Probant vs. Kameleoon
Both tools compared point by point, based on Kameleoon’s public documentation, accessed November 7, 2026. Every statement about Kameleoon links to its source.
What Kameleoon offers.
Kameleoon is a French company with an address in Paris.1 It presents itself as a unified platform for experimentation, feature management, and AI personalization, for marketers, product managers, and developers, across web, mobile, and server.2 It highlights prompt-based experimentation (PBX).3 It claims more than 1,000 brands.4
Probant is for sites that want A/B test verdicts that hold up, with a plan set before launch and guardrails monitored through rollout. It measures what each conversion is worth: the margin on an order, a qualified or signed lead, a paid subscription.
The differences, point by point.
On the left, what Kameleoon’s public documentation says, each sentence with its source. On the right, what Probant does. “Not mentioned” means nothing was found in that documentation, which does not prove a feature is missing.
| Criterion | Kameleoon | Probant |
|---|---|---|
| Statistical method | Three approaches to choose from: frequentist, Bayesian, and sequential.5 The documentation presents the frequentist framework as the default.6 In Bayesian mode, non-informative priors by default.7 | Frequentist, with a sequential plan set before launch: target effect, duration, and checkpoints. The verdict fits in one sentence, with a 95% range. Where to check |
| Peeking | The sequential test gives a confidence interval that is valid at any point in the experiment.8 It relies on a Gaussian mixture confidence sequence, from “always valid” inference (Howard et al., 2021).910 | Decisions are only made at planned checkpoints, with O’Brien-Fleming boundaries: very strict at first, close to the usual threshold at the end. During a rollout, ranges are always valid (mSPRT), so you can check them at any time. Where to check |
| Traffic split (SRM) | Native detection with a chi-square test, p-value threshold of 0.001.11 Checks every hour for the first 24 hours, then daily.12 When SRM occurs, the documentation asks you to pause the experiment first.13 | The traffic split is checked with a chi-square test every time the report is read. If the split is off, the test moves to “blocked” status and no winner is declared. Where to check |
| Guardrails | Rollback conditions: Kameleoon pauses the experiment when a goal crosses the set threshold, checked about every 5 minutes.1415 A feature can also be turned off automatically, rule by rule or for the whole environment.16 | Guardrail metrics are tracked with their range; a guardrail is only breached if the whole range crosses the limit. A breached guardrail blocks the test’s verdict, stops a running test on its own, and during a rollout moves visitors back to the original automatically. Where to check |
| Gradual rollouts | Progressive delivery to a growing audience, by percentage or in predefined steps.1718 Rollback based on KPI performance.19 | Steps of 25% → 50% → 100% by default, 10% → 25% → 50% → 100% in cautious mode. Automatic rollback when a guardrail is breached; an exposed visitor stays exposed in the following steps. Where to check |
| Multi-armed bandit | Multi-armed bandit and contextual bandit listed in their plans.2021 | Optimization test: Thompson sampling, reallocated at most once a day, with a minimum share per variant. It maximizes a conversion rate, or an amount per visitor such as revenue: purchase probability and average order value drawn from their distributions, with values capped at a percentile as in an A/B test. Optionally, shares per device (mobile, tablet, desktop): one draw per device, each with its own learning and floor. A visitor always keeps their variant: a new split only applies to visitors who arrive after it. The screen says so: it maximizes conversions (or the amount per visitor) during the test, without proving that a variant is better. The difference is read period by period (and device by device), with no declared winner; the CSV export gives the same reading. Where to check |
| Revenue and margin per visitor | Revenue per visit or per visitor and average order value, with a credible interval in Bayesian mode.22 Winsorization of revenue goals and ratio metrics.2324 Margin per visitor: not mentioned in their public documentation (as of November 7, 2026). | Revenue and margin per visitor, each with its range; the gain is stated in euros. Margin is calculated on the server from reconciled orders, the cost of each product (Shopify or CSV), and store fees; no margin verdict below 90% of revenue covered by known costs. Where to check |
| Winner’s curse | Not mentioned in their public documentation (as of November 7, 2026). | The reported gain is corrected for the winner’s curse by Bayesian shrinkage toward the effects of the store’s past tests. Where to check |
