Comparisons/Free alternatives

Free Google Optimize alternatives after the sunset

Choose Mida for a hosted free experimentation path with a defined traffic boundary, GrowthBook for a free cloud or open-source self-hosted workflow, and Amplitude when experimentation belongs with broader product analytics. VWO is a trial path, not a durable free plan. None reproduces every Google Optimize integration exactly.

Pricing and free-plan details verified September 17, 2026

The current decision

What are you actually replacing?

You need website experimentation after Google Optimize ended, and you want to know which free paths still support a real test workflow.

Google Optimize and Optimize 360 were discontinued on September 30, 2023. The replacement decision is about the workflow you need now, not finding a hidden way to keep the old product running.

Google Analytics Help: Optimize sunset: Google Optimize and Optimize 360 were sunset on September 30, 2023.

Free means different things

A permanent free tier, a limited free plan, and a temporary trial are different buying propositions. This page keeps those states separate.

Decision matrix

Which option fits the constraint?

Product	Free path	What is available	Experiment fit	Implementation constraint	Best fit
GrowthBook Cloud	Free tier with limits	Free Cloud Starter supports up to 3 users, unlimited feature flags, and unlimited experiments.	Feature flags, A/B testing, Experiment analysis, Cloud or self-hosted deployment	The free hosted plan is limited to 3 users. The hosted plan does not remove the need to implement and measure experiments correctly.	Technical teams that want hosted experimentation tied to feature delivery without running the platform themselves.
GrowthBook self-hosted	Open source / self-hosted	GrowthBook documents a free self-hosted path with unlimited users, unlimited feature flags, and unlimited experiments.	Feature flags, A/B testing, Experiment analysis, Self-hosted deployment	Self-hosting requires engineering ownership, deployment, and operational maintenance. You still need a reliable source of experiment metrics.	Teams that can self-host and want control over deployment, data, and experimentation infrastructure.
Mida	Free tier with limits	Free Sandbox includes up to 100,000 monthly tested users, 50,000 daily tracked events, 60 days of retention, 2 active experiences, and 2 active goals per project.	Website A/B testing, Split URL testing, Visual experimentation, Goal measurement	The free plan is bounded by tested users, events, active experiences, goals, and retention. Server-side testing and advanced capabilities are not included in the free plan.	Small marketing sites that need hosted website experimentation without running infrastructure.
Amplitude	Free tier with limits	Free plan includes 2 million events per month, 10,000 monthly session replays, and limited experimentation, guides, and surveys.	Product analytics, Session replay, Feature flags, Web experimentation, Surveys	Experimentation capabilities and volumes are limited on Free. The platform is broader than a simple marketing-page test tool.	Product teams that want experiments measured alongside product analytics and replay.
VWO	Free trial only	VWO documents a 30-day full-featured trial. Its former free Starter plan has been sunset.	Visual experimentation, Targeting, Experiment reporting, Personalization workflow	The current free path is time-limited. A trial is not a durable $0 operating plan.	Teams that need to evaluate a mature hosted experimentation workflow before buying.

This is a fit guide, not a ranking. The right choice depends on the evidence you need and the constraint you cannot ignore.

Evidence and tradeoffs

What each option can and cannot replace

GrowthBook Cloud

Free tier with limits

Official source ↗

What is free

Free Cloud Starter supports up to 3 users, unlimited feature flags, and unlimited experiments.

Capabilities

Feature flags, A/B testing, Experiment analysis, Cloud or self-hosted deployment

Important limits

• The free hosted plan is limited to 3 users.
• The hosted plan does not remove the need to implement and measure experiments correctly.

Fit and boundary

Technical teams that want hosted experimentation tied to feature delivery without running the platform themselves. A strong fit for teams that can supply the implementation and measurement layer. It is not a one-click visual editor clone.

Source checked September 17, 2026

GrowthBook pricing: Free Cloud Starter plan with up to 3 users, unlimited feature flags and experiments, plus a free self-hosted path.

