you picked Matomo or Piwik Pro)
- Faster page loads (if you picked Plausible or Fathom)
The honest take: Most organizations should run GA4 + an alternative in parallel, not migrate.
Concept Mapping: GA4 to Alternatives
Section titled “Concept Mapping: GA4 to Alternatives”Use this table to understand what you’re trading away. Every GA4 concept maps to something, but the quality of the mapping varies.
| GA4 Concept | Plausible/Fathom | Matomo | Piwik Pro | Trade-off |
|---|---|---|---|---|
| Event | Goal/custom event | Event | Event | Near 1:1, simpler in lightweight tools |
| Custom dimension | Event property (limited) | Custom dimension | Custom dimension | Limited in lightweight tools |
| Audience | Not available | Segment | Segment + CDP | Biggest loss: no audience building |
| Conversion | Goal | Goal/conversion | Goal | Direct mapping |
| Ecommerce event | Revenue event (simple) | Full ecommerce module | Full ecommerce | GA4 flexible, alternatives more structured |
| User-ID | Not available | Visitor/user ID | User segment | Limited cross-device tracking in alternatives |
| BigQuery export | API/CSV download | Raw database/API | Data export API | Matomo easier (direct DB access) |
| Consent mode | Not applicable | Supported | Supported | Lightweight tools less consent-aware |
| Predictive audience | ML-driven | Not available | Not available | Unique to GA4 |
The gap widens for advanced use cases. Simple site analytics? Alternatives are solid. Complex ecommerce with Google Ads? GA4 is still better.
The Parallel Running Approach (Recommended)
Section titled “The Parallel Running Approach (Recommended)”Don’t migrate. Run both for 2-4 weeks and see if the alternative tells you what you need.
- Install the alternative alongside GA4. Both fire simultaneously.
- Configure equivalent goals/events in the alternative to match GA4’s key conversions.
- Run for 2-4 weeks minimum. Collect a full business cycle of data (a weekend, a weekday cycle).
- Compare metrics:
- Sessions/pageviews (expect 5-15% variance)
- Traffic sources (biggest discrepancies here)
- Top pages (should align closely)
- Conversions (should track similarly)
- Investigate differences. Where do they diverge and why? (Bot filtering, sampling, cookieless counting, etc.)
- Evaluate if the alternative answers your questions. Does it show you what you need to know?
- Make a decision: Remove GA4, keep both, or go back to GA4.
Why Metrics Will Differ
Section titled “Why Metrics Will Differ”Expect 5-20% variance between GA4 and alternatives. Here’s why:
| Difference | GA4 | Alternative | Impact |
|---|---|---|---|
| Bot filtering | Yes (Server-side) | Sometimes (configurable) | GA4 often lower |
| Sampling | Yes (at high traffic) | No (usually) | Alternative may be higher |
| Cookie vs cookieless | First-party + cross-domain | First-party only | Alternatives lower on repeat visits |
| Session timeout | 30 min default | Varies | Can differ significantly |
| Data processing delay | Real-time | Real-time | Usually same |
| Referrer handling | GA4 specific | Standard | Nuanced differences |
The alternative isn’t “wrong” — it’s just different. Both numbers are defensible.
Implementation Steps
Section titled “Implementation Steps”Once you’ve decided to migrate (not just parallel-run), follow this sequence.
