Datadog customers
Product and engineering organizations that want experiments, rollout controls, user behavior, session replay, and production performance in one environment.
Independent tool overview
Eppo (now Datadog Experiments) at a glance
Eppo is now Datadog Experiments, an active experimentation product that combines Eppo's warehouse-native statistical analysis with Datadog feature flags, product analytics, RUM, session replay, and observability. New buyers should evaluate the Datadog product and pricing; existing Eppo customers should plan around Datadog's ongoing migration path.
Visit the official Eppo (now Datadog Experiments) site ↗
- Current product
- Datadog Experiments
- Acquisition
- Datadog acquired Eppo in May 2025
- Core use
- Product experimentation, feature rollout, and causal impact analysis
- Data sources
- Datadog analytics/RUM plus supported data warehouses
- Statistical methods
- Sequential, fixed-sample, and Bayesian, with CUPED++
- Public starting price
- $450 per billable experiment per month on annual billing
- Last reviewed
- August 31, 2026
Overview
What Eppo (now Datadog Experiments) is
Datadog acquired Eppo in May 2025 and has since launched Datadog Experiments. The original Eppo site now leads with the successor product, while Eppo documentation and customer workspaces remain relevant during migration. This is a product transition, not a discontinued experimentation capability.
The platform supports randomized A/B and multivariate tests, governed metrics and protocols, diagnostics, progressive delivery, and multiple statistical methods. Teams can measure behavior and performance through Datadog Product Analytics and RUM or connect Snowflake, BigQuery, Redshift, or Databricks to analyze source-of-truth warehouse metrics such as revenue, retention, and lifetime value.
Datadog Feature Flags can assign users consistently to variants and connect experiment state with errors, latency, and session replay. Current documentation also describes analyzing experiments randomized outside Datadog by supplying warehouse exposure data, although the pricing FAQ says Experiments uses Datadog Feature Flags; buyers with an existing flag provider should confirm the supported workflow and total price before migrating.
A trustworthy experimentation program still depends on sound hypotheses, stable assignment identifiers, predeclared primary and guardrail metrics, adequate sample size, and disciplined interpretation. Automated significance, CUPED variance reduction, or Bayesian output cannot rescue poor instrumentation, repeated peeking, changing eligibility, or post-hoc metric selection.
Use cases
Who Eppo (now Datadog Experiments) is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Warehouse-centered data teams
Teams that need experiment metrics calculated from governed business data in Snowflake, BigQuery, Redshift, or Databricks.
Mature experimentation programs
Organizations that benefit from reusable protocols, advanced statistics, diagnostics, metric governance, and cross-team self-service.
Capabilities
Core Eppo (now Datadog Experiments) features
Warehouse-native analysis
Queries assignment and metric data in the customer's warehouse and writes analysis outputs without requiring a separate raw-event copy as the analytical source of truth.
Datadog-native metrics
Builds outcome metrics from Product Analytics and Real User Monitoring, with performance and behavioral context available alongside experiment results.
Feature flags and controlled rollout
Uses deterministic assignment, targeting, environments, feature gates, kill switches, and staged exposure to ship variants safely.
Statistical engine
Supports sequential frequentist, fixed-sample frequentist, and Bayesian analysis, sample-size planning, segmentation, and CUPED++ variance reduction.
Health diagnostics and guardrails
Checks sample ratio mismatch, traffic balance, metric instrumentation, and other conditions that can invalidate or delay an experiment.
AI and personalization experiments
Can compare models, prompts, ranking systems, or user experiences and supports contextual-bandit use cases for real-time personalization.
Process
How the Eppo (now Datadog Experiments) workflow works
Step 1
Write the decision before the test
State the hypothesis, eligible population, unit of randomization, primary metric, guardrails, minimum detectable effect, and ship/stop criteria.
Step 2
Connect trustworthy inputs
Instrument stable assignments and map governed warehouse, RUM, or Product Analytics metrics; validate identities, timestamps, joins, and missing data.
Step 3
QA and canary the variant
Test in a separate environment and with internal accounts, then use limited exposure while checking errors and performance before full experiment enrollment.
Step 4
Run without changing the rules
Hold eligibility, allocation, variants, and decision metrics stable for the planned duration; investigate diagnostics rather than quietly repairing the design midstream.
Step 5
Record and monitor the decision
Document effect sizes and uncertainty, segment findings carefully, ship or stop according to the protocol, and continue watching operational guardrails after rollout.
