Rehearse a campaign before you send it

Every pre-send check you probably run answers a question about the email, not about the reader. Does the layout hold in Outlook? Do the links work? Does the message land in the inbox? All useful, and none of them tell you that your subject line reads like a list of chores.

That gap is what a rehearsal is for. In Mailnaptic it is called Rehearse with subscriber digital twins, it sits in the campaign builder, and it is available on the Pro and Business plans. It is labelled Beta in the app, and we will be plain about what that means further down. This post explains what the feature does, how to read the output, and where it should not be trusted.

What your existing checks cover

Most teams stack several checks before a send. Each one is good at one thing, and the trouble starts when we ask it for another. The table below is the version we wish someone had shown us earlier.

Rendering tools such as Litmus show you screenshots of how your email displays across many clients, which is why they catch broken tables and clipped buttons. Seed lists are test addresses at Gmail, Yahoo, Microsoft and others, and Twilio's own guidance on them is blunt: seed accounts do not behave like real humans, they have no engagement history, and results can produce both false alarms and false comfort. They are a directional signal for placement. If placement is what you worry about, start with our guide to the real Gmail and Yahoo sender thresholds instead.

A rehearsal does not replace any of these. It covers the one question they leave open: would this particular audience want to read this particular draft?

What a rehearsal simulates

When you press Run rehearsal, the builder sends the subject, preheader, rendered body HTML, from name and your selected lists and segments to the rehearsal service. It then simulates a sample of representative subscribers drawn from that audience, which is why you should set the audience before you rehearse. A rehearsal run against the wrong segment is a rehearsal of the wrong email.

Each twin reacts the way a person might. In the app the actions you will see are click, reply, read, skim, unsubscribe, and the quiet default of doing nothing. Twins can also point at a friction line (a phrase that put them off) and a hook line (a phrase that pulled them in). Those lines are the most useful output, because they are specific and you can edit them.

The numbers come from a baseline plus the twins. The app anchors the prediction to your audience's own history where it exists: a 180 day average open rate for that audience, or failing that your recent sent campaigns, or failing that industry averages. The twin votes then move the prediction up or down. Once you have rehearsals that were followed by real sends, the app reports that it has calibrated against them and shows how many.

Quick or Thorough, and why the sample is small

The Sample dropdown has two settings: Quick (10) and Thorough (25). A run takes roughly 10 to 20 seconds according to the loading message in the builder.

Ten is small. That is fine for a first pass, because you are hunting for friction, not estimating a rate. It also means the ranges are wide, and the app says so by showing a 95% confidence interval next to each figure. If you see "50% opens, range 24 to 76%", the honest reading is that the sample cannot distinguish a poor subject line from a great one. Use Quick while you are editing and Thorough once, on the version you intend to send.

How to read the result

Read it top to bottom, but spend your time at the bottom.

Then open the individual reactions. They are the closest thing to a read-aloud test with your readers.

Apply the fixes, then run it again

Under Suggested fixes each item has a title, an instruction, an "Applies to" target (subject, preheader or body) and an Apply button.

What Apply does depends on the target, and the difference matters. For a subject or preheader, the app asks the AI rewrite service to rewrite your current text following the instruction and puts the result straight into that field. Check the new text before you move on, because it replaces what was there. For a body fix, the app does not edit your blocks. It generates a rewritten version and tells you to use the block level AI rewrite in the builder, so body fixes are still manual work.

After changing anything, run the rehearsal again. If you do not change anything, the app shows the cached result from the last identical rehearsal and tells you to edit the subject or body to re-simulate. That is sensible, but it means pressing the button twice will not give you a second opinion.

A habit we would suggest: change one thing per fix. If you apply a subject fix, a preheader fix and a rewrite of the first paragraph in one go, you will not know which of them helped. The two lines that work as a pair deserve their own care, which we cover in subject lines and preheaders as a pair.

Comparing two versions

If the campaign has A/B testing switched on, a Compare A/B checkbox appears next to the sample selector. Tick it and the rehearsal runs variant B (its subject, preheader and content) alongside variant A, and a switcher at the top of the results lets you flip between the two, each with its own rates, ranges, fixes and reactions. A/B testing itself, including content variants, is available on Growth and above.

Used well, this is the cheapest way to prune a test before it costs you real subscribers: rehearse both, drop the one that draws unsubscribes, and spend the live test on two candidates you actually believe in. For the live part, read email A/B testing without fooling yourself, because a rehearsal does not remove the need for sample sizes and patience.

What a rehearsal cannot do

This is the section to remember. Digital twins are not people and they have never read your emails before. The feature page says so, and so does the Beta note in the app: ranges are confidence intervals for a small sample, and accuracy depends on how much history your account has.

A good test of whether you are using it right: if a rehearsal result would change what you write, it was useful. If you are using it to feel safe about sending something you already doubted, it was not.

Where it fits in your pre-send routine

Rehearse after the draft is written and the audience is chosen, and before the final review. It is one step in a longer list, and the rest of that list is in nine things to check before you send a campaign. If an AI wrote the first draft, read what AI email writing gets wrong before you apply any suggested fix, since the fixes are AI rewrites too.

If you are still building the rest of your program, the overall sequence is in our 30 day plan for a small business, and rehearsal fits in the week you write your first campaigns.