Home › Micro Degree Best Practices › Email Marketing and Marketing Automation
Best Practice ManualEmail Marketing and Marketing Automation
A practical reference for planning, building, testing and measuring email campaigns and automated customer journeys in everyday marketing work.
01Purpose and scope
- This manual supports the course Email Marketing and Marketing Automation. It describes how working marketers plan, build, launch and maintain email programs, from one-off newsletters to multi-step automated journeys that respond to what a subscriber does. It is meant as a reference for the routine decisions that come up on a real team, not as a theory text.
- The scope covers broadcast campaigns, triggered messages, welcome and onboarding series, abandoned cart or browse sequences, re-engagement flows and post-purchase messages. It also touches the data and tooling that make those messages possible, because most email problems are really list, data or timing problems in disguise. Transactional email such as receipts and password resets is mentioned only where it shares infrastructure with marketing mail.
- Practice changes a good deal by context. A subscription software company with a trial period, a retailer with seasonal peaks, and a nonprofit that sends appeals all use the same tools differently, and the limits on what they can safely send also differ. Treat this manual as a starting structure that you adjust to your own audience, volume, and legal exposure.
02Core principles
- Email is a permission-based channel. Every message should go to people who agreed to receive that kind of message, and the agreement should be easy to withdraw. Permission is the foundation for deliverability, for compliance, and for the trust that makes later messages get opened.
- Relevance beats volume. A message sent to a narrow group with a clear reason to care usually does more than a broad send with generic copy. Segments should reflect something the business actually knows, such as product interest, purchase history, or the stage of a free trial, rather than guesses from thin data.
- Each automation should have one job and a defined end. A welcome series exists to introduce the brand and drive a first action, and it should exit subscribers who take that action. Flows without exit conditions keep sending to people who have already converted, which annoys them and makes reporting misleading.
- Measure what the message was for. Open rate is affected by privacy features in some mail apps and is weak evidence on its own. Click-through, conversion, revenue per recipient, unsubscribe and complaint rates, and downstream retention tell you more about whether a message did its job.
- Sending frequency is a decision, not a default. Each additional message costs attention and raises the chance of unsubscribes or spam complaints. Set frequency caps across flows so that one subscriber does not receive three unrelated messages on the same day.
03Planning and preparation
- Start by naming the business goal and the audience for each program. A program might aim to convert trial users to paid plans, to bring lapsed buyers back, or to keep existing customers informed about a service change. Write the goal in one sentence and name the metric that will show whether it worked. Without this, the team tends to build a sequence first and argue about its purpose later.
- Audit the data before designing anything. Check how addresses were collected, whether consent was recorded with a timestamp and source, how the list is split by status, and how often fields such as first name, country or product type are blank or wrong. Decide which fields are required for each segment and which need cleanup. A segment built on a field that is empty for forty percent of people will quietly underperform.
- Map the stakeholders. Marketing owns the message strategy, but the data team usually owns the customer records and event feeds that trigger automation, legal or a privacy officer reviews consent language and retention rules, and customer support often hears about confusing messages first. Agree early on who approves new flows and who can pause a live one.
- Choose tools with the workflow in mind. Email service providers, marketing automation platforms and customer data platforms overlap in capability. Confirm that your choice can handle suppression lists, event-based triggers, branching logic, template versioning, sending domain authentication, and reporting by segment. Check the vendor documentation for limits on contact counts and sending rates, since those can change and affect cost.
- Prepare a content calendar for broadcasts and a flow map for automations. The flow map should show each trigger, each branch, each wait period, each exit condition, and each message, with the owner of the copy and the approval status. A simple diagram on a shared page is usually enough, and it becomes the reference when someone asks why a subscriber received a message.
04Recommended workflow
-
1
: Define the program brief. Record the goal, audience, offer, success metric, required approvals, and launch constraints such as a product release date or a sales promotion that must not overlap with the send. Keep the brief short enough that a reviewer can read it in five minutes.
-
2
: Build or confirm the audience. Create the segment using stored fields and recorded events, not the email content. Before saving it, preview the count and sample a few records to confirm the logic matches the intent. Apply suppression lists for unsubscribes, bounces, recent complaints, and people already in a conflicting flow.
-
3
: Draft the message. Write a subject line and preheader that say what the message contains, avoid misleading urgency, and keep the primary call to action visible without scrolling on a phone. Use a single primary action per message where possible. Include the sender name, a valid physical postal address where required, and a clear unsubscribe link.