Cost
Eppo (now Datadog Experiments) pricing and free plan
Datadog lists Experiments from $450 per successfully launched experiment per month when billed annually, or $575 on demand. An experiment becomes billable after at least 500 total subjects and three days. The product page offers a 14-day Datadog trial. Datadog-managed assignment uses Feature Flags, whose separate usage pricing is free below one million monthly flag-configuration requests and then begins at $55 per additional one million on annual billing for the 1M–10M band. Confirm migration terms and all required Datadog products with sales.
Experiments — annual
$450 per billable experiment/month
Published starting price with annual billing.
- Billable after 500 total subjects and at least 3 days
- Includes an allotment of feature-flag credits
- Warehouse and adjacent Datadog product usage may add cost
- Confirm migration pricing for existing Eppo contracts
Experiments — on demand
$575 per billable experiment/month
Published on-demand rate without the annual-billing starting price.
- Same published billability threshold
- Usage-based flag and analytics charges may apply
- Best evaluated against expected experiment concurrency and cadence
Trial
14-day free trial
Datadog advertises a trial of its product suite from the Experiments page.
- Use the trial to validate SDK assignment and metric joins
- Confirm which products and usage are included before enabling production traffic
Pricing checked . Check current pricing at the source ↗
Assessment
Eppo (now Datadog Experiments) strengths and limitations
Where it stands out
- Connects experimental lift with application errors, latency, behavior, and warehouse business metrics
- Supports rigorous, governed methods without forcing every product manager to build custom analysis notebooks
- Warehouse-native mode preserves a company's existing metric source of truth and makes query logic inspectable
- Feature flags, diagnostics, and observation in one platform can shorten the path from safe rollout to a defensible decision
What to consider
- Eppo branding, documentation, workspaces, and Datadog's new product coexist during migration, so current and prospective customers should get the supported path in writing
- Per-experiment pricing can become expensive for high-cadence programs, especially after feature-flag, warehouse-compute, RUM, Product Analytics, or support costs
- The pricing FAQ implies Datadog Feature Flags are required, while product documentation supports analysis of externally randomized tests; confirm the exact entitlement for bring-your-own assignment
- Warehouse-native pipelines still consume warehouse compute and depend on well-modeled, materialized source tables
- Advanced statistics reduce variance and improve monitoring but do not replace experiment-design expertise or protect against selective interpretation
Compare
Eppo (now Datadog Experiments) alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Data Analysis
Databricks Mosaic AI
Consider Databricks Mosaic AI when model development, evaluation, and governance inside a broader lakehouse are more important than a general product-experimentation program.
Explore Databricks Mosaic AI →Coding
Galileo (now Splunk Agent Observability)
Consider Splunk Agent Observability, formerly Galileo, for narrower AI-agent evaluation and observability rather than product-wide randomized testing.
Explore Galileo (now Splunk Agent Observability) →Data Analysis
DataLab
Consider DataLab when analysts mainly need a flexible collaborative environment for custom analysis rather than managed feature flags and experimentation governance.
Explore DataLab →Questions
Eppo (now Datadog Experiments) FAQs
Is Eppo still active?
Yes, but it is transitioning into Datadog Experiments. Datadog acquired Eppo in May 2025, continues to support Eppo customers, and is migrating the product and customer base into its platform.
What happened to Eppo?
Its warehouse-native experimentation, feature-management, and statistics capabilities became the foundation of Datadog Experiments, now integrated with Datadog analytics and observability products.
How much does Datadog Experiments cost?
The published starting price is $450 per successfully launched experiment per month with annual billing or $575 on demand. An experiment is billable after it reaches 500 subjects and runs at least three days; related feature-flag and analytics usage can add cost.
Does Datadog Experiments copy raw warehouse data?
The warehouse-native workflow queries supported customer warehouses and uses them as the source of truth. It needs a service account with read access to relevant inputs and write access to an output schema, so teams should apply least privilege and monitor compute usage.
Can it analyze an experiment assigned outside Datadog?
Current documentation describes a bring-your-own-randomization workflow using exposure data in the warehouse. Because the pricing FAQ also says Experiments leverages Datadog Feature Flags, confirm commercial entitlement and implementation details for your account.
Is this suitable for small websites?
Usually only if the site has enough eligible traffic, meaningful conversion volume, and an experimentation budget. Small samples can leave tests underpowered, while the published per-experiment price is aimed more at established product teams.
Bottom line
Our Eppo (now Datadog Experiments) verdict
Eppo's successor is a strong fit for organizations that already use Datadog or need warehouse-native metrics and rigorous experimentation at scale. The combination of feature delivery, user behavior, production performance, and business outcomes is unusually complete, but small teams and high-volume testing programs should model the per-experiment and adjacent usage costs carefully. Existing Eppo customers should treat migration details—not product viability—as the central diligence question.
Visit Eppo (now Datadog Experiments) website ↗
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.