GrowthBook self-hosted

Open source / self-hosted

Official source ↗

What is free

GrowthBook documents a free self-hosted path with unlimited users, unlimited feature flags, and unlimited experiments.

Capabilities

Feature flags, A/B testing, Experiment analysis, Self-hosted deployment

Important limits

• Self-hosting requires engineering ownership, deployment, and operational maintenance.
• You still need a reliable source of experiment metrics.

Fit and boundary

Teams that can self-host and want control over deployment, data, and experimentation infrastructure. A durable free path when infrastructure is acceptable. It is not the lowest-ceremony option for a small marketing site.

Source checked September 17, 2026

GrowthBook pricing: Free Cloud Starter plan with up to 3 users, unlimited feature flags and experiments, plus a free self-hosted path.

Mida

Free tier with limits

Official source ↗

What is free

Free Sandbox includes up to 100,000 monthly tested users, 50,000 daily tracked events, 60 days of retention, 2 active experiences, and 2 active goals per project.

Capabilities

Website A/B testing, Split URL testing, Visual experimentation, Goal measurement

Important limits

• The free plan is bounded by tested users, events, active experiences, goals, and retention.
• Server-side testing and advanced capabilities are not included in the free plan.

Fit and boundary

Small marketing sites that need hosted website experimentation without running infrastructure. Closest hosted fit for straightforward website tests. It does not recreate Google Optimize integrations or remove the need for a sound test hypothesis and measurement plan.

Source checked September 17, 2026

Mida pricing: Free Sandbox plan with 100,000 monthly tested users, 50,000 daily tracked events, 60 days retention, 2 active experiences, and 2 active goals per project.

Amplitude

Free tier with limits

Official source ↗

What is free

Free plan includes 2 million events per month, 10,000 monthly session replays, and limited experimentation, guides, and surveys.

Capabilities

Product analytics, Session replay, Feature flags, Web experimentation, Surveys

Important limits

• Experimentation capabilities and volumes are limited on Free.
• The platform is broader than a simple marketing-page test tool.

Fit and boundary

Product teams that want experiments measured alongside product analytics and replay. Useful when experimentation is one part of a larger product measurement system. It is not the lightest replacement for a single landing-page test.

Source checked September 17, 2026

Amplitude pricing: Free plan with 2 million events per month, 10,000 monthly session replays, and limited experimentation, guides, and surveys.

VWO

Free trial only

Official source ↗

What is free

VWO documents a 30-day full-featured trial. Its former free Starter plan has been sunset.

Capabilities

Visual experimentation, Targeting, Experiment reporting, Personalization workflow

Important limits

• The current free path is time-limited.
• A trial is not a durable $0 operating plan.

Fit and boundary

Teams that need to evaluate a mature hosted experimentation workflow before buying. Useful for evaluation, not for planning a permanent free experimentation stack.

Source checked September 17, 2026

VWO free-plan announcement: The former VWO Testing Starter plan has been sunset; the current free path described is a 30-day trial.

Startup guidance

What changes when budget and sample are small?

• Choose the implementation model before comparing features. A hosted tool reduces infrastructure work; self-hosting trades subscription cost for engineering ownership.
• A free traffic allowance does not make a test valid. You still need enough eligible traffic, a defined outcome, and a decision rule before interpreting results.
• If the real need is a page diagnosis rather than traffic allocation, run the Nebula audit before adding an experimentation layer.

When not to switch

The free option is not always the right option

• Do not migrate just to preserve the old product name. If the test workflow is occasional and the implementation cost is larger than the expected learning, document the decision and wait.
• Do not treat a short VWO trial as a free long-term replacement. Choose a durable free path only when its limits and operating model fit the workload.

Where Nebula fits

Evidence about behavior is not diagnosis

Experimentation tells you what happened when traffic saw different variants. Nebula evaluates the live page conditions that should be understood before deciding what to test or rebuild.

Audit my landing page