How do Scribble automations work?
An automation is a saved rule: what it watches, who it watches, what it sends, how far ahead it fires, what it may cost, and whether it is on. A scheduler runs every rule once a day, looks at the date lead time days from now, and creates one card for each person whose date lands then. Nothing sends the moment you save a rule — it joins the next daily run.
On this page
#The six decisions a rule is made of
| Field | What it does | Default |
|---|---|---|
| Trigger | What the rule watches. See triggers. | Birthday |
| List | Who it watches. Blank means everyone in the address book. Lists change and the rule follows them. | Everyone |
| Message | The text written on the card, with merge fields. | Empty |
| Design | Which card. Blank uses your organisation's default. | Default card |
| Signed by | What {{sender_name}} becomes. | Your organisation name |
| Days ahead | How far in advance of the date the card is made. See lead time. | 5 — too short, raise it |
| Catch-up window | How many missed days a run looks back over. | 7 |
| Cap for the year | The most cards this rule may create in a calendar year. | None |
| On / off | Whether the daily sweep includes it. | On |
#What happens on a run
Work out which dates to act on
The rule takes today, adds days ahead, and that is the occurrence date it is looking for. It then repeats that for each day in the catch-up window, so a run that has been missed is picked up rather than lost.
Find everyone whose date matches
Only contacts in the rule's list, and only ones that actually carry the date the trigger needs. A birthday rule cannot match a contact with no birthday on file.
Check the reasons not to send
In order: the annual cap, the credit balance, the do-not-mail list, and whether the message has a value for every merge field this person uses. Each skip is recorded with a reason against that person.
Claim the occurrence
Before an order exists, the rule inserts a claim row for
(rule, contact, occurrence date). A unique constraint on those three columns means a second attempt collides and stops. This is the entire duplicate guard, and it is enforced by the database, not by code that could race.Create the card
An ordinary order, handed to the same pipeline a consumer checkout uses. One credit is spent, the card gets a code, six production steps and a set of derived dates, and it appears on Sent cards next to everything else.
#Testing a rule without waiting for a date
Use Run now on the automation. It reports who it matched, who it would write to and who it skipped and why. Read the skip reasons before you trust a rule: no credits left, on your do-not-mail list and the message needs {{company}}, which this contact has no value for are all things you would rather find out now than in December.
#Common questions
- How often does the scheduler run?
- Once a day. A rule saved this afternoon joins tomorrow's run.
- Can a person get two cards from one rule?
- No. Every card claims a unique row for the rule, the contact and the occurrence date before the order exists, so a retry, an overlapping run or a manual Run now cannot produce a second card for the same occurrence.
- Can a person get two cards from two different rules?
- Yes. The guard is per rule. If someone is on both a birthday list and a client thank-you list and both fire in the same week, they get two cards. That is usually what you want, but it is worth checking overlapping lists.
- What happens if I run out of credits mid-run?
- Anyone not yet written to is skipped with a reason. They are picked up on the next run once you top up, provided their occurrence date is still inside the catch-up window.
- Do automations cost more than campaigns?
- No. One credit is one card either way, at the rate you bought your credits at.
Last checked against the product on . Something wrong or missing? Tell us.