ABsmartly

What Is ABsmartly?

Most A/B testing tools are built for marketing teams who want drag-and-drop experiment creation and pretty dashboards. ABsmartly is not that product, and the company is upfront about it. Founded in Amsterdam in 2020 by engineers who led large-scale experimentation at Booking.com and Netflix, A/B Smartly B.V. (KvK 78151929) built a platform around the statistical and engineering requirements those environments imposed — group sequential testing, warehouse-native architecture, open-source SDKs, and serious statistical controls. The visual editor is absent by design.

The pricing signals the target buyer immediately: €60,000 per year starting point for 50 million events per month. ABsmartly is not competing with Flagsmith, VWO, or Optimizely's growth tier for teams running weekly marketing tests. It competes with Statsig, LaunchDarkly's experimentation layer, and the custom-built experimentation platforms that Booking.com, Spotify, and Zalando previously had to build in-house. The proposition is straightforward: if your team is running experiments at scale and the statistical rigour of standard A/B testing is not sufficient, ABsmartly offers the infrastructure that previously required a data science team to build.

Key Features

Group Sequential Testing

Standard A/B testing uses a fixed-horizon design: you set a sample size before the experiment, run it to completion, and then look at results. Teams routinely violate this by peeking at results early and stopping experiments when they cross a significance threshold — a practice that dramatically inflates false positive rates. Group Sequential Testing is a mathematically rigorous alternative that allows multiple planned interim analyses with pre-specified stopping rules, without inflating error rates.

ABsmartly's implementation can accelerate time-to-decision by up to 80% compared to fixed-horizon designs. For teams running dozens of concurrent experiments on high-traffic properties, this is the difference between weekly and monthly experiment cycles. It is not a bolt-on feature — Group Sequential Testing is the architectural centre of the statistics engine, which is why engineers from Booking.com chose it as the design basis.

Open-Source SDKs

ABsmartly publishes open-source SDKs for JavaScript/TypeScript, Python, Go, Java, .NET, Swift, and Vue 2 on GitHub under the absmartly organisation. Security teams can audit client-side integration code — rare for commercial experimentation platforms and important for regulated industries or organisations with strict third-party code review requirements. Engineering teams can also contribute to SDK codebases, which has practical value for organisations with unusual integration requirements.

The SDK architecture supports on-premise event collection if organisations want to retain full control over the data pipeline rather than routing events through ABsmartly's cloud infrastructure.

Warehouse-Native Integration

Experiment data connects to Snowflake, BigQuery, and Databricks directly, feeding into the analytics infrastructure that engineering and data science teams already use. Rather than maintaining a separate experimentation data silo that requires periodic export to the data warehouse, ABsmartly integrates into the existing stack. Amplitude and Segment connections handle the analytics and CDP layer.

This architecture matters for organisations running complex multi-touchpoint analysis where experiment assignments need to be joined with behavioural data, revenue data, or customer attributes stored in the warehouse — a requirement that Optimizely's self-contained reporting cannot serve.

All-Inclusive Flat Pricing

One plan: all features, unlimited seats, unlimited experiments, unlimited goals. ABsmartly does not gate Group Sequential Testing, multi-armed bandit, or advanced targeting behind a higher tier. The events-based model scales with usage volume rather than punishing growth with per-seat increases. Support is included by phone, email, and Slack with no additional charge — the 60-day proof-of-value engagement helps new customers validate the statistical approach against their real data before committing.

Pricing

ABsmartly pricing starts at approximately €60,000 per year for 50 million events per month. All features are included at this entry point: Group Sequential Testing, multi-armed bandit, real-time analytics, all SDKs, and support. Volume above 50 million events per month is quoted based on scale.

The 60-day proof-of-value period allows enterprise buyers to integrate the SDKs, run test experiments on real traffic, and validate statistical claims before signing the annual contract. For technical buyers evaluating a €60K+ commitment, this is a meaningful risk reduction compared to standard 14-day trials.

Compared to Optimizely, which can reach $100,000+ per year for comparable statistical capabilities at the enterprise tier, ABsmartly's pricing is defensible for engineering teams that use the platform at full capacity. For teams running fewer than 10 experiments per year, neither product is the right fit.

EU Compliance & Privacy

A/B Smartly B.V. is a Dutch company registered under KvK 78151929, operating under EU GDPR by default. EU data residency is standard — not an enterprise add-on that requires custom negotiation. A Data Processing Agreement is available for all customers. The open-source SDK architecture also allows customers to implement on-premise event collection if they want to keep all data within their own infrastructure without routing through ABsmartly's cloud.

