Evidence / Experiment-result catalog

abtest.design

Use reported experiments to form a better hypothesis.

Documentation reviewed

Source review
Oct 18, 2026

Understand the resource

What it is.

abtest.design curates reported A/B test outcomes across flows such as onboarding, checkout and paywalls. It is a useful discovery layer for experiments, while each claim needs its original methodology and context before it can inform a decision.

When it is useful

A team is considering a change because another product appears successful.

A proposed application

Put it to work.

This example illustrates a possible use. The proposed workflow and expected result are separate from the bounded checks recorded below.

Try this

Follow a relevant case to its original source, record the audience and metric, then design a test for your own users.

Result to aim for

A specific hypothesis and measurement plan.

Before you start

What you need.

Original experiment details and a valid measurement plan for the target product.

The practical boundaries

What to watch for.

Published winners are a selected sample; effects do not transfer automatically between products.
Sample size, uncertainty, duration and guardrail metrics may be absent, so a percentage alone is insufficient evidence.

Review record

What we could establish.

Public primary documentation reviewed. See the account-free evaluation for executed checks and remaining workflow limits.

Source coverage and runtime coverage are separate

The source-review notes describe documentation and identity. The account-free evaluation below records exactly which public interaction or local fixture was checked, what happened, and what remains untested.

Check current access, terms, licensing and data handling before using a resource with your own material.

Runtime and security evaluation

What the evaluation covers.

Bounded runtime checkedAccount-free · Oct 19, 2026

The VEED pricing-options case opened with its example imagery and related tests. Public case-study navigation worked; its reported uplift was not independently reproduced.

Methods: Public HTTP response · Public browser sample. A completed check covers the stated sample; it is not a safety certification.

Target: https://abtest.design/
Version or capture: Unversioned public page observed on 2026-10-19
Integrity scope: sha256:02a8d096b4112c644ec0fe83e77ba045be6f323cad83635dcdb49e03053d621b — initial HTTP 200 response body only; not whole-application integrity.

Test conditions

Chromium 155.0.8059.82, 1365×900 desktop viewport.
Separate fresh, ephemeral browser contexts for entry and interaction; no owner profile, credentials or stored session.
Only GET/HEAD network requests allowed; forms, downloads, WebSockets and service workers blocked. Some analytics endpoints also blocked.
Short bounded waits; no mobile, assistive-technology, cross-browser or performance certification.

Individual checks and observations

Public HTTP response
pass
Public HTTP response: Anonymous GET with redirects followed, response-body SHA-256 and selected security-header summaries; no cookies retained.
Observed: The standalone client received 200 at https://abtest.design/. This establishes response delivery only.
Fresh browser entry
pass
Public browser sample: Open the catalog URL in a new anonymous Chromium context with read-only network guards and no inherited profile.
Observed: The document returned 200 and identified itself as “abtest.design”. 105 image elements were present; no account was used.
Observed public sample or navigation
pass
Public browser sample: Follow the observed Showing pricing options upfront New VEED • +15% trial conversion link to https://abtest.design/tests/showing-pricing-options-upfront.
Observed: The destination returned 200; title: “Showing pricing options upfront”. 54 page image elements had decoded during the bounded wait. This count is not a catalog size or performance benchmark.

Security findings and boundaries

Passive inspection of the initial public document’s response headers and anonymous browser request hostnames, with read-only guards. No source exploitation, account access, dependency audit or private data.

Observed · info

Public response protections

The initial browser page used HTTPS. Its response included HSTS. No enforcing Content-Security-Policy response header was observed. This records the sampled response only; meta policies, backend authorization and exploit resistance were not assessed.

Observed · info

Request surface in the anonymous visit

abtest.design attempted requests to 6 additional hosts, including www.googletagmanager.com, events.framer.com, framerusercontent.com, fonts.gstatic.com. Host counts include requests stopped by the test guard; they do not establish what data a service retained.

Practical advice · caution

Before using this reference

Before borrowing the change, look for the original sample, duration, uncertainty and guardrails. A selected winner is not a transferable result.

Practical caveats

Before borrowing the change, look for the original sample, duration, uncertainty and guardrails. A selected winner is not a transferable result.
Published winners are a selected sample; effects do not transfer automatically between products.

What remains untested

No experiment was run and no source analytics, account data, statistical calculation or original allocation method was inspected.
Results describe the observed public pages and a bounded action, not the entire product or the referenced designs.

Security limits

No authentication, authorization, tenant separation, data retention or backend handling was tested.
Header presence or absence is a sampled configuration observation, not a security certification or an exploit finding.
6 blocked-request records were retained across entry and interaction checks. Guards may alter normal page behavior; the logs do not represent unrestricted traffic.

Evaluation sources

Download this evaluation record (JSON)

Follow the evidence

Sources and further reading.

Primary resource page

Keep exploring