A/B test your emails and send windows, with automatic winners

"Which subject line works better?" is a question you can only answer by trying both. Until now, trying both inside a campaign meant guessing, or running two campaigns and eyeballing the results. FirstSales now runs the experiment for you. Write two versions, and it splits the sends, measures what actually earns replies, and promotes the winner on its own.

What changed

There are two kinds of test you can run, and both end the same way, with FirstSales picking the winner so you don't have to babysit it.

Content variants

On any email step in a campaign, you can now add a second version, a different subject, a different body, or both, and FirstSales splits your sends across the variants. Add as many as you want to test. Each is labeled by letter (A, B, and so on). There's a guardrail so you can't accidentally delete a step's only remaining version. It works in both campaign modes, autopilot and template, with the panel adapting to the mode you're in.

Send windows

Some audiences reply in the morning. Some reply at night. In a campaign's Settings, you can now turn on a send-window test that pits two different sending time-windows against each other. Set a start and end hour for Group A and for Group B, and FirstSales learns which window your audience actually responds to. The form checks that your windows make sense, so you can't set a start after an end or make both windows identical.

The automatic winner

This is the part that changes how you work. Once a test is running, you don't have to watch it. FirstSales collects real results as your campaign sends, and your per-step analytics show the variants side by side, each as its own row, with a Leading badge on whichever is ahead by reply rate. A send-window test shows a "collecting data" state while it gathers responses.

When one version is a clear, confident winner, and enough messages have gone out for that result to be trustworthy, FirstSales promotes it automatically. The rest of your audience starts getting the winning email, the analytics row updates with a "Winner, auto-promoted" badge, and the losing variant is dimmed and marked "No longer sending." Send-window tests lock in their winner with a confidence figure shown right in analytics.

Want to keep experimenting? A Resume test button on the step undoes the promotion and puts it back into testing.

New

A/B test your emails and send windows, with automatic winners

How to use it

Content-variant test:

1. Open your campaign and go to the campaign builder.
2. Select the email step you want to test.
3. In the email template panel, click Add variant to create version B (A stays your original). Write its subject and body. Add more variants if you like.
4. Keep the campaign running. FirstSales splits sends across the variants.
5. Watch the campaign's per-step Analytics. The front-runner shows a Leading badge, and the system promotes the winner on its own.

Send-window test:

1. Open your campaign and go to the Settings tab.
2. Turn on the send-window test.
3. Set Group A's start and end hour, and Group B's. Make the two windows different and valid.
4. Save. FirstSales sends across both windows and gathers results.
5. Open Analytics to watch it collect, then lock in a winner with a confidence figure.

Why it matters

Most teams know they should A/B test and don't, because the manual version is tedious. Set up two campaigns, wait, remember to check back, read the numbers, manually switch everyone to the winner, hope you didn't call it too early. Every one of those steps is a place to give up. Automating the whole loop (split, measure, decide, promote) means the test happens whether or not you have the discipline to run it by hand, and the winner reaches your audience the moment it's proven rather than whenever you next remember to look.

Calling the winner automatically also protects you from the most common testing mistake: declaring victory too soon, off a handful of sends, on a difference that's really just noise. FirstSales waits until enough has gone out for the result to hold up before it commits, and if you disagree, Resume test hands the decision back to you. It pairs naturally with the template sequence library and with the self-learning AI and AI Instructions that shape your drafts in the first place.