-
4
: Set up triggers and branches for automated flows. Define the event that enters a subscriber, such as a signup or a cart abandonment, and check that the event is captured reliably. Add wait steps that reflect realistic behavior, not arbitrary intervals. Add exit conditions for conversion, unsubscribe, and any event that makes the message irrelevant.
-
5
: Test before launch. Send test messages to internal accounts across major mail clients and screen sizes. Check personalization fallbacks for blank fields, links with tracking parameters, images that have alt text, and rendering in dark mode. Run a test enrollment with a small set of sample profiles and confirm each branch fires as designed.
-
6
: Review and approve. The approver should check the brief against the content, confirm the legal footer and consent basis, verify that the segment and suppression logic are correct, and confirm the send time does not collide with other programs. Record the approval with a name and the version reviewed.
-
7
: Launch and monitor. Watch the first hours of a broadcast for delivery problems, a spike in complaints, or unexpected bounce patterns. For flows, check enrollment volume and early step performance against the forecast. Be ready to pause a send or a flow quickly, and know who has permission to do so.
-
8
: Analyze and adjust. Compare results against the success metric in the brief, segment the results by source and cohort, and note what you would change next time. Feed the findings into the next brief rather than storing them in a slide that no one revisits.
05Quality checks, risk controls and compliance
- Authenticate your sending domain. Set up SPF, DKIM and DMARC records as your email provider instructs so that receiving servers can verify your messages. Major mailbox providers have published sender requirements that change over time, so check the current guidance from each provider and from your email service before relying on it. Monitor your domain reputation and keep complaint rates low, since a poor reputation can send even good mail to spam folders.
- Honor opt-outs quickly and completely. Unsubscribe requests should be processed across all lists and flows, not just the one the link came from. Many jurisdictions set deadlines for processing opt-outs. In the United States, the CAN-SPAM Act requires commercial email to include a working opt-out mechanism and a valid postal address, and it sets a processing window for opt-out requests. Other rules are stricter. The EU and UK data protection regimes generally require a lawful basis for processing personal data and, for many marketing messages, consent that is specific and freely given. Canada and several other countries have their own consent rules. Where rules depend on jurisdiction, confirm the current requirements with your legal team and the authoritative regulator, since requirements change.
- Keep consent records. Store the source, date, and wording of each opt-in, and keep suppression records even after a person is deleted, so that you do not accidentally re-add them through an import. Retention periods for personal data should follow your privacy notice and applicable law.
- Check the content for accuracy. Every price, deadline, eligibility rule and product claim in an email should match the source of truth. Automated flows are especially risky because they can send old offers long after conditions have changed. Build a rule that flows referencing time-limited offers are reviewed whenever the offer changes.
- Control for risk in automation. Set maximum enrollment rates, global frequency caps, and alerts for unusual volume. A broken trigger that enrolls an entire database into a flow can cause a large number of complaints within hours. Keep a documented rollback procedure.
- This guidance does not replace qualified legal, privacy or professional judgment. Where your organization operates in regulated sectors such as financial services, health care or education, the rules on promotional messages can be much stricter, and you should obtain approval from your compliance function before sending.
06Collaboration and documentation
- Email programs involve many people who each see a different part. Copywriters, designers, data analysts, developers and legal reviewers need the same flow map and the same brief. Use a single shared location for briefs, flow diagrams, approved copy, and launch records so that people are not working from old attachments.
- Name owners for each live program. Every automated flow should have a person responsible for its performance and its content, and a backup who can pause it. Review active flows on a regular schedule to retire ones that no longer match the business, since abandoned flows often keep running for years.
- Document the naming and tagging conventions. Consistent names for segments, tags, and events make it possible to audit who is in what, and they prevent duplicate segments that drift apart. A short naming guide with examples is worth more than a long policy nobody reads.
- Coordinate with sales, support and product. If a flow encourages a trial user to book a demo, sales should know. If a service change is announced by email, support should have the answer ready before the message lands. Share the calendar across teams so that a promotional send does not arrive the same morning as a billing notice.
- Keep a change log for templates and flows. When a message is edited or a branch is added, record what changed, why, and who approved it. This makes it possible to explain a performance shift and to roll back cleanly.
07Measures and continuous improvement
- Set a small set of measures for each program. For broadcasts, track delivery rate, unsubscribe rate, complaint rate, click-through rate and conversions attributed to the send. For flows, track enrollment volume, step-by-step progression, exit reasons, and revenue or retention compared with a holdout group. A holdout group, a set of eligible subscribers who do not receive the flow, is the most reliable way to estimate the flow's effect, though it costs some potential revenue.
