Segments, Lists and A/B Testing

Sending everything to everybody is the fastest way to lose a subscriber. Admin → Marketing → Audience & Segments is where you cut your list into groups worth mailing: named lists you pick by hand, segments that keep themselves up to date, and tags that feed both. This guide covers building each one, targeting a campaign at several of them at once, and testing two subject lines against each other.

Lists and segments are different things

A list is a group you choose by hand: the members you invited to the preview, the artists in last year's show. You add and remove people yourself.

A segment is a rule. It describes who belongs, and membership updates on its own as subscribers change. "Everyone tagged Collector who has opened something in the last 90 days" is a segment.

Both appear on the Audience segments page, with a subscriber count and when each was last recalculated.

Build a named list

Select New list, give it a name, then search your subscribers and select who belongs. The button confirms the number, for example Add 24 people.

A list is editable after you create it. Open it and use Add people to search and add more, or Remove on any row to take someone out. Each list has a Members view showing the roster with email, name, status and signup date.

You can also fill a list from the subscriber table. Select subscribers there and choose Add to list from the bulk action bar, then pick an existing list or type a name to create one.

Build a segment

Select New segment and add rules. A rule tests one of these:

Tags, Signup Date, Last Activity, Email Engagement, Open Rate, Click Rate, Campaign Interaction, Location, Subscriber Status or a Custom Field.

Combine several rules to narrow the group. Once saved, the segment recalculates and the Last recalculated column tells you when it last ran. Open a segment to see its members, edit its rules with Edit rules, or take the roster away with Export CSV. The same Export CSV button sits on each row of the segments list.

Some segments are maintained by Crafted Call rather than by you. Those open read-only, as View rules.

Tags

Tags are labels on a subscriber, and they are the raw material for most segments. Apply them in bulk from the subscriber table with Add tags and Remove tags, or during an import.

The Tags card on the Audience & Segments page manages the tags themselves:

• Rename changes a tag everywhere it is used. Renaming onto a tag that already exists offers to merge the two.
• Merge folds one tag into another with Merge into.
• Delete removes a tag from every subscriber who has it.

Tags that back a list are protected: they cannot be renamed or merged, because the list's membership is defined by them.

Pick a naming convention before you have fifty tags. Prefixes such as audience- and source- keep the list readable a year later.

Target a campaign

On a campaign's To step, choose all subscribers or pick specific segments. Selecting several segments combines them: anyone in any selected segment receives the campaign, and the step shows the combined total.

Don't send to (optional) takes exclusion segments. Anyone in an excluded segment is skipped even when another selected segment includes them, so exclusion always wins. Use it for a "do not contact" group, or to avoid mailing people who already registered.

Unsubscribed, bounced and unconfirmed subscribers are never included, whatever your segments say. Check the recipient count on the review step before sending.

A/B test a subject line

Open a draft campaign's setup page and turn on Subject line A/B test. Enter Subject Line A and Subject Line B, then set Test Size, the share of recipients each variant goes to. The rest of your audience is the holdout, and the split preview shows the three group sizes from the real send audience.

Subject lines are the only variable you can test. Content, send time and button wording are not A/B variables in Crafted Call, so change one of those in a follow-up campaign rather than expecting a test to isolate it.

Results arrive on the campaign's analytics page: open and click rates per variant, a p-value, and a 95% confidence interval on the difference. A winner is only actionable once the difference is significant, at which point Send winner to holdout becomes available. Turn on Auto-send winner to holdout if you would rather it go out without you.

Small audiences rarely reach significance. A few points of difference on a few hundred recipients usually means nothing, and the results panel says so rather than declaring a winner.

Practical advice

• Keep segments large enough to learn from. Under about a hundred recipients, differences are noise.
• Tag people at the moment they arrive, whether from an import, a signup form or a submission. Retroactive tagging is a chore.
• Re-check a segment's count before a big send. A rule that made sense in March may match almost nobody in September.
• Export a segment before a mailing you care about, so you have a record of who received it.

Where to go next

• Subscriber Management and Signup Forms for adding, importing and cleaning subscribers.
• Importing Your Audience from Mailchimp to bring a Mailchimp audience across with its tags intact.
• Email Campaign Basics to send to what you have built.
• Email Consent by Country for the consent rules that apply where you operate.