The Dutch entity has no US parent and is not subject to US Cloud Act obligations. For European organisations in regulated industries — financial services, healthcare, public sector — that require EU data sovereignty and the ability to audit client-side integration code, ABsmartly satisfies both requirements simultaneously.

Who It's Best For

If your engineering team runs experiments at tens of millions of events per month and your data science team is fluent in Group Sequential Testing, ABsmartly provides the infrastructure without requiring a multi-year build.

If your security policy requires the ability to audit client-side SDK code before production deployment, open-source SDKs on GitHub satisfy that requirement — which very few commercial experimentation platforms offer.

If you are a mid-market SaaS product team running monthly A/B tests on a marketing page, Kameleoon or Webtrends Optimize are more cost-effective and accessible choices. ABsmartly's pricing and developer-only setup are not suited to low-volume experimentation workflows.

If you are evaluating alternatives to a custom-built internal experimentation platform at a company like a large e-commerce retailer or bank, ABsmartly is one of the few European vendors that can replace that infrastructure commercially.

The Verdict

ABsmartly is a contrarian product in the best sense: it deliberately serves a narrow market extremely well rather than competing for the broad middle with compromises. The Group Sequential Testing engine is not a feature checkbox — it is the statistical argument for the product's existence, and it reflects the credibility of its Booking.com and Netflix lineage. Open-source SDKs and warehouse-native architecture close the remaining objections from engineering-led buyers who have historically had to build this infrastructure themselves.

The honest constraints: €60K entry pricing excludes SMBs and most scale-ups. No visual editor excludes non-technical experimentation. A small Amsterdam team running at bootstrap scale means integration breadth and development velocity lag LaunchDarkly or Statsig. For the specific buyer — high-volume engineering team, serious statistical requirements, EU data residency mandatory — ABsmartly is one of very few European vendors that can credibly fill the gap. The 60-day proof-of-value removes most of the procurement risk.

Frequently Asked Questions

What is Group Sequential Testing and why is it the core of ABsmartly?

Group Sequential Testing allows experiment results to be analysed at multiple planned interim points without inflating false positive rates — a statistical problem that standard fixed-horizon A/B testing creates when teams peek at results before completion. ABsmartly's Group Sequential Testing engine can reduce time to a valid conclusion by up to 80% compared to fixed-horizon designs. It was chosen as the architectural foundation because the founders built experimentation systems at Booking.com and Netflix where standard A/B testing at scale produced too many false positives from early stopping.

Does ABsmartly have a visual editor for non-technical users?

No, and this is a deliberate design decision. ABsmartly is built for engineering teams and requires SDK integration for all experiment setup. Non-technical marketers and product managers who want a no-code drag-and-drop test builder should evaluate Kameleoon, Webtrends Optimize, or VWO instead. ABsmartly's buyer is the data scientist or senior engineer who wants rigorous statistical infrastructure, not the growth marketer who wants to swap a headline.

Are ABsmartly's SDKs genuinely open source?

Yes. SDKs for JavaScript/TypeScript, Python, Go, Java, .NET, Swift, and Vue 2 are published under open-source licences at github.com/absmartly. Teams can audit client-side code, run on-premise collection, and contribute to SDK repositories. The server-side platform is commercial SaaS. For security teams that require code audits before production deployment, this makes ABsmartly one of very few commercial platforms that satisfies the requirement.

How does ABsmartly compare to LaunchDarkly?

LaunchDarkly is a US-based feature flag platform with experimentation as a secondary capability — backed by significant VC funding and a marketplace of 100+ integrations. ABsmartly is a Netherlands-based experimentation-first platform with feature flags as secondary. LaunchDarkly is more accessible for mixed engineering and product teams; ABsmartly is more statistically rigorous for data science-led experimentation. LaunchDarkly has much broader integration coverage; ABsmartly has EU data residency by default and open-source SDKs that LaunchDarkly does not.

What does the 60-day proof-of-value engagement include?

The proof-of-value period allows enterprise buyers to integrate ABsmartly SDKs into their real production environment, run test experiments on live traffic, and validate the Group Sequential Testing results against their existing A/B testing data before signing an annual contract. ABsmartly provides technical support and onboarding during the engagement. For a €60K+ annual commitment, this substantially reduces procurement risk compared to a standard 14-day trial on a demo environment.