FAQ
Frequently Asked Questions
Straight answers to the questions we hear most often. If yours is not below, write to [email protected].
How it works
What does Suadela Engine actually do on my page?
It reads the copy and visitor behaviour on a page you own, generates alternative versions of headlines, subheads, calls-to-action, and body blocks according to codified copywriting principles, deploys those versions as variants on live traffic, measures which performs best, and consolidates the winner into the page. It runs continuously, not as a one-off rewrite.
Do visitors see that a test is running?
No. Each visitor sees a single coherent version of the page, one of the live variants. Switching happens server-side at render; there is no flicker, no opt-in prompt, no test banner.
What does installation require?
A single script tag added to your page, the same way you would install Google Analytics or a session-recording tool. No backend integration and no database changes are required beyond the initial paste.
What kinds of pages does it work on?
Landing pages, sales pages, pricing pages, homepage hero sections, and lead-capture pages; any page whose primary goal is a measurable action such as a sign-up, purchase, demo request, or lead submission. It is not designed for blog articles, documentation, or pages without a conversion event.
What platforms are supported?
Any HTML page we can place a script tag on, which includes WordPress, Webflow, Framer, Wix, Squarespace, Shopify, and custom stacks. Native integrations for WordPress, Webflow, and Shopify ship at launch; on other platforms installation is manual.
Statistics and performance
Does this work if my site has low traffic?
Yes, and low-traffic sites are where Suadela is designed to outperform traditional A/B-testing tools. Classical testing requires large sample sizes to detect small differences because the minimum detectable effect scales as 1/√N (one over the square root of the sample size). On a page receiving a few hundred visits a week, tests of minor tweaks (button colour, single-word swaps) are statistically dead on arrival: you never accumulate enough data to separate signal from noise.
Suadela addresses this by starting with variants that are structurally distinct (different promise, different framing, different structural beats) rather than minor tweaks, so that the winner declares itself even with a few hundred visits. As traffic accumulates, the engine narrows to progressively finer refinements. This sequential screening-then-optimisation strategy is the standard approach in classical experimental design (Fisher 1935; Box, Hunter & Hunter 1978); it is not an AI technique, just one rigorously applied to landing-page copy.
How many variants can run at the same time?
Up to eight during the exploration phase, narrowing to three or four during refinement. Traditional A/B-testing tools (Optimizely, VWO, Google Optimize) split traffic equally between variants, so adding variants divides sample size proportionally; under that regime most tools effectively cap testing at two.
Suadela uses adaptive allocation, a family of methods from multi-armed-bandit theory (Thompson sampling, upper-confidence-bound strategies). Traffic automatically shifts toward variants that are winning and deprioritises clear losers within the first few dozen visits. Under this regime, cumulative regret grows roughly as O(√(K·N·log N)) in the number of arms K rather than linearly, so additional variants are much cheaper than under an equal-split design.
How long before I see results?
Early directional signal is typically visible within the first few hundred visits, usually one to three weeks depending on traffic volume. A statistically consolidated winner for a significant structural change commonly lands in the 2,000-to-10,000-visit range. Smaller refinement tests run later in the lifecycle may require more traffic; the dashboard shows the current confidence level for each comparison in real time.
How does the engine decide when a variant has won?
Through Bayesian posterior probability of winning, not a fixed p-value cutoff. Each variant carries a continuously updated estimate of its conversion rate given the data observed so far. When the posterior probability that one variant outperforms the others crosses a threshold (by default 95 percent, adjustable per customer), the engine consolidates it. This approach is robust to the stopping-bias problem that affects naive p-value-based testing.
What happens if the winning variant stops winning later?
The engine keeps a small rotating challenger cohort that periodically probes adjacent variants. If the current winner is dethroned due to market shift, seasonality, or traffic-source drift, the dashboard flags the reversal and the engine promotes the new leader. You are notified before any consolidation is pushed.
How is Suadela different from just using ChatGPT or Claude to write A/B-test variants?
Three structural differences. Training distribution: general-purpose language models generate from a wide average of public internet text; Suadela generates from a curated knowledge base of the canonical copywriting literature. Scope: a chat session produces a batch of text once, whereas Suadela runs the full generate-deploy-measure-consolidate loop as a single product, including the adaptive-allocation and Bayesian-winner-detection engine described above. Memory: a fresh chat has no record of prior tests; Suadela accumulates aggregated, de-identified performance signal over time.
See the long-form pitch for the full argument and the specific authors and frameworks in the knowledge base.
Copy control and safety
Can I stop the engine from making claims I do not want on my page?
Yes. The dashboard lets you define banned claims, required disclaimers, tone rules, mandatory phrases, and protected copy blocks. These constraints are applied before any variant goes live and are enforced on every generation pass.
What stops the engine from writing something factually wrong?
The engine rewrites only material already present on your page or in the product description you provide. It does not invent features, metrics, testimonials, or prices. Numeric claims, product names, and quoted statements are preserved verbatim; only framing, structure, and persuasive form are varied. Final responsibility for accuracy remains with you as publisher, consistent with Terms section 6.
Who owns the copy the engine generates, and who is legally responsible for it?
You own the copy that runs on your page and may reuse it after your subscription ends. You are the publisher and the legally responsible party for regulatory compliance (advertising, consumer-protection, and sector-specific rules in your jurisdiction). Falco Labs applies your configured constraints on a best-effort basis; see Terms sections 6 and 7 for the full allocation of rights and responsibilities.
Data and privacy
What data is collected about my visitors?
Anonymous on-page interaction signals: clicks, scroll depth, time on section, exit points, and the variant served, tied to a first-party cookie. No personally identifying information is collected unless you explicitly connect a CRM or pass identifiers via the installed script. See the Privacy Policy for the complete list.
Is Suadela GDPR-compliant?
Yes. Under GDPR, you are the data controller and Falco Labs is the data processor. A Data Processing Addendum is available on request. The complete list of third-party subprocessors is published at /legal/subprocessors.html.
Is my page copy stored or used to train other models?
No. Your page copy and your dashboard configuration are used only to operate the engine on your account. Aggregated, de-identified performance data (for example, which categories of headline structure outperform on which traffic sources) feeds back into the methodology, as stated in Terms section 7; the specific text of your variants is never shared with or reused on another customer's account.
Pricing and policies
How much does it cost?
Starter is $347 per month and Pro is $797 per month, billed monthly in USD. Founding-Member early-access pricing is available during the pre-launch window; see Terms section 3.1 for the mechanics.
What does joining the waitlist commit me to?
Nothing. The waitlist requires only an email address. No card is collected, no charge is made, and no subscription is created. Waitlist members receive a launch email describing the Founding-Member window and can decide whether to subscribe at that point.
Can I cancel anytime, and is there a refund?
Yes. Cancellation is one click in the dashboard and takes effect at the end of the current billing period. The first paid subscription month is covered by a 14-day money-back guarantee; see the Refund Policy for the procedure.
What happens to my Founding-Member price if I cancel and re-subscribe later?
The Founding-Member price is forfeited. It is held only for the duration of a continuous, uninterrupted subscription. If you cancel, downgrade, or lapse for more than 30 days and later re-subscribe, the then-current public price applies. See Terms section 3.1.
← Back to home