What automation actually means

Marketing automation is not “send emails automatically”. That is just scheduling.

Automation means: when something happens, do the next thing — and choose what to do based on how the person responded.

Somebody downloads your guide. Three days later they get a follow-up. If they opened it, they get a case study. If they did not, they get a different message with a different subject line. If they clicked twice, a salesperson is notified.

That branching is the difference. You are not sending a campaign. You are building a path with forks in it.

How Odoo builds a campaign

Every campaign has three parts.

The target

Which records the campaign applies to — contacts, leads, or another model — and a filter that narrows it.

For example: leads from the website, created in the last 90 days, not yet qualified.

Get the filter right first. A perfectly designed campaign sent to the wrong list is worse than no campaign, because it burns your reputation with people who never asked.

The trigger

What starts a record on the journey. It can be the moment the record is created, the moment it matches your filter, or a manual start.

The activities

The steps themselves, each with a delay and a condition.

An activity might be: send an email, create a to-do for a salesperson, add a tag, update a field, or move the record to a different stage.

Each one waits for its delay, checks its condition, and acts.

Conditions are the whole point

Every activity can have child activities that fire based on what happened.

The available branches include: email opened, not opened, link clicked, not clicked, replied, or bounced.

A simple example:

• Day 0 — send the introduction email
• Day 3, if opened — send the case study
• Day 3, if not opened — resend with a different subject line
• Day 7, if clicked — create an activity for the salesperson to call
• Day 14, if nothing at all — tag as cold and stop

That is a real campaign. Five steps, three branches, and it treats an engaged person differently from an unengaged one.

Where the data comes from

This is the advantage of running automation inside Odoo rather than in a separate marketing tool.

Your campaign can filter on things a standalone email platform never sees:

• Customers who bought a specific product last year
• Leads in a particular pipeline stage for more than 30 days
• Customers with no order in six months
• Contacts in a specific country or industry
• People who opened a support ticket about a particular issue

No export, no list upload, no sync that drifts out of date. The campaign reads live business data.

Campaigns worth building first

Start with these four. They cover most of the value.

Welcome sequence. New lead arrives, gets a short series introducing what you do. Three emails over two weeks is plenty.

Lead nurture. Leads that are not ready yet get useful content periodically until they engage or go cold.

Re-engagement. Customers with no activity for six months get a check-in. Cheap to run, and it consistently surfaces business you thought was gone.

Post-purchase. After delivery, a message asking how it went. Catches problems before they become complaints and reviews.

Rules that keep it honest

Four rules worth writing down before you build anything.

Every email must be useful on its own. If somebody would not miss it, do not send it.

Set an exit. People who go cold should leave the campaign. Do not keep emailing somebody who has ignored six messages.

Do not run overlapping campaigns. Somebody in three campaigns gets three emails a week from you and unsubscribes from all of them.

Respect consent. Only email people who agreed to hear from you. Include an unsubscribe link and honour it immediately. This is a legal requirement in most countries and a reputation requirement everywhere.

Reports worth using

Campaign flow. How many records are at each step, and where they drop out. The drop-out point tells you which email is failing.

Open and click rates per step. Which messages work. Compare like with like — a subject line test only means something across similar audiences.

Conversions. How many people who entered the campaign became opportunities or customers. The only number that ultimately matters.

Unsubscribes by campaign. A spike means you are sending too much or sending the wrong thing. Treat it as a warning, not a statistic.

What usually goes wrong

Too many emails. The most common failure by a wide margin. Automation makes sending easy, and easy becomes excessive.

No exit condition. Records stuck in a campaign forever.

Filter too broad. Sending to everyone rather than to the people the message actually suits.

Built and forgotten. A campaign written eighteen months ago still running, referencing a product you discontinued.

Automating a bad message. If the email does not work when you send it by hand, automating it just sends a bad email more efficiently.

Getting started

1. Clean your contact data and confirm you have consent
2. Build one simple campaign — a welcome sequence, three emails
3. Run it for a month and look at the numbers
4. Add one branch based on what you learned
5. Only then build a second campaign
6. Review every campaign quarterly and switch off what is not working

Resist the temptation to build five campaigns at once. One campaign you understand is worth more than five you cannot explain.