| Bayesian mode | Bayesian mode as one of three approaches.5 Non-informative priors by default.7 | Optional, off by default: a Bayesian reading next to the verdict, with the probability that the variant does better, the credible interval, and the expected loss of each choice. The prior is centered on zero and as wide as the effects measured by the store’s completed tests; while there are too few of them to estimate it, the Bayesian reading is not shown. The decision always follows the sealed plan; when the two readings disagree, the report explains why. Where to check |
| Multiple comparisons | Holm–Šidák correction, chosen for being more powerful than Holm-Bonferroni.25 | Segment comparisons are corrected with the Holm method. With several variants, the error risk is split across comparisons (Bonferroni). Where to check |
| Multivariate testing | All combinations of variations are created automatically.26 More sections require more traffic for a reliable result.27 | Full factorial design generated by Probant: two to four sections, each with two to four levels, up to sixteen combinations, each served as a variant. Main effects of each section across all traffic and interactions between two sections, corrected with the Holm method on the comparisons stated in the plan. Based on the workspace’s measured traffic, the plan says whether the test can reach a conclusion and suggests removing a level or a section; no best combination is declared without a check. Where to check |
| Variance reduction (CUPED) | CUPED available, based on data from existing visitors.28 Only with the frequentist method, not in sequential or Bayesian mode.29 | CUPED applied to live test reports, using each visitor’s history before their first exposure; without history, the difference stays unadjusted and the report says so. Where to check |
| Visual editor | The graphic editor loads directly on top of the site, in the browser, from a page where the script is installed.3031 | Two ways to edit, producing the same changes: a page preview loaded by Probant, without its scripts, or the live page on the site, scripts included. On the site, the editor is loaded by the script already installed, through a short-lived signed link for one draft and one variant, with no browser extension. It only sends the chosen changes to Probant. Where to check |
| Widgets | Widget Studio creates banners, pop-ins, and countdowns, until a date or for a few hours.3233 Free version of Widget Studio included; premium version on request.34 | Banner, countdown, pop-up, badge, reassurance, action bar, delivery date, social proof, promo code, free shipping, and survey, set up without code in the editor and saved in the workspace library for other tests. A widget is just a set of ordinary changes: escaped text, links over https or to the site, and styles applied without a style attribute, so a strict CSP does not need loosening. The countdown follows the site’s time zone and disappears at the deadline; the pop-in opens on arrival, on exit intent, or after scrolling, closes with the keyboard, and does not reappear during the visit if consent allows. Their small modules only load for visitors in the variant. The delivery date is calculated in the visitor’s browser (cutoff time, business days, public holidays); social proof shows the exact number of purchases received over a rolling week, and only above a threshold. The survey asks up to 8 questions (closed answers, recommendation scale, open question) with conditional logic, on the page or through a link outside the site (email, QR code), based on 70 templates (recommendation, satisfaction, effort, checkout blockers, cancellation, concept test, multi-question surveys…); open answers are redacted before sending (email addresses, phone numbers, cards, IBANs), kept for 180 days, and grouped by theme; each question is read by variant, with its range and response rate, and comparisons across all questions are corrected together; a response through a link is attached to the variant the visitor saw, recalculated by Probant, or stays outside the test. Where to check |
| Script and flicker | The asynchronous tag hides the whole page with a CSS rule while the script downloads, for up to 750 ms by default.3536 Kameleoon states that this hiding can produce a blank page and delay First Contentful Paint.37 29 kB compressed application file (Brotli).38 | Script under 7 KB compressed, loaded asynchronously, that does not hide the page while it loads. Server-side assignment available, for a display with no flicker at all. Where to check |
| Granular permissions | Predefined roles (Super Admin, Admin, Technical, Marketing, Production), assigned per user and per project; several roles can be combined.39 | Four fixed roles, plus custom roles the owner builds from a catalog of named permissions: prepare, launch and stop, roll out, approve a launch, write variant JavaScript, manage flags, metrics, margin, members, and keys. No privilege escalation: billing, roles, transfer, and deletion of the workspace stay with the owner; no one can grant more than their own permissions, and an API key cannot exceed its author’s. Where to check |