- Interpret open rates with care. Apple Mail Privacy Protection and similar features can load tracking images automatically, which inflates open counts. Use opens as a rough trend indicator, and rely on clicks and conversions for judgments about content.
- Use A/B tests with discipline. Change one variable at a time, such as a subject line or the first sentence of the body, and decide the sample size and the result threshold before the test starts. Small lists rarely produce reliable answers, so for those, use judgment and qualitative feedback instead of forcing a winner.
- Review deliverability on a schedule. Watch bounce categories, spam complaint trends, and inbox placement signals if your tools provide them. A rise in hard bounces often points to list hygiene problems, while a rise in complaints usually points to content, frequency, or list sourcing.
- Improve through short cycles. Make one or two changes per program per cycle, measure them, and record what happened. Over time the team builds a shared record of what works for its own audience, which is more useful than general advice from any single source.
08Common mistakes and how to prevent them
- Buying or renting lists. Purchased lists typically contain people who never agreed to hear from you, which produces complaints, bounces and reputation damage. The short-term reach is rarely worth the long-term cost, and in many jurisdictions it is not lawful for marketing purposes. Grow the list through clear signup forms and value exchange instead.
- Sending to everyone. Blasting the full list with every message trains subscribers to ignore you. Use engagement-based segments, and consider a sunset policy that moves long-inactive contacts into a re-engagement flow and then suppresses them if they do not respond.
- Automations without exit conditions. A flow that keeps sending a trial reminder after the person has paid is a common and avoidable error. Add exit events for conversion and for unsubscribes, and test them with real sample profiles.
- Ignoring mobile rendering. Many subscribers read on phones. Wide tables, small buttons, and images with text baked in fail on small screens. Test on real devices and keep the main call to action near the top.
- Untracked links and unclear attribution. Without consistent UTM parameters and a clear attribution window, revenue claims are hard to defend. Agree on a convention before the first send and apply it to every link.
- Letting old flows run. Offers expire, products change, and staff leave. Schedule reviews of active flows and assign an owner to each one so that outdated messages are caught before subscribers see them.
09Practical checklist
- Confirm the program goal, audience, success metric and approver in a one-page brief before building anything.
- Verify consent source and date for every segment, and check that suppression lists are applied.
- Authenticate the sending domain and confirm the current guidance from your email provider and major mailbox providers.
- Include a working unsubscribe link, a valid postal address where required, and an accurate sender identity in every message.
- Test rendering on phones and desktop clients, and check fallback text for blank personalization fields.
- Run a test enrollment for each automated flow and confirm entry triggers, wait steps, branches and exit conditions.
- Confirm that every price, deadline and offer in the content matches the current source of truth.
- Set frequency caps and enrollment limits, and document who can pause a live program.
- Record approvals, template versions and flow changes in the shared log.
- Compare results to the success metric within an agreed review period, and capture changes for the next brief.
- Review active flows on a regular schedule and retire those that no longer match the business.
★Learning sources
- Micro Degree: Email Marketing and Marketing Automation
- Micro Degree: Marketing Strategy Development
- Micro Degree: Programmatic Advertising Management
- Micro Degree: Lead Generation Operations
?Frequently asked questions
How many emails should a new subscriber receive in the first week?
There is no universal number, and the right answer depends on how much the subscriber signed up for and how much value each message delivers. A welcome series often runs three to five messages over the first couple of weeks, with a frequency cap so that broadcasts do not stack on top of it. Watch unsubscribe and complaint rates in the first cohort to judge whether the cadence is too heavy.
Can I email people who bought from me but never opted into marketing?
This depends on jurisdiction and on what the original purchase notice said. In the United States, the CAN-SPAM Act generally allows commercial messages to existing customers if they include an opt-out, but other countries apply stricter consent rules to marketing. Confirm the specific requirements with your legal or privacy team and the relevant regulator before sending.
Why did my open rate drop after a privacy change in mail apps?
Some mail apps now pre-load tracking images, which can inflate opens for some subscribers and make the metric unreliable as a trend. A drop in opens may reflect this change rather than a drop in interest. Compare clicks, conversions and unsubscribe rates, which are less affected, and segment by mail client when the data allows.
Should I use a holdout group for every automated flow?
A holdout group is the most reliable way to estimate what a flow contributes, but it costs potential revenue from the people who do not receive the message. Many teams use holdouts for larger or higher-revenue flows and rely on simpler before-and-after comparisons for small ones. The decision should be made with the revenue owner before the flow launches.
This manual explains practice in general terms. It does not replace qualified professional judgment, approved procedures or required authorization.
Marketing Best Practices
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain