Ecommerce email flows

What to automate and what to keep as a campaign

A campaign starts with an editorial decision: communicate a launch, a selection or a promotion to a particular audience. An automation starts with a customer’s situation. The question is not how many emails you can prepare in advance, but which conversation should begin when that situation occurs and when it stops making sense.

In a fashion store, a collection launch might be a campaign. Helping a new subscriber understand sizing could belong in a welcome sequence. Both messages may reach the same person, but serve different purposes. The calendar and flows need shared planning so the result does not feel like separate departments sending independently.

Think of each automation as a decision the team should be able to explain: who enters, what they need, which information is known and which change requires the message to stop or adapt. Drawing more steps does not improve that decision by itself. Without a distinct reason for the next email, there is no next step to write yet.

A map for choosing your store’s flows

Use this table as a working starting point. Each row connects a situation with a task for the message and a condition requiring the journey to be reviewed. It does not prescribe an email count or a universal delay: those decisions depend on the catalogue, where the contact came from and available data.

The guides to welcome, abandoned cart, post-purchase, browse abandonment and winback explain the conditions and examples for each journey.

Prioritise need, data and execution capacity

Before ordering the work, identify the problems occurring in your store. Forms may collect contacts without a useful welcome, buyers may ask how to use a product, or cart reminders may arrive after purchase. Each calls for a different intervention. Priority should not be determined solely by the sequence most commonly found in a template.

For each candidate journey, compare customer need, how often the situation occurs, data quality, the content you can produce and the cost of getting it wrong. A flow for very few people may solve an important problem; one with a large audience may be poorly conceived if you cannot identify their situation accurately enough.

For example, if a store repeatedly receives questions about garment care, post-purchase content could have a clear purpose. If the event supporting a cart reminder arrives inconsistently, that data should be fixed before expanding reminders. This example illustrates a decision criterion, not a result achieved by Abalola.

Choose a scope you can review from end to end. It is better to record why a journey remains pending than publish it without a verifiable definition. An email marketing strategy helps order these decisions and assign owners before turning them into design and configuration tasks.

Check the data and whether sending is appropriate

For each entry condition, find an actual example of the supporting data. Check when it arrives, which person it belongs to and what it contains. An event with a familiar name does not prove that it represents precisely the action you imagine. The store platform, integration and configuration can change what is received.

Separate observation from eligibility. Knowing that someone showed interest does not by itself establish whether a communication should be sent through a particular channel. Document how preferences and exclusions are managed and which system is responsible for preserving them when data synchronises.

In Klaviyo, profile filters are checked again during the journey, whereas trigger filters are not re-evaluated at sending. That technical distinction matters when a later purchase should change the decision.

Customer segmentation should reflect conditions you can verify. Give each rule a business explanation, an example that should enter and one that should remain outside. When data is missing, decide how the journey should behave instead of allowing an empty variable to determine the content accidentally.

Set priorities between messages

Imagine someone subscribes, views a product, begins checkout and purchases on the same day. Viewed separately, every automation may seem reasonable. From their inbox, they may tell incompatible stories. Coordination requires looking at the complete sequence of actions.

Write down which conversation should change when a new signal appears. A reminder assuming a pending purchase needs reconsideration after the order. A winback invitation needs review if the customer has just returned. Spacing contradictory messages apart is not enough: first fix the condition that makes them relevant.

Include scheduled campaigns in this review. Identify who may receive them while in each journey and which exclusions the specific proposition needs. Not every campaign needs identical rules; an editorial communication and a limited offer may call for different decisions.

Keep a simple view of the important overlaps: situation, possible messages, agreed priority and owner. Update it when adding a journey. Avoid a network of exceptions nobody can explain. If many rules are needed to resolve the same collision, review the conversation’s design.

Validate the complete journey before expanding it

Write each email around a specific task. Check that subject, content and destination fulfil the same promise. If the message mentions a sizing guide, the link should help find that guide. If it presents a dynamic product, review what happens when data is missing or the item becomes unavailable.

Test the sequence of decisions too. An attractive preview does not confirm the correct person enters, data arrives on time or a purchase stops the reminder. Use test histories and record expected results before checking configuration.

The review should leave enough evidence for someone else to understand what was tested and what remains pending. It does not need a lengthy report for every change; it needs a clear record of the case, result and decision to publish or correct. An email marketing audit can organise this work when existing flows have poorly documented rules.

Cases worth including

• A valid entry and an entry that should be excluded, each with a distinct, verifiable reason.
• A purchase or status change during a delay between messages.
• Missing product data, an incorrect link or a destination that is no longer available.
• A contact overlapping with another journey and a scheduled campaign.
• A preference change or exclusion that should affect subsequent sends.
• Re-entry and duplicate events, checking the outcome the business expects.
• The end of the journey, with an explicit decision about which communication may follow.

Activate with monitoring and keep a maintenance plan

Before activation, review how many contacts could enter and who will monitor the first results. A valid configuration is not a consequence-free publication: a broad segment or historical data may change the expected scope depending on the mechanism used. Check the specific behaviour before introducing older audiences.

If the sender or domain has little previous activity, include a deliverability review in the plan. Sending growth should respond to the channel’s condition and observed signals. Copying a warming calendar without examining that situation does not replace a decision about volume and audience.

After publication, separate data problems from content problems. Few entries may mean the situation is uncommon, the integration is failing or a condition excludes too many people. Many sends without responses may require a review of proposition, selection or frequency. A different subject line cannot fix an event that never arrives.

Use email marketing metrics tied to each journey’s objective and comparable periods. Attributed revenue describes what the platform records, but does not by itself prove additional sales. Review unsubscribes, issues, incentive costs and the tasks the messages were intended to solve too.

Plan a review when products, shipping policies, promotions or integrations change. Keep owners, the last check date and a clear way to pause the journey if an inconsistency appears. Automation reduces repetitive work; keeping the conversation useful still requires editorial judgement and follow-up.

Sources and further reading

• Klaviyo · Getting started with flows
• Klaviyo · Email deliverability best practices