| Data hosting | Customers are grouped in clusters by country, each physically hosted in its country: French customers share the French cluster.4041 EU or US data residency depending on the plan, or custom on Enterprise.42 City and host of the French cluster: not mentioned in their public documentation (as of November 7, 2026). | Database and compute in Amsterdam, with Scaleway, in the European Union. Scaleway, which hosts the database and compute, is a French company. Published subprocessor list: Scaleway (Amsterdam, Netherlands (nl-ams region), in the European Union), Brevo (European Union: servers in France and Germany, encrypted backups in Belgium (Google Cloud), according to Brevo), Mistral AI (European Union by default (European API endpoint), according to Mistral), ClickHouse (option not used), Stripe (Contract with Stripe Payments Europe, Ltd. (Ireland); data transferred to Stripe, LLC (United States), according to Stripe), Sentry (European Union: Frankfurt (Germany), including backups, according to Sentry). Depending on the feature used, Mistral may temporarily transfer data outside the European Union to the subprocessors it publishes; its data processing agreement limits these transfers to countries the EU recognizes as adequate, or covers them with standard contractual clauses. Stripe transfers payment data to Stripe, LLC, in the United States, under the Data Privacy Framework and, failing that, standard contractual clauses. Whatever region is chosen, Sentry may keep the data of Probant’s Sentry account (members, settings, audit log, project descriptions) in the United States; reported errors stay in Frankfurt. Where to check |
| Consent by default | Several policies: “consent required,” where Kameleoon does not run normally before agreement, or “consent not required,” where it starts right away.4344 A deferred mode shows variants without writing or sending data before consent.45 | Consent required by default: no cookie, no assignment, and nothing sent before the visitor agrees. Consent given mid-page: the page does not change in front of the visitor; their variant starts on the next page. First-party cookie, lasting 390 days at most, never extended. Where to check |
| Pricing model | Entry-level PBX Starter plan from $495 per month, for 10 experiments and 50,000 tested visitors per month.46 Enterprise on quote.47 On Starter, experiments stop when the quota is reached.48 30-day free trial, no credit card required.49 | Creating a Probant workspace is free; paid plans cost €590 to €1,990 per month depending on the number of tested visitors (2 months free with annual billing), with no overage charges. Agency plan: 1,000,000 tested visitors per month, shared across client workspaces, from 3 workspaces, for €1,490 per month or €14,900 per year. Where to check |
What Kameleoon also does.
Probant shares these points with Kameleoon.
An “always valid” sequential test
A confidence sequence valid at any time, from the same family as the mSPRT ranges Probant uses in rollouts.910
A multiple comparisons correction
Holm–Šidák, a variant of the Holm method.25
Native SRM detection
Automatic stops on a threshold
Automatic pause of an experiment, or shutdown of a feature, when a threshold is crossed.1416
Gradual rollouts
Hosting in France
French customers are hosted in a cluster located in France.4041
A “consent required” mode
Kameleoon does not run normally until consent is given.43
When to choose Kameleoon over Probant.
Based on Kameleoon’s public documentation.
You want personalization and feature management alongside tests
Kameleoon combines experimentation, feature management, and AI personalization across web, mobile, and server.2
You want to choose your statistical method
Frequentist, Bayesian, or sequential, your choice.5
You want CUPED in frequentist mode
You want a public entry price and a trial
What sets Probant apart.
These points are not in Kameleoon’s public documentation, accessed November 7, 2026. A missing mention does not prove a feature is missing; if these points matter to you, ask Kameleoon.
Margin per visitor
Probant measures revenue and margin per visitor, each with its range.
KameleoonMargin per visitor is not mentioned in their public documentation (as of November 7, 2026).
Learn moreWinner’s curse correction
The reported gain is corrected by Bayesian shrinkage toward the effects of the store’s past tests.
KameleoonA winner’s curse correction is not mentioned in their public documentation (as of November 7, 2026).
Learn moreSRM that blocks the verdict
A traffic split imbalance moves the test to “blocked” status, and no winner is declared.
KameleoonKameleoon’s documentation describes an alert and asks you to pause the experiment yourself.13 Automatic blocking is not mentioned there (as of November 7, 2026).
A precise hosting location
Database and compute in Amsterdam, with Scaleway, in the European Union. Scaleway, which hosts the database and compute, is a French company.
KameleoonKameleoon hosts its French customers in France.40 The cluster’s city and host are not mentioned in their public documentation (as of November 7, 2026).
Frequently asked questions.
Does Kameleoon run sequential tests, like Probant?
Yes. Its sequential mode gives an interval that is valid at any point in the experiment.8 The documentation presents the frequentist framework as the default.6 Probant sets its checkpoints before launch (O’Brien-Fleming) and uses always-valid ranges during rollouts.
