Every workflow has exactly one entry point, and the Trigger node is it: manual (you press Run) or timer (hourly, daily, weekly, or a custom cron schedule). The choice sounds trivial but it's actually the single biggest factor in whether an automation earns your trust or produces a mess — here's how to use both modes well.
Manual mode: where every workflow should start
Manual mode does exactly one thing: waits for you to click Run. No schedule, no surprises. This should be the mode every new workflow lives in until you've watched it produce correct output more than once — because there's no dedicated approval-gate node in the canvas (see node-based automation, explained), manual mode on the Trigger is your review step during development. Flip to a timer only once you trust what the workflow produces without you watching.
Timer mode: the four options
Hourly — runs once every hour. Right for genuinely time-sensitive signals: breaking news, fast-moving prices, anything where a day-old post would already feel stale.
Daily — once every 24 hours. The most common choice for recap and digest formats — a morning weather card, an evening recap, a daily price snapshot.
Weekly — once every 7 days. Fits weekly-cadence formats: a "this week in X" digest, a weekly milestone check, a Monday price-recap post.
Custom (cron) — anything the three presets don't cover, written as a cron expression.
Reading a cron expression
A cron expression is five fields: minute, hour, day of month, month, day of week — each either a number, a range, a step (*/4), or * for "any." The field defaults to 0 9 * * *, which breaks down as: minute 0, hour 9, any day of month, any month, any day of week — plain-English, "every day at 9:00."
A few that come up constantly in content automations:
| Expression | Meaning |
|---|---|
0 9 * * * |
Every day at 9am (the default) |
0 */4 * * * |
Every 4 hours |
0 9 * * 1 |
Every Monday at 9am |
0 9,17 * * * |
Twice daily, 9am and 5pm |
0 9 * * 1-5 |
Weekdays only, 9am |
*/30 * * * * |
Every 30 minutes |
Matching the schedule to the format, not the other way around
The most common timing mistake isn't a wrong cron expression — it's picking a schedule that doesn't match how often the underlying data actually changes. A few real pairings:
- Breaking news reaction → hourly, or a tight custom interval (
*/30 * * * *) if your niche moves fast. Paired with an If node filtering for genuinely new, relevant stories — polling often is fine as long as most polls simply find nothing worth posting. - Daily price/weather recap → daily, timed for when your audience is actually online, not necessarily midnight or market-open.
- Weekly digest → weekly, on whatever day your format's "week" naturally ends (Friday for a work-week recap, Sunday for a weekend-inclusive one).
- Milestone/threshold posts (see idea #12 in dynamic data automation ideas) → checked more often than they'll actually fire (hourly or daily), relying entirely on an If node to suppress the vast majority of runs where nothing crossed the threshold.
A habit that prevents the most common failure mode
Set the schedule after the workflow is proven on manual runs, not before. A workflow wired straight to an hourly timer from the start means an unnoticed bug produces 24 bad runs before you check on it the next day — the same bug caught during manual testing costs you one run. Timer mode is for automations you trust, not automations you're still building.