- Most A/B testing tools build parallel tracking that creates data discrepancies with your analytics
- Varify.io has no own tracking — it uses your existing GA4, BigQuery, Matomo, Piwik Pro or PostHog
- One data source = zero discrepancies, no extra cookies, no additional consent needed
- BigQuery integration gives exact numbers (no HyperLogLog++ estimates) with product-level filtering
Most A/B testing tools build their own tracking parallel to your existing analytics — creating duplicate snippets, cookie overhead, and unavoidable data discrepancies between your testing numbers and your analytics numbers. Varify.io takes a fundamentally different approach: it has no own tracking. Varify delivers the experiment variants, but measurement happens entirely in your existing analytics tool — Google Analytics 4, BigQuery, Matomo, Piwik Pro, or PostHog. That means: one data source, no data discrepancies, no additional consent needed.
Other tools also offer analytics integration, but at different depths: Optimizely and VWO have GA4 connectors alongside their own tracking, Convert offers 90+ integrations, and Amplitude/PostHog replace your analytics entirely. The right approach depends on whether you want to keep your existing analytics or replace it. For a broader comparison, see our full A/B testing tools guide.
Three approaches to analytics integration in A/B testing
Approach 1 — Parallel Tracking (most tools)
The testing tool builds its own tracking alongside your analytics. Examples: VWO, AB Tasty, Kameleoon, Crazy Egg. Advantage: Independent from your analytics setup, own dashboards, full control over data collection. Disadvantage: Two tracking systems = two truths. Numbers in your testing tool will never exactly match your GA4 numbers (different cookie logic, session definitions, attribution). Plus: more JavaScript, more cookies, more consent requirements.
Approach 2 — Analytics-Native (Varify's approach)
The testing tool only delivers variants; measurement is delegated entirely to your existing analytics. Examples: Varify (GA4, BigQuery, Matomo, Piwik Pro, PostHog). Advantage: One data source, zero discrepancies, no extra tracking snippet, no additional consent. Your testing results live in the tool you already trust. Disadvantage: You need a working analytics setup. If your GA4 is poorly configured, your testing results will reflect that.
Approach 3 — Analytics Replacement (all-in-one platform)
The testing tool IS your analytics. Examples: PostHog, Amplitude. Advantage: Maximum integration — experimentation, analytics, session replay, feature flags in one tool. Disadvantage: You must migrate your entire analytics. For teams already running GA4, Matomo, or Piwik Pro, that's a massive switching cost.
The right approach depends on one question: Do you want to keep your existing analytics, or are you willing to replace it? If keep → Approach 1 or 2. If replace → Approach 3.
How deep does the integration go? — Comparison table
| Integration | Varify.io | Optimizely | VWO | Convert | PostHog | Amplitude |
|---|---|---|---|---|---|---|
| GA4 | Native | Connector | Parallel | Parallel | Import | |
| BigQuery | Native | Webhook | Sync | |||
| Matomo | Native | Limited | Limited | |||
| Piwik Pro | Native | Limited | Limited | |||
| PostHog | Native | Built-in | ||||
| Own tracking | None | Yes | Yes | Yes | Yes | Yes |
| Data discrepancies | None | Possible | Likely | Possible | If parallel | If parallel |
| Extra cookies | None | Yes | Yes | Yes (1st-party) | Configurable | Yes |
Source: Claude Research, May 1, 2026
Varify.io
Varify's analytics-native architecture is unique on this list: no other tool delegates measurement entirely to existing analytics. This solves the data discrepancy problem every company knows. The trade-off: you need a cleanly configured analytics setup. Varify is as good as your GA4/Matomo — not better, not worse. The BigQuery integration raises the ceiling: raw data access without SQL, exact numbers instead of estimates.
VWO
VWO has a solid GA4 integration but runs its own tracking as the primary data source. In practice: your VWO dashboard shows different numbers than GA4. That's not a bug — it's two measurement systems with different session logic. Teams using VWO must decide which source is the truth.
Convert
Convert has the broadest integration catalog (90+ tools) and minimizes discrepancies through a consistent first-party cookie approach. For teams connecting many analytics and marketing tools, Convert is strong. But the integration is additive — Convert tracks in parallel, it doesn't replace your analytics.
PostHog
PostHog takes the most radical approach: it wants to replace your entire analytics. Experimentation, product analytics, session replay, feature flags — all in one platform. If you're ready to retire GA4, the integration is perfect. If you want to keep GA4, PostHog isn't an analytics integration — it's an analytics replacement. That's a fundamental decision, not a tool choice.
Amplitude
Amplitude offers the deepest combination of analytics and experimentation. Like PostHog, it wants to be the primary data platform. For product teams needing retention, funnels, and experiments in one interface, it's strong. For marketing teams wanting to keep GA4, the migration effort is too high.
Optimizely
Optimizely offers enterprise-grade experimentation with a GA4 connector, but analytics integration is an add-on, not the core architecture. Experiment results live primarily in Optimizely's own dashboard. For companies already using Optimizely as their experimentation platform, it works. As a pure analytics integration solution, it's oversized.
A/B testing that uses your existing analytics. Not another tracking tool.
Varify connects to GA4, BigQuery, Matomo & more — no parallel tracking, no data gaps. From €149/month.
Varify's analytics integrations in detail
GA4 Integration
All GA4 events and metrics are automatically available as experiment goals via the Tracking Setup Wizard. No manual goal setup — everything you track in GA4 can be selected as a goal in Varify. Supports server-side tracking via GTM. Varify's own reporting shows daily progress graphs, statistical significance, and CSV export.
BigQuery Integration
Direct access to GA4 raw data — without writing SQL. This is the most precise evaluation method for A/B tests because BigQuery stores raw events instead of aggregated data. No HyperLogLog++ estimation (like GA4's standard reporting) — exact numbers. Real-time data collection. Especially relevant for e-commerce tests with product-level metrics: product-level filtering, outlier smoothing, duplicate user event exclusion.
Matomo Integration
Direct evaluation in Matomo — setup in a few clicks. For European companies that replaced GA4 with Matomo for privacy reasons, this is the seamless testing complement.
Piwik Pro Integration
Native evaluation support. Piwik Pro is particularly widespread in regulated industries (finance, healthcare, public sector).
PostHog Integration
Events appear in Varify approximately 3–4 minutes after triggering. Full event and click metrics.
The data discrepancy problem — and how to solve it
The problem: You read in your testing tool "Variant B has 15% more conversions" and check GA4 — there it's only 8%. Which number is right? Both. And neither. The discrepancy arises because two tracking systems with different cookie logic, session definitions, and attribution models measure independently. It's not a configuration error — it's a systemic effect of parallel tracking.
What teams typically do: Declare one source the "truth" and ignore the other. Build expensive reconciliation processes. Or worse: delay decisions because nobody trusts the numbers.
Varify's solution: By having no own tracking, the problem doesn't exist. There's only one data source — your analytics tool. The numbers in Varify's reporting ARE the numbers from GA4/Matomo/BigQuery. No reconciliation needed.
Honest limitation: This approach assumes your analytics is cleanly configured. If your GA4 setup has issues (missing events, wrong attribution, consent gaps), Varify shows the same issues. That's why Varify recommends an A/A validation test across your domain before your first real test.
