Documentation menu

Journeys

A journey is a multi-step automation: a graph of trigger, wait, branch and send steps a contact moves through over time. It's the same category as ActiveCampaign or Klaviyo's flow builder. The one real difference is the send step — it mails one of your actual interactive email projects, so a welcome series can include a real poll or a scored quiz instead of static copy.

Triggers

A journey enrolls a contact when one of these happens:

  • List joined — a contact is added to a specific list (CSV import, API push, or signup form).
  • Form submitted — a specific signup form is submitted.
  • Segment entered — a contact starts matching a saved segment.
  • Date field — an anniversary/birthday-style trigger anchored to a contact field, recurring every year.
  • Interaction — a specific block (or any block) fires a poll vote, form submit, RSVP, or similar.
  • Purchase — a product-block checkout completes, or a Shopify or WooCommerce order traced back to one of your emails is paid.
  • Checkout abandoned — a Buy Now checkout on one of your own Product block emails was started but never completed, after a delay you set (minimum 5 minutes). This is the own-checkout path: no Shopify or WooCommerce connection required.
  • Store cart abandoned — a connected Shopify store reports a checkout that was started and left, looking back up to 30 days. This is the external-store path.

Cart vs. Checkout — Intent Split: Both triggers can run as separate journeys on the same account. Use checkout abandoned to target friction in your own in-email checkout (FAQs, trust signals, a countdown discount); use store cart abandoned to catch a stall in a connected external store's own shopping cart. The stage field on each abandonment record tracks which kind it is, so analytics and webhooks correctly attribute each event.

  • Meeting booked — a recipient books a meeting from an email's booking block, optionally of one booking type. Later steps can use the tags bookingTypeName, slotStart and meetUrl.
  • Deal stage changed — a deal is created or moved, on the board, by an auto-rule or by another journey, optionally narrowed to one pipeline and stage. The journey runs for the deal's contact, and later steps can use the deal's details as tags (dealTitle, dealValue, dealStage, …).
  • Response completed — a contact answers every block an email counts towards a complete response (the studio's Responses panel sets which ones; see the completion rate), optionally narrowed to one email. It fires once per contact and email, when the last counted block is answered. Changing an answer later doesn't fire it again.
  • Response not finished — a contact answered some of one email's counted blocks but not all of them, a delay you set after their first answer (1 to 168 hours). It doesn't fire if they finish first or the survey closes (a closing date or a response cap), it only watches answers given while the journey is on, and it fires once per contact and email. The usual next step sends the same email again: the answers they already gave are kept, and the page their next answer lands on shows what's left. Both response triggers work only for contacts, that is recipients of an email sent to a contacts list.
  • Store order — a connected Shopify or WooCommerce store reports an order that was paid, or one that was fully shipped, a delay you set after it (0 to 2,160 hours). Narrow it to one store or to orders with certain products (up to 20). It fires once per order, only for a buyer who is already in that store's contacts list. A WooCommerce order counts as shipped when its status is Completed. Later steps can use the tags orderId and orderEvent.
  • API — your own backend enrolls one of its contacts by calling POST /api/v1/journeys/<id>/trigger with { email, listId }, using the same bearer API key as the Send API — for events MailInApp has no way to see on its own.

Steps

Build the graph visually on the canvas. The core steps are send (mail an interactive project), wait (a fixed delay, or "each contact's own best hour" — see send-time optimization), and branch (split on a condition, reusing the same audience-filter builder segments use). Beyond those three you also get split (sends each contact down one of several paths, see the split step), update field, add/remove from list, webhook, and create ticket. A goal step exits the run early once a condition is met (stop an abandoned-cart journey the moment the contact pays, for instance), and exit ends it outright.

Three more steps act on your leads. Notify me alerts you about the contact by push, email or Slack (see lead alerts). Enrich contact fills their blank details from your data providers. Update deal opens a deal for them, or moves theirs to a stage on your pipeline.

Rather not build it by hand? Describe the journey and Mia AI builds it, emails included, and leaves it paused for you to review.

Branch and update-field steps can read either a contact field or the value from whatever interaction enrolled the run. A "send poll → branch on which option they picked" flow needs no field created up front.