Day 1-2: Audit GA4 config
- List all custom events you track
- Document all custom dimensions and metrics
- Audit ecommerce properties (if applicable)
- List all custom audiences you’ve built
- Export historical data you might need
Day 3-5: Install alternative
- Set up account in your chosen tool
- Configure equivalent goals/events
- Create custom dimensions
- Set up ecommerce tracking (if needed)
- Run both tools to verify both are tracking
Day 6-7: Plan the cutover
- Decide on a cutover date (end of month is cleanest)
- Communicate to stakeholders: “We’re switching analytics tools”
- Archive GA4 reports (document the last reports)
Week 2: Both tools running
- Monitor both dashboards daily
- Compare session counts, pageviews, top pages
- Check that custom events are firing in both
- Document metric differences
Week 3: Team training
- Show team the alternative’s dashboard
- Walk through key reports
- Explain where to find what they used in GA4
- Collect questions and feedback
Adjust as needed:
- If ecommerce data is off, debug the event mappings
- If traffic sources differ wildly, check bot filtering settings
- If you find blockers, escalate them now (not after cutover)
Day 1: Remove GA4
- Stop firing GA4 events (comment out GA4 tag in GTM/analytics code)
- Keep the alternative running
- Announce to team: “Cutover complete. All analytics now in [tool].”
Days 2-7: Monitor
- Watch for anomalies
- Check that events still fire
- Alert if something breaks
- Have a rollback plan (keep GA4 code commented, not deleted)
Week 2-4: Rebuild dashboards
- Recreate key GA4 reports in the alternative
- Adjust for any data modeling differences
- Share with stakeholders
- Document how to read the new dashboard
Month 1-3: Settle in
- Answer questions as they arise
- Refine your event tracking if gaps emerge
- Optimize custom dimensions based on what matters
- Update any downstream integrations (Slack alerts, etc.)
What You Will Lose
Section titled “What You Will Lose”Be brutally honest with your team about what’s gone.
1. Google Ads Integration (Critical for paid search)
Section titled “1. Google Ads Integration (Critical for paid search)”GA4 can:
- Send conversions back to Google Ads automatically
- Build audiences in GA4, share them to Google Ads
- Optimize Google Ads campaigns based on GA4 value
Alternatives cannot do this. If you run Google Ads:
- You’ll need to manually track conversions via conversion tracking in Google Ads (event-based or click-based)
- You can’t sync audiences from the alternative to Google Ads
- You’ll rely on GA4/Ads native setup, not your analytics tool
This is why many organizations keep GA4 even if they use Matomo or Piwik Pro as their primary tool.
2. Audience Sharing to Other Platforms
Section titled “2. Audience Sharing to Other Platforms”GA4 can share audiences to:
- Google Ads
- DV360
- SA360
- Facebook (via Google Conversions API)
- LinkedIn (via conversion sync)
Alternatives require a separate CDP or manual export. If you need audience activation:
- Piwik Pro has built-in CDP (basic)
- Matomo requires Segment/Tealium
- Plausible/Fathom require manual export
3. BigQuery Export
Section titled “3. BigQuery Export”GA4 exports to BigQuery with a click. Alternatives:
- Matomo: Direct database access (better than BigQuery)
- Piwik Pro: Manual data export or API
- Plausible/Fathom: API or CSV export
If you do advanced analysis in BigQuery, switching to Matomo is easier (you get the raw database). Switching to a lightweight tool requires learning their API.
4. Predictive Audiences & ML Features
Section titled “4. Predictive Audiences & ML Features”GA4 has:
- Predicted churn (customers likely to stop visiting)
- Predicted purchase value (LTV estimate)
- Automatic anomaly detection
No alternative has this. If you rely on GA4’s predictions, you’ll lose them. You’d need a CDP or ML-focused tool like Mixpanel.
5. Cross-Device Tracking
Section titled “5. Cross-Device Tracking”GA4 (with User-ID) can track:
- Same user across devices (phone → desktop → tablet)
- User journey across devices
Alternatives can’t do this reliably. If you need cross-device tracking, you’re stuck with GA4 or a premium CDP.
What You Will Gain
Section titled “What You Will Gain”1. No Data Sampling (Matomo, Piwik Pro)
Section titled “1. No Data Sampling (Matomo, Piwik Pro)”GA4 samples at 10M+ events/month. Alternatives don’t.