Where does Kameleoon host data?
French customers are hosted in the French cluster, located in France.4041 The plans offer EU or US data residency.42 Probant hosts its database and application in Amsterdam, with Scaleway.
Sources.
Official pages from Kameleoon, accessed November 7, 2026. Quotes are reproduced in their original language.
- Source 2: kameleoon.com/llms.txt
Kameleoon is a unified experimentation, feature management, and AI-driven personalization platform used by marketers, product managers, and developers to run A/B tests, manage feature flags, and deliver personalized experiences across web, mobile, and server-side environments.
- Source 3: kameleoon.com/llms.txt
Kameleoon pioneered Prompt-Based Experimentation (PBX), an AI capability that lets teams build, launch, and analyze experiments from natural-language prompts.
- Source 5: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/choosing-the-right-statistical-method-for-ab-testing
Kameleoon supports three statistical approaches for interpreting your results: Frequentist, Bayesian, and sequential testing.
- Source 6: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/statistical-significance
To help you decide, Kameleoon uses the frequentist statistical framework, which assesses how likely the observed data is if the "null" hypothesis is true.
- Source 7: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/bayesian
However, Kameleoon's A/B testing uses non-informative priors by default.
- Source 8: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/choosing-the-right-statistical-method-for-ab-testing
Provides a valid confidence interval at any point during the experiment's runtime.
- Source 9: storage.googleapis.com/kameleoon-storage-documentation/user-manual/pdf/statistics-at-kameleoon.pdf
Kameleoon has chosen to use a Gaussian mixture asymptotic confidence sequence for its sequential tests
- Source 10: storage.googleapis.com/kameleoon-storage-documentation/user-manual/pdf/statistics-at-kameleoon.pdf
It is based on always valid inference which was proposed by Howard et al. (2021)
- Source 11: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/sample-ratio-mismatch
Kameleoon uses a threshold of 0.001 for the p-value; an experiment tests positive for SRM if the p-value falls below this threshold.
- Source 12: docs.kameleoon.com/user-manual/experiment-analytics/analyze-results/results-page-tabs/health-menu
They run hourly during the first 24 hours after you activate an experiment, and daily afterward for all active experiments.
- Source 13: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/sample-ratio-mismatch
Open the experiment Results page. Pause the experiment first to prevent further biased data collection.
- Source 14: docs.kameleoon.com/user-manual/experimentation/web-experimentation/configure-and-launch/automatically-rollback-an-experiment
Define a performance threshold for a goal, and Kameleoon automatically pauses the experiment when the goal's performance crosses that threshold
- Source 15: docs.kameleoon.com/user-manual/experimentation/web-experimentation/configure-and-launch/automatically-rollback-an-experiment
Kameleoon checks each condition approximately every 5 minutes.
- Source 16: docs.kameleoon.com/user-manual/experimentation/feature-experimentation/using-the-rollout-planner/validation-and-rollback/automatically-rollback-a-feature
Choose whether to turn off only the selected rule or turn off the toggle for the entire environment.
- Source 17: docs.kameleoon.com/user-manual/experimentation/feature-experimentation/using-the-rollout-planner/rules-and-targeting/define-progressive-delivery-rules
Progressive rollouts let you spread out large feature releases by gradually rolling them out to an increasing audience.
- Source 18: www.kameleoon.com/plans
Percentage-based rollouts or set pre-defined stages and timelines for rollouts
- Source 22: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/bayesian
When you hover over non-binomial metrics, such as Revenue per visit/visitor or Average cart value, an overlay appears.
- Source 23: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/winsorization
If your custom goal includes a Revenue metric, Kameleoon applies Winsorization to it.
- Source 25: storage.googleapis.com/kameleoon-storage-documentation/user-manual/pdf/statistics-at-kameleoon.pdf
we chose to implement the Holm–Šidák method as it is more powerful than the Holm-Bonferroni method
- Source 26: docs.kameleoon.com/user-manual/experimentation/web-experimentation/advanced-experiment-types/setting-up-multivariate-tests
Kameleoon automatically creates all possible combinations from the variations you design.
- Source 27: docs.kameleoon.com/user-manual/experimentation/web-experimentation/advanced-experiment-types/setting-up-multivariate-tests
the more sections you create, the more traffic you need to get usable, reliable results.
