A/B ExperimentsNEU
Every team argues about which subject line, which cadence or which pitch works best, and almost always decides on gut feeling. SellioCRM replaces the argument with measurement on two levels. Inside a campaign flow, any email, WhatsApp or SMS step can carry two to five variants, and the CRM compares open and reply rates per variant. On the Experiments screen, you compare what does not fit in a single message: two cadences, two sales pitches or two landing page versions, measuring conversion. In both cases the split is stable, the same lead always lands in the same variant, and a winner is only pointed out once every variant has at least twenty participants. Applying the winner takes one click.
- A/B testing on a flow message step: two to five variants with their own subject, body and call to action
- Per-variant comparison with open and reply rates, and a suggested winner with a minimum sample
- Cadence, pitch or generic experiments, with Participants, Conversions, Rate and Lift
- Deterministic split: the same lead always in the same variant, never counted twice
- Landing pages outside the CRM measured through the public tracking endpoint
In Sellio kostenlos, Sie zahlen nur für die Maschine
A/B tests and Experiments are part of the CRM and are free as long as your company's infrastructure usage always stays below US$ 20 per month.
Decisions with numbers
Subject line, cadence and pitch stop being opinions and become measured rates.
No false winners
The minimum sample of twenty per variant keeps two lucky sends from becoming house policy.
A fair test
The same lead never receives different versions on resends, and volume is distributed evenly.
Learning applied instantly
Applying the winner takes one click and covers every future participant.
From email to page
The same engine measures messages inside the CRM and pages outside it.
A winner is only suggested once every variant has at least twenty sends or participants; before that the screen says the sample is still short instead of asserting a number the statistics cannot back
The split is not a draw on every send: it is computed from the enrollment or record, so reprocessing the campaign never moves anyone between variants
Applying the winner sends only that variant to future recipients, without touching anyone who already received one
An Experiment's conversion is never guessed: someone records it, and that is what keeps the number auditable
Variants inside the flow
On a campaign flow message step, you write two to five versions, each with its own subject, body and call to action. Personalization variables keep working in each one. The CRM splits leads evenly and shows sent, open and reply figures per variant.
Improvement suggestions
The step comparison points to the next move. Keep running while the sample is still small, apply the winner when there is a clear reply-rate margin, try a bolder change when the rates are tied, and revisit subject and sender when opens are low across the board.
Experiments beyond the message
In Analytics, Experiments, create an experiment of type Pitch, Cadence or Generic with labels for variants A and B. Each participant enters a variant in a stable way, and the Results table shows Participants, Conversions, Rate and Lift against the first variant, which acts as the control.
Landing pages too
To test two page versions outside the CRM, have each page call the public experiment tracking endpoint with your workspace's lead capture token: a view event on load and a conversion event on form submit. The results land in the same table.
- 1
In the campaign flow, open a message step and add two to five variants.
- 2
Let the campaign run and follow sent, open and reply figures per variant.
- 3
For cadences, pitches or pages, create an experiment in Analytics, Experiments, with labels A and B.
- 4
Record conversions and read Participants, Conversions, Rate and Lift.
- 5
When the screen shows a sufficient sample, click Apply winner.
Two email subject lines
Variant A asks whether the quote still stands; variant B offers to hold the price for seven days. With more than twenty sends each, B shows up as suggested thanks to the higher reply rate.
Opening with pain or with a number
A Pitch experiment compares the two call scripts for three weeks and shows a 61% Lift for the second.
Short or long cadence
Two contact cadences enter as variants of a Cadence experiment, and the conversion rate decides which becomes the standard.
Two landing pages
Each page sends view and conversion events to the public endpoint, and the outcome appears in the same Results table.
What is the difference between the campaign A/B test and Experiments?
The campaign test compares versions of one message and measures opens and replies automatically. An Experiment compares anything, such as cadences, pitches or pages, and measures a conversion you declare.
Can a participant switch variants?
No. The assignment is computed from the record or enrollment and kept, even if the campaign is reprocessed.
When does the winner appear?
Only once every variant has at least twenty sends or participants and the best one beats the others. Before that, the screen says the sample is still short.
How many variants can I test?
On a flow message step, two to five. On the Experiments screen, the experiment is created with two variants.
Who can open the Experiments screen?
Anyone with access to the campaigns module. On per-user plans, the A/B testing feature must be included in the contracted edition.
Marketing Pipeline
Qualify every lead with a score you actually understand, know where it truly came from, and follow the funnel and journeys in depth.
Kampagnen-Hub und Anzeigen-ROI
Eine Kampagne, die jedes Asset vereint, mit Budgetsteuerung und ROI pro Anzeige.
E-Mail und Nurturing
Visueller Baustein-Editor, Sequenzen, Journeys und A/B.