A/B Split Testing
Find the Follow-Up That Works Better.A/B testing email sequences, with every lead kept on one version.
Not sure which follow-up works better? Run both. An A/B split divides your leads between two versions of a sequence by the weights you set, 50/50 or 90/10. Each person stays on their side from start to finish, and which side they got is recorded.
- 50/50 or any weighting
- Nobody switches sides
- Each side recorded
How it works
- Leads reach the split50/50 by default, or any weights you set
- Weighted by youA 90/10 split is just two numbers
- Each lead gets a sideWorked out from their own record, never redrawn
- Their side is recordedSo you can count both groups
Know Which Follow-Up Works Better.
Each lead’s side is worked out from their own record, so nobody lands on both versions or switches halfway.
Split 50/50 for a straight comparison, or send 10% to something new while 90% stays on what works.
Each side is a full branch with its own emails, waits and questions. One side can simply stop.
Which side each person got is recorded, so you count each group from the record, not from guesswork.
Set Up a Test in Three Steps.
Drop in a Split Step
Add it where the two versions should part ways. It starts with two sides at 50/50.
Set the Weights
Type a weight on each side, like 90 and 10. Each side’s share is its weight out of the total.
Build Each Path
Give each side its own emails, waits and questions. The two paths never have to meet again.
Split the Path, or Hold Some People Back.
Divide people between two paths by weight. Each path has its own emails, waits and branches, and each person keeps their side.
Hold a percentage of people back before they take any step, and compare them with everyone who got the sequence.
How A/B Testing Email Sequences Works in Adoptiv.
A/B testing email sequences means sending two versions of your follow-up to two groups of leads and comparing how each group does. It is how you split test sales emails, timing or channels with real leads instead of opinions.
In Adoptiv you A/B test a sales cadence with a split step in the sequence builder. Each side carries a weight, so a 90/10 split is two numbers, not a different kind of step. Each side is a full path of its own, with its own emails, waits and branches, so you can test a whole approach, not just a subject line.
The side each person gets is worked out from their own record, not a coin toss. Nobody lands on both versions, nobody switches halfway, and running the same step again lands them on the same side. Before moving on, the split records which side it chose, so you can count each group from the record.
For a control group, set a percentage on the sequence to hold people back before they take any step. How to A/B test a sequence well is still your call: Adoptiv runs no significance test and declares no winner. Pair a split with a goal step to record which leads on each side reach a won deal, and run your next sequence experiment from there.
A Fair Test, Built into the Sequence.
50/50 on a new split, or any two weights you type.
Each lead’s side comes from their own record, so it stays the same.
Each side gets its own emails, waits, texts and questions.
Hold a percentage back before the first step, to compare against.
Which side each person got is saved before they move on.
For Teams That Would Rather Test Than Guess.
A Fair A/B Test vs Splitting Lists by Hand.
| What | Two lists split by hand | With Adoptiv |
|---|---|---|
| Who gets which version | Two lists split by hand: Whichever spreadsheet tab they landed in | With Adoptiv: Worked out from each lead’s own record |
| Leads on both versions | Two lists split by hand: Easy to happen by mistake | With Adoptiv: Not possible: each person keeps one side |
| Uneven splits like 90/10 | Two lists split by hand: Count and divide the list yourself | With Adoptiv: Two numbers on the split step |
| Knowing who got what | Two lists split by hand: Rebuilt from sent mail | With Adoptiv: Recorded on each person’s run |
| Picking a winner | Two lists split by hand: Your judgment | With Adoptiv: Still your judgment: no significance test is run |
A/B Split Testing Questions, Answered.
What is A/B testing for email sequences?
Sending two versions of a sequence to two groups of leads, so you can compare which does better. In Adoptiv, a split step divides people between two paths by weights you set.
Can I do a 90/10 split?
Yes. Each side has a weight, so a 90/10 split is just two numbers on the step. A new split starts at 50/50.
Can a lead end up in both versions?
No. Each person’s side is worked out from their own record rather than drawn at random, so they get the same side every time and never switch partway.
Does Adoptiv pick the winner for me?
No. It does not run a significance test, declare a winner or move everyone to the better side. You compare the two groups and decide.
What can I test?
Anything that fits on a path: different emails, different waits, a text instead of an email, or a call task first. Each side is a full branch of the sequence.
What is a control group in a sales sequence?
A share of people held back from the sequence, so you can compare them with people who got it. In Adoptiv you set the percentage on the sequence, and those people stop before the first step.
Is the split random?
No. Each lead’s side is worked out by a fixed calculation on their own record. It divides people by your weights, but never changes, so nobody moves sides.
What should I A/B test in a follow-up sequence?
One thing at a time is easiest to read: a subject line, the gap between steps, a text instead of an email, or a call task earlier. Each side can hold a full path.
Can one side of a test just stop?
Yes. The two sides never have to meet again, and one can lead straight to an end step.
Still have a question? A real person replies in under 60 seconds in US business hours. Talk to our team →
14-day free trial
Find Out Which Follow-Up Works.
Put an A/B split in your next sequence. Free for 14 days, no credit card.
14-day trial · no card · migration included