- Source 28: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/cuped
Controlled-experiment Using Pre-Experiment Data (CUPED) uses data from existing visitors to accelerate reliable experimentation and personalization results.
- Source 29: docs.kameleoon.com/user-manual/experiment-analytics/statistical-methods/cuped
CUPED only works with the Frequentist statistical method. It doesn't work with Sequential testing or Bayesian.
- Source 30: docs.kameleoon.com/user-manual/experimentation/web-experimentation/graphic-based-experiments/getting-started-with-the-graphic-editor
It loads directly on top of your website in the browser, letting you click on any element and modify it without writing code.
- Source 31: docs.kameleoon.com/user-manual/experimentation/web-experimentation/graphic-based-experiments/getting-started-with-the-graphic-editor
Navigate to a webpage on which you have installed the Kameleoon script.
- Source 32: docs.kameleoon.com/user-manual/assets/widgets/create-a-widget-with-the-widget-studio/create-a-widget
A widget is a banner, pop-in, countdown, survey, or other piece of content
- Source 33: docs.kameleoon.com/user-manual/assets/widgets/create-a-widget-with-the-widget-studio/create-a-widget
choose Till date to end the countdown on a date several days away, or Timer to count down in hours.
- Source 35: docs.kameleoon.com/developer-docs/web-experimentation/technical-concepts/flicker-management-and-performance
The snippet includes a “blocking CSS rule” that initially hides all the content on the page while the browser downloads the script.
- Source 36: docs.kameleoon.com/developer-docs/web-experimentation/technical-concepts/flicker-management-and-performance
By default, this variable has the value 750 milliseconds.
- Source 37: docs.kameleoon.com/developer-docs/web-experimentation/technical-concepts/flicker-management-and-performance
Hiding the page until the Kameleoon application file loads or the timeout expires can cause a white page and a slower First Contentful Paint (FCP).
- Source 38: docs.kameleoon.com/developer-docs/web-experimentation/technical-concepts/flicker-management-and-performance
The application file, created using TypeScript, has a size of just 29 kB (compressed with Brotli), which minimizes loading time on the first page.
- Source 39: docs.kameleoon.com/user-manual/account-and-team-management/users-and-teams/user-roles
You assign roles per user and per project. If a user has multiple roles, Kameleoon merges their permissions and grants the combined access.
- Source 40: docs.kameleoon.com/developer-docs/web-experimentation/self-hosting/simple-self-hosting-and-on-premises-options
Kameleoon groups customers into clusters based on their geographic location (country). For example, French customers share data within the French data cluster, while German customers share data within the German clusters.
- Source 41: docs.kameleoon.com/developer-docs/web-experimentation/self-hosting/simple-self-hosting-and-on-premises-options
Kameleoon always physically hosts a country's cluster in a data center on that country's soil to ensure data confidentiality and compliance with regional data privacy laws.
- Source 43: docs.kameleoon.com/developer-docs/privacy-and-compliance/consent-management
Consent required: With this policy, consent must be explicitly obtained from the visitor. Until it is given, the legal consent will be considered to be "unknown" and Kameleoon will not operate normally.
- Source 44: docs.kameleoon.com/developer-docs/privacy-and-compliance/consent-management
Consent not required: With this policy, consent is not explicitly required for this website and module. Kameleoon assumes it is granted automatically and starts immediately in full-featured mode.
- Source 45: docs.kameleoon.com/developer-docs/privacy-and-compliance/consent-management
In this mode, Kameleoon does not send any data to tracking servers nor write any data to the device. However, it will still display experiments (or personalizations) normally, if the visitor triggers the associated segment.
- Source 46: www.kameleoon.com/plans
Starting from $495/month Build up to 10 experiments and test 50,000 visitors per month.
- Source 47: www.kameleoon.com/plans
Custom Unlimited Experiments. Unlimited tested traffic. Advanced capabilities.
- Source 48: www.kameleoon.com/plans
At 90% of your quota, you’ll be notified by email and in-app. At 100%, experiments stop running.
Migrate from Kameleoon
Your Kameleoon history can be imported and is reread by Probant’s engine: every verdict is recalculated from visitors and conversions.
- Open demo, no sign-up
- Free to start, no overage charges
- Database and compute in Amsterdam, with Scaleway