There's no dry-run or simulate mode. The only safeguard before a journey goes live is that it always starts paused, so you review the whole graph and its content before switching it on. A step reached twice in the same run — a real branch rejoin, or a loop back to an earlier wait — collapses on the canvas into a "continues at" marker instead of being drawn twice. A genuine graph stays readable that way, rather than turning into an unreadable tree.

The split step

A split sends each contact who reaches it down one of two or more paths. Its Routing setting decides how, in one of three ways.

  • Path 1
  • Path 2

Day 0

50%50%

Day 2

50%50%

Day 7

50%50%

Day 21

50%50%

The shares you set, for as long as the journey runs. Nothing is compared.

Fixed shares: share of new contacts per path
WhenPath 1Path 2
Day 050%50%
Day 250%50%
Day 750%50%
Day 2150%50%
Maximize a result: share of new contacts per path
WhenPath 1Path 2
Day 050%50%
Day 219%81%
Day 712%88%
Day 218%92%
Run an experiment: share of new contacts per path
WhenPath 1Path 2
Enrollment, 20 days50%50%
After the decision0%100%
One split, three ways to route new contacts. Only an experiment ends in a result, and only a person sends everyone down the winning path.

Fixed shares

Each path gets the share you set under Weights (relative share), for as long as the journey runs. Use it to spread contacts over different versions of a sequence. It measures nothing and declares nothing: to compare the paths fairly, run an experiment.

Maximize a result (optimization)

Pick a Result to maximize and how long it counts after the split, between 1 and 720 hours:

  • Clicks a link in an email after the split
  • Answers an email after the split (a vote, rating, form, RSVP or booking)
  • Buys in an email after the split
  • Reaches a goal step, which needs a goal step after the split

As results come in, more new contacts go down the path converting best. Every path keeps at least 5% of new contacts, so it can keep learning if things change. A maximizing split can have up to 8 paths.

This is optimization, not an experiment. The shares change as it learns, so the paths' numbers aren't a fair comparison, and no winner is declared. Learning so far shows each path's contacts counted, how many converted, how many are still in their window and its current share of new contacts, never a rate or a winner. Start learning over clears what it has learned.

Maximizing is included from the Lite plan up. Maximizing for purchases needs the Pro plan or higher. If your plan stops including it, the split goes back to its fixed shares.

Run an experiment (pre-registered test)

Each path becomes an arm of an experiment, and you choose one as the control. Set up the experiment opens its setup, where you fix the question, the metric and how long contacts are assigned before anything is measured. Until the experiment launches, contacts are split by the weights. Once it launches, each contact who reaches the split while it enrolls is assigned a path and keeps it.

A journey can run one experiment split at a time. Experiments are included from the Lite plan up, with as many arms as your plan allows. Experiments inside journeys covers the setup, the results and what happens after the decision.

Sending everyone down one path

After an experiment, you can record a path as the winner. Send everyone down path N then appears on the results page and on the split. From then on, every contact who reaches the split takes that path, and the canvas reads "Everyone → path N (experiment winner)". Contacts the experiment already assigned keep their path.

This never happens on its own. Undo: split by the shares again puts the split back the way it was.

The goal step after a split

A goal step ends a contact's run when its rules match, such as "created a first album". After a split, it's also what the goal results count: Reaches a goal step when maximizing, and Goal rate in an experiment. Goal rate is the share of contacts who reached the split and whose run ended at a goal step within their window.

Safety rails

  • Quiet hours hold a due send until the next allowed hour, checked against one timezone you set for the whole journey (not a per-contact field) — every enrolled contact shares the same quiet window.
  • A per-contact, per-journey cap of 3 sends a day defers rather than drops an over-limit send until the next day.
  • A running experiment keeps its journey. A journey can't be deleted while one of its experiments is observing. A draft experiment is deleted with the journey; a finished one stays, with its results.
  • Auto-disable after 3 failed ticks in a row — a journey whose sends have all failed across a whole processing cycle turns itself off rather than silently burning through your list, the same protection recurring schedules already have.

Plans

Free includes 1 journey with up to 5 steps; every paid tier raises both caps — see pricing for the full table. A journey starts paused when created, so you can review the whole graph before turning it on.