A/B Testing
Test what works.
No cookies needed.
GhostlyX A/B testing lets you run split tests on any page without cookies, fingerprinting, or personal data collection. Deterministic visitor assignment, Bayesian statistics, and real-time results in a single lightweight script.
Get started freeTotal visitors
Total conversions
Probability B beats A
Visitors
Conversions
Rate
Statistical significance reached. The winning variant can be deployed with confidence. Strong signal detected, but more data is recommended before declaring a winner. Not yet statistically significant. Continue running to gather more data.
Switch between experiments to explore variant performance and Bayesian probability. All data is simulated.
Zero
Cookies or tracking scripts added
Bayesian
Statistical engine with probability reporting
< 1 kB
Experiment script, loaded on demand
Make data-driven decisions without compromising privacy
Most A/B testing tools rely on cookies and third-party scripts to track individual visitors across sessions. GhostlyX takes a fundamentally different approach. Variant assignment is deterministic, based on an anonymous visitor hash and experiment ID. The same visitor always sees the same variant without storing anything in their browser.
Results are calculated using Bayesian statistics with Monte Carlo simulation. Instead of waiting for arbitrary p-value thresholds, you get a clear probability percentage (for example, "97% chance Variant B beats Control") that is straightforward to act on.
Conversion tracking works automatically. Define a goal (a pageview path or custom event), and a background job attributes conversions to the correct experiment variant. No manual instrumentation needed beyond the initial setup.
Deterministic variant assignment
Visitors are assigned to variants using a stable hash function. No cookies, local storage, or browser fingerprinting. The same visitor always sees the same variant.
Bayesian probability engine
Results use Monte Carlo simulation with 10,000 samples and a Jeffreys prior. You see a clear probability that one variant beats another, not confusing p-values.
Automatic conversion attribution
A background job matches experiment visitors against goal completions. Conversions are attributed to the correct variant automatically.
Multiple variants per experiment
Run tests with up to 4 variants on Business plans, or unlimited variants on Scale. Configure traffic allocation per variant.
Pageview and event goals
Track conversions based on visitors reaching a specific URL path or triggering a custom event. Both goal types work out of the box.
Real-time results dashboard
Watch visitor counts, conversion rates, and probability scores update in real time as your experiment collects data.
Privacy at every layer
Traditional A/B testing tools track individual users with cookies and session storage. GhostlyX proves you can run meaningful experiments without any of that.
No cookies or local storage
Variant assignment is computed from a deterministic hash on every page load. Nothing is stored in the browser. There is no cookie consent requirement.
No visitor identifiers
Experiment data is stored with anonymous visitor hashes only. There are no session IDs, user IDs, or cross-site identifiers in the experiment tables.
GPC and DNT respected
If a visitor has Global Privacy Control or Do Not Track enabled, GhostlyX excludes them from all experiments. They always see the control variant.
Opt-out page available
Visitors can exclude themselves from all GhostlyX tracking (including experiments) through a dedicated opt-out page on your domain.
Aggregated statistics only
The results dashboard shows aggregate conversion rates and probabilities. Individual visitor behaviour is never exposed or reconstructable.
Lightweight on-demand loading
The experiment script loads only when experiments are active for the current page. No extra JavaScript is added to pages without running tests.
Set up an A/B test in four steps
Enable experiments on your tracking script
Add data-experiments="true" to your existing GhostlyX script tag. No additional libraries or third-party scripts are needed.
Create an experiment in the dashboard
Give your experiment a name, define the page where it runs, and add your variants. Each variant includes a name and optional JavaScript or CSS to apply when active.
Set a conversion goal
Choose a pageview goal (visitor reaches a specific URL) or an event goal (visitor triggers a custom event). GhostlyX attributes conversions to variants automatically.
Monitor results and ship the winner
Watch your experiment results in real time. Once the Bayesian probability reaches your confidence threshold, deploy the winning variant permanently.
A/B testing plan limits
Business
$29/mo
10 active experiments
Up to 4 variants per experiment
Scale
$69/mo
Unlimited experiments
Unlimited variants per experiment
A/B testing is not available on Free or Pro plans. View all plans
Frequently asked questions
What is GhostlyX A/B testing?
GhostlyX A/B testing lets you run split tests on your website to compare different versions of a page or element. Visitors are assigned to variants using a deterministic hash based on anonymous visitor identifiers, ensuring consistent experiences without cookies or personal data. Results are calculated using Bayesian statistics with probability-to-beat-baseline reporting.
Does GhostlyX A/B testing use cookies?
No. GhostlyX A/B testing uses a deterministic hashing algorithm to assign visitors to variants. The assignment is based on an anonymous visitor hash and the experiment ID. No cookies, local storage, or fingerprinting are used. Visitors see the same variant consistently across page loads without any client-side storage.
How does variant assignment work without cookies?
GhostlyX uses a deterministic hash function that combines the anonymous visitor hash with the experiment ID. This produces a stable number between 0 and 1 that maps to a variant based on configured traffic splits. Because the inputs are stable, the same visitor always gets the same variant, even across different sessions, without storing anything in the browser.
What statistical method does GhostlyX use for A/B tests?
GhostlyX uses Bayesian statistics with Monte Carlo simulation (10,000 samples) and a Jeffreys prior. This approach calculates the probability that one variant beats another, rather than relying on p-values. You get a clear percentage (for example, "97% probability that Variant B beats Control") which is easier to interpret than traditional frequentist methods.
What types of goals can A/B tests track?
A/B tests in GhostlyX can track conversions based on pageview goals (visitor reaches a specific URL path) or event goals (visitor triggers a custom event). Conversion attribution is handled automatically by a background job that matches experiment visitors against goal completions.
Which plans include A/B testing?
A/B testing is available on the Business plan (up to 10 active experiments with up to 4 variants each) and the Scale plan (unlimited experiments and variants). Free and Pro plans do not include A/B testing.
How do I set up an A/B test on my site?
Add the data-experiments="true" attribute to your existing GhostlyX tracking script. Then create an experiment in the dashboard, define your variants, set a conversion goal, and start the test. The tracking script automatically loads a lightweight experiment module that handles variant assignment and applies changes to the page.
Does A/B testing respect Do Not Track and Global Privacy Control?
Yes. If a visitor has Do Not Track or Global Privacy Control enabled, GhostlyX will not include them in any experiments. They will always see the control variant and their visits will not be counted in experiment statistics.
A/B testing that respects your visitors
Run meaningful experiments without cookies, fingerprinting, or personal data. Set up in minutes, get statistically rigorous results.