How to Get Started with Statsig

You can start with Statsig by creating a free account from the homepage, then using the Docs, SDKs, and Integrations links in the site navigation to connect your product and begin running experiments, managing feature flags, and analyzing product data. This path fits teams that want to self-serve; if you need a guided rollout or enterprise terms, use "Talk to Sales" or "Contact Us" instead.

Start with the free account

The homepage presents a single primary entry point: Create a free account. The same page also offers See how it works for a product tour and Contact Us for direct outreach.

What to expect at this stage:

• You are signing up for the platform itself, not a single product — Statsig describes itself as giving your team "5+ products in a single platform."
• The account is the prerequisite for everything else: experiments, flags, analytics, and session replay all live behind it.
• No pricing or payment details are stated on the homepage, so treat cost and plan limits as something to confirm on the Pricing page rather than assuming.

Pick the product you actually need first

Statsig is broad, and trying to adopt everything at once is the most common way to stall. The platform groups its capabilities like this:

Supporting building blocks include Dynamic Configs, Parameter Stores, Autotune, Layers, Holdouts, Power Analysis, Metrics Explorer, Logs Explorer, User Segments, and Events Ingestion.

A practical starting order for most teams: Feature Management → Experimentation → Product Analytics. Flags give you control over what ships, experiments tell you whether it worked, and analytics tells you what to test next. Session Replay and the analytics variants are easier to justify once you have traffic and metrics in place.

Connect your stack with SDKs and Integrations

Two navigation entries do the heavy lifting here: SDKs and Integrations.

• SDKs are how your application reads flags, configs, and experiment assignments at runtime. You'll need to install the SDK for your platform and initialize it with credentials from your account.
• Integrations connect Statsig to the rest of your tooling — data warehouses, event pipelines, and analytics destinations.

The platform also lists Warehouse Native as a platform capability, which matters if your data already lives in a warehouse and you want experiment computation to happen there rather than in Statsig's own pipeline.

Expected result after this step: your app can evaluate a flag or experiment, and events flow back so results can be computed.

Learn the concepts before you scale

Statsig provides several self-serve learning resources directly in the navigation:

• Docs — the reference for setup, SDK usage, and feature behavior.
• Statsig University — structured learning material.
• Sample Size Calculator — useful before you launch a test, to check whether your traffic can detect the effect size you care about.
• Statsig Lite — a lighter-weight option listed alongside the other support resources.
• Blog, Customer Stories, Product Updates — context on how other teams use the platform.

The homepage also cites 50% and 20% as headline figures, and states the platform is "Trusted by thousands of companies, from OpenAI to series A startups." Treat those as positioning claims rather than benchmarks for your own results.

Common sticking points

• Skipping the metric definition step. Experimentation and Product Analytics both depend on a trusted set of product metrics. Build those early; the platform explicitly frames this as "Build a complete set of product metrics."
• Adopting all products at once. The integrated-platform pitch is real, but sequencing matters more than breadth on day one.
• Assuming the free account covers everything. The homepage says "Create a free account" but does not state what that tier includes. Check Pricing before you plan a rollout.
• Ignoring statistical setup. Holdouts, Layers, and Power Analysis exist because naive tests mislead. Use the Sample Size Calculator before launching.

When to talk to sales instead

Self-serve works when you have engineering capacity to wire up an SDK and interpret results yourself. Reach out via Talk to Sales or Contact Us if you need procurement, security review, migration help from another experimentation platform, or a rollout plan across multiple teams.

One note on context: the site states "Statsig is part of the Amplitude family," so if your organization already uses Amplitude, that relationship is worth raising in the sales conversation.