If you have high traffic and need accurate data on every segment:
- Matomo self-hosted: 100% data, you own it
- Piwik Pro: 100% data, they own it
- Plausible/Fathom: 100% data, lightweight
2. Full Data Ownership (Matomo)
Section titled “2. Full Data Ownership (Matomo)”Your data is 100% yours. You can:
- Query it directly (MySQL)
- Export it to your data warehouse
- Sell it or use it however you want
- Keep it forever (no 14-month window)
GA4 keeps data on Google’s servers with a 14-month rolling window.
3. Simpler Dashboards (Plausible/Fathom)
Section titled “3. Simpler Dashboards (Plausible/Fathom)”Lightweight tools have 1/10th the features of GA4 but 10x the clarity. If you just care about:
- How many visitors today?
- What are my top pages?
- Where’s my traffic coming from?
Then Plausible or Fathom’s simplicity is a feature, not a limitation.
4. Cookieless Tracking Without Consent (Matomo, Piwik Pro)
Section titled “4. Cookieless Tracking Without Consent (Matomo, Piwik Pro)”Matomo’s cookieless mode:
- No consent banner needed
- Fingerprint-based visitor tracking
- Less accurate but GDPR-compliant
GA4 requires cookies (consent) to do similar tracking.
5. Faster Page Loads
Section titled “5. Faster Page Loads”Lightweight tools (Plausible, Fathom) have ~1KB scripts. GA4’s ~80KB. If you’re obsessed with Core Web Vitals:
- Switching to a lightweight tool gives you a 0.1s LCP improvement
- Not huge, but measurable
When Migration is the Wrong Answer
Section titled “When Migration is the Wrong Answer”Keep GA4 if:
GA4 → Ads conversion import and audience sharing is GA4’s killer feature. Don’t lose it for an alternative.
You have expertise invested in GA4 scripting, custom events, audience logic, and dashboards. Switching costs are high.
GA4 → BigQuery is seamless. Alternatives require more work. Unless you’re switching to Matomo (raw DB access is better).
GA4’s ML features are unique. No alternative has them. If they matter, stay.
GA4’s User-ID feature is still the best. Alternatives can’t match it without a CDP.
Instead of migrating: Run GA4 + an alternative in parallel.
Keep GA4 for Google Ads integration and historical data. Use the alternative for what GA4 doesn’t give you (data ownership, simplicity, cookieless tracking, etc.).
Comparing Alternatives During Migration Evaluation
Section titled “Comparing Alternatives During Migration Evaluation”| Tool | Best For | Hardest Transition |
|---|---|---|
| Plausible | Simplicity, privacy-first, EU | Losing custom dimensions |
| Fathom | Easiest onboarding, fastest load | No ecommerce module |
| Matomo | GA4-feature parity, data ownership | Self-hosting maintenance |
| Piwik Pro | Regulated industries, all-in-one stack | Enterprise pricing, steeper learning curve |
Post-Migration: What Breaks
Section titled “Post-Migration: What Breaks”Expect to fix:
- Slack alerts — If you had GA4 → Slack integrations, set them up in the new tool or via Zapier
- Downstream dashboards — Looker Studio, Tableau, etc. connected to GA4 will break. Rebuild them.
- Custom reports — Custom GA4 reports you’d saved need rebuilding
- Team questions — “Where do I find [metric]?” Answer 50 times a week
- Historical analysis — You can’t compare year-over-year with the old tool (unless you ran parallel long enough to backfill)
Plan for this. Don’t surprise your team.
Recommended Path Forward
Section titled “Recommended Path Forward”If GA4 is working: Keep it. Add an alternative if you need something specific (data ownership, simplicity, etc.). Run both.
If GA4 is broken: Fix it first (custom events, consent mode, etc.). Migration won’t fix broken GA4 implementations.
If your use case doesn’t fit GA4 (data residency, regulated industry, philosophically anti-Google): Migrate to Matomo or Piwik Pro, but keep GA4 for Google Ads integration if you run ads.
If GA4 is simple overill (small site, basic analytics): Migrate to Plausible or Fathom. Simpler tool, simpler life.