Optimization
Let the data pick the winner — and send it.
Compete versions of a broadcast — subject lines, content, or send times — on a slice of your audience. Mumara measures the metric you choose, declares the winner, and sends the rest of the list the version that actually performed.
Just want to break a sent campaign's results down by list, SMTP, link, and browser? That's the A/B Comparison tab — no test required.
How a test runs
Sample, measure, decide, send.
-
Step 1
Build the variants
Create the competing versions on a broadcast — several subject lines, content treatments, or send times. Test more than two; it is not limited to A vs B.
-
Step 2
Configure the test
Choose the sample percentage to test on, the winning metric (open rate or unique click-through), and how long the test runs before a decision.
-
Step 3
Measure on the sample
The sample portion goes out and Mumara tracks engagement across variants in real time until the duration elapses.
-
Step 4
Send the winner
The best performer on your metric sends to the remaining audience automatically — or just reports the result if you would rather make the call.
Why it beats guessing
Opinions about subject lines are cheap. Tests are decisive.
Everyone has a hunch about which subject line will win, and everyone is sometimes wrong. Testing on a sample turns the hunch into a measured outcome before the bulk of the audience ever sees it — so the version most people receive is the one the data actually preferred, not the one that sounded best in the meeting.
And because the winning variant's production send routes through the Bridge and Pool you choose, the proven message goes out on the right reputation lane — the optimization and the infrastructure decision handled together.
Split Tests vs A/B Comparison
Decide before, or understand after.
The two sound alike but do opposite jobs in Mumara ONE — worth keeping straight.
Split Tests is experimental and forward-looking. It sends competing variants to a sample, measures a metric, picks the winner, and sends the rest of the audience the winning version. The point is to choose the better variant before most people see anything.
A/B Comparison is diagnostic and after-the-fact. It takes a campaign you already sent and breaks the result down across lists, sub-campaigns, SMTPs, top-clicked links, and browsers. Nothing is sent as a result — it is a reporting lens.
Use Split Tests when you want the platform to make the call before the bulk send; use A/B Comparison when you want to understand a send you already made. This page is the former.
What it removes
The ways campaigns underperform quietly.
-
Sending the wrong version to everyone
Pick a subject by gut and the whole audience gets the loser. A sample test means the bulk receives the proven winner instead.
-
Only ever testing two things
A plain A vs B limits you to one comparison. Multivariate testing competes several variants at once, so you learn more per send.
-
Manual winner-picking and re-sending
Watching a test and then hand-building the winning send is slow and error-prone. The winner sends automatically the moment the metric decides.
-
Optimizing on bot-inflated opens
A test judged on raw opens can crown the variant bots liked. Pair with engagement verification to test on signal, not noise.
Common questions
What buyers usually ask.
What can I test?
Subject lines, content variants, and send times on a broadcast — and more than two at once, since it is multivariate. The variants compete on a sample of the audience before the full send.
How is the winner chosen?
You pick the winning metric — open rate or unique click-through — and the test duration. When the duration elapses, Mumara declares the variant that performed best on that metric and sends it to the remaining audience automatically.
Can I decide manually instead of auto-send?
Yes. You can let Mumara send the winner out automatically, or configure the test to determine and report the winner so you make the final call before the production send.
How is this different from the A/B tab?
Split Tests competes variants on a sample and sends the winner to the rest — it decides something and sends mail. The A/B tab in a campaign's statistics is a breakdown of an already-sent campaign by list, SMTP, link, and browser — it analyses, it does not send.
Does the winning send use my dedicated IPs?
Yes. The production send of the winning variant routes through the Bridge and Pool you choose, so the proven message goes out on the dedicated IPs you intend.
Related
Pages that pair with this one.
-
A/B Comparison
The diagnostic counterpart — break a sent campaign down by list, SMTP, top clicks, and browser.
Read more -
Broadcasts
Split Tests run on broadcasts — build the competing variants in the broadcast editor.
Read more -
IP Pools
Route the winning variant's production send through a chosen Pool and its dedicated IPs.
Read more -
Mumara AI
Generate subject-line variants to test with the in-editor AI composer.
Read more -
Reporting & Analytics
Test results roll into reporting alongside the rest of the campaign's numbers.
Read more
Mumara ONE · Split Tests
Stop guessing. Start measuring.
Compete variants on a sample, let the metric decide, and send the winner to the rest automatically — on the Pool you choose.