How To Build Behaviour-Based Email Journeys

Behaviour-based email journeys send or withhold messages according to what a person has done, what has changed in their relationship with a business, and what they still need to do. They are more useful than a fixed sequence because they respond to context: a new subscriber needs a different message from a customer who has just placed an order, abandoned an application or stopped using a service.

That does not mean every tracked action deserves an email. A strong journey makes a limited number of well-judged interventions, using signals that are reliable and relevant to the recipient. It also knows when to end, pause or give way to another message. In practice, the hardest parts are usually data quality, message collisions and contact pressure rather than the automation builder itself.

What makes a journey behaviour-based?

A behaviour-based journey has four working parts:

• A trigger: an event or a meaningful change of state, such as creating an account, viewing a product category, beginning a trial or failing to complete a payment.
• Eligibility rules: conditions that decide whether the person can receive the message, including marketing permission, suppression status, geography, account status and whether the event is sufficiently recent.
• Decisioning and timing: logic that selects the next message, delay or branch according to subsequent activity.
• Exit rules: conditions that stop the journey once its purpose has been met, the person has moved into a higher-priority journey, or continued contact is no longer appropriate.

The useful distinction is between an event and an attribute. “Downloaded the implementation guide at 10:14” is an event. “Customer on an annual plan” is an attribute. Both can inform a journey, but events are normally better at explaining why an email should be sent now. Attributes help determine what should be said and whether the message is suitable.

For example, a customer who has added an item to a basket has created an event. A basket reminder may be sensible if there has been no purchase after a suitable interval. But someone who merely viewed a product page once has provided a weaker signal. Treating the two actions as identical creates journeys that feel intrusive and can generate unnecessary mail volume.

Begin with the customer moment, not the available data

Automation platforms make it tempting to start with a list of events and build a branch for each one. Start elsewhere: define the customer moment, the practical obstacle and the help the email can provide.

A useful journey brief answers the following questions in plain language:

• What has happened that makes an email useful at this point?
• What is the recipient trying to achieve?
• What is the single next action that would indicate progress?
• What information, reassurance or reminder would genuinely help?
• What should happen if they have already completed that action before the email is due?

Consider an account-setup journey for a software service. The objective is not “send three onboarding emails”. It may be “help a new user complete the first configuration step within seven days”. The trigger might be account creation; the most important branch is whether configuration has happened. A message that explains the next setup step can be helpful. Repeating the same prompt after the user has already finished it is a data and orchestration failure.

This framing also prevents an automation from becoming a disguised campaign calendar. Behaviour should shape the message, rather than merely provide a pretext to send the next promotional email.

Choose signals with enough meaning and reliability

Not every observable action is a dependable indicator of intent. A journey should favour first-party events that are closely connected to the customer’s task: an order, renewal date, account milestone, support interaction, form completion or confirmed preference change.

Be cautious with signals that can be ambiguous, delayed or generated by technology rather than a person. Email opens, for example, are a poor foundation for high-stakes branching. Google notes that it does not track open rates and cannot verify the accuracy of open rates reported by third parties in its sender guidance. Use stronger evidence where possible, such as a logged-in product action, a meaningful click, a purchase or an explicit preference update.

Before using an event in a journey, document its definition. “Viewed product” may mean a page loaded for one second, a visitor scrolled halfway down it, or a known subscriber viewed it twice within 24 hours. Those definitions lead to very different audiences and very different levels of confidence.

Make permission and suppression checks part of the entry gate

For marketing journeys, permission is not an administrative detail to bolt on after creative approval. It is a live eligibility condition. The journey should check that the person is currently eligible before every marketing send, not just when they originally entered the workflow.

In the UK, email marketing is governed by PECR alongside data protection law where personal information is used. The ICO explains that organisations must be fair, lawful and transparent, and that a person who objects to direct marketing must be removed from marketing use; a suppression or “do not contact” list is needed to make that workable. The ICO’s guidance on electronic mail marketing and data protection is the appropriate starting point for UK teams.

Store the operational evidence as well as the marketing status: when and how permission was collected, the wording or source, the email address concerned, the preference categories selected, and every later opt-out. A central suppression service should take precedence over local lists inside individual platforms. If one system receives an unsubscribe, a separate journey tool must not keep sending because it has not yet caught up.

Tracking can require separate consideration. The ICO notes that tracking pixels are subject to PECR’s rules on storage and access technologies, rather than the electronic-mail rules that apply to the message itself. Explain clearly what behavioural data is collected, why it is used and which choices recipients have; do not assume that a marketing opt-in automatically resolves every tracking question.

Design the logic as a decision system

A journey diagram should show decisions and exclusions as clearly as emails. The simplest way to do this is to define a hierarchy before building individual branches.

Set priority between messages

Some communications should interrupt or suppress others. A password reset, fraud warning, order-delivery update or service interruption notice should not compete with a promotional browse reminder. Equally, an active complaint or unresolved support case can be a reason to pause marketing automation altogether.

Create a contact policy that ranks message types and gives the automation a way to check for conflicts. It should cover both scheduled campaigns and triggered programmes. Otherwise, a person might receive a welcome message, price-drop alert, replenishment reminder and weekly promotion in a single day simply because each team designed its workflow in isolation.

Use waiting periods that match the decision

Delays are not decoration. They allow time for the behaviour to develop and reduce the risk that the email arrives after the person has acted. A failed-payment reminder might need to arrive promptly. A replenishment prompt should be based on a plausible usage window. A high-consideration purchase may justify a longer interval and a softer follow-up.

Every wait should be followed by a fresh check. For a basket journey, that normally means confirming that there has been no purchase, no cancellation and no account-level reason to suppress marketing. For onboarding, it means checking whether the person has completed the milestone the message was intended to encourage.

Give journeys a memory

Without recurrence rules, event-triggered programmes can repeatedly enrol the same person. Decide how often someone may enter, whether a new event restarts the clock, and whether the latest event replaces earlier ones. A product-view journey may permit one enrolment per category in a defined period. An expiry or renewal journey may need a stricter, account-level rule.

Also define expiry. If the action is too old, the message may no longer be useful or accurate. A reminder about a limited stock item several days later can be actively unhelpful if the item has sold out.

Write for the observed behaviour without overstating what you know

Good behavioural email is specific enough to be useful but does not pretend to know more than it does. “Here is a guide to completing your setup” is often better than “We noticed you are struggling”, unless the evidence genuinely supports that conclusion.

Content should answer the recipient’s likely question at that stage. In a welcome journey, this may be what to expect and how to set preferences. In a post-purchase journey, it may be how to get value from the product, access help or manage delivery. In a lapsed-usage journey, it may be a clear reminder of a feature that solves a known task.

Keep the message’s stated purpose honest. Google’s guidance says that marketing and subscribed messages sent at bulk volume to personal Gmail accounts need one-click unsubscribe and a clearly visible unsubscribe link in the message body. It also advises keeping content categories separate, rather than adding promotions to sales receipts. Google’s sender guidelines are a useful operational reference: service updates and promotional messages should be identifiable as such in both message purpose and sending setup.

Protect the journey from data and delivery failures

Automated email magnifies defects. A broken integration can enrol thousands of people into the same journey; a stale order feed can send a post-purchase promotion to someone who has returned an item; an identity-resolution error can join activity to the wrong subscriber. Build controls around the flow rather than relying on someone spotting the problem in a dashboard.

Useful safeguards include a duplicate-event check, maximum daily enrolment thresholds, alerts for unusual event volumes, test records that do not receive live marketing, and a way to pause a journey without deleting its logic. Keep a searchable send log showing the trigger, branch, eligibility result, content version and suppression decision for every message.

Authentication and reputation matter as much as the logic. Google requires all senders to personal Gmail accounts to use SPF or DKIM, while bulk senders need SPF, DKIM and DMARC, aligned authentication for direct mail, and one-click unsubscribe for marketing or subscribed messages. Google also advises keeping reported spam rates below 0.3%, with a lower operational target to provide resilience. These requirements mean that sending more triggered mail is not automatically safer or more effective: poorly targeted automation can create negative feedback at scale.

When a sender is experiencing delivery problems, it is sensible to stabilise authentication, audience quality, complaint signals and contact pressure before expanding behavioural programmes. This is one area where a managed deliverability review can be valuable. Email Foundry’s work on authentication, suppression, provider-level pacing and contact-pressure controls is relevant when the underlying sending environment needs attention; it is not a substitute for a clear customer proposition or lawful data use.

Measure whether the journey helps, rather than whether it merely sends

Start with the outcome the journey was designed to influence: completed setup, retained subscription, resolved payment, repeat purchase, feature adoption or fewer support contacts. Then examine whether the programme creates avoidable friction.

Useful operational measures include:

• the share of eligible people who enter the journey;
• time from trigger to intended outcome;
• conversion or completion after each branch;
• repeat enrolment and duplicate-send rate;
• unsubscribe, complaint and hard-bounce rates by journey and mailbox-provider cohort;
• the number of messages withheld by frequency caps, priority rules or suppression checks.

Do not evaluate a journey only against people who received it. Where practical, compare outcomes with an appropriate holdout group or a controlled alternative. This helps separate genuine incremental value from activity that would have happened anyway. Review results by audience segment and signal quality: a branch may work well for recent customers but poorly for people with weaker, older signals.

Build gradually and review the whole system

A mature programme rarely begins with dozens of elaborate automations. Start with a small number of journeys that correspond to important customer moments and have dependable data: welcome and preference setting, account or product onboarding, post-purchase support, renewal, or a carefully controlled incomplete-task reminder.

Each new journey should be reviewed against the existing contact policy. Ask what it suppresses, what suppresses it, which data it depends on, how it exits, and how a recipient can stop receiving marketing. Revisit those answers when products, consent wording, data sources or mailbox-provider requirements change.

The result should feel less like a machine sending emails after every click and more like a coordinated service: timely when there is a genuine reason to write, quiet when there is not, and trustworthy enough that people continue to welcome the next message.