Make the campaign understandable to its next owner
A future editor should be able to explain who receives a campaign, what changes its content and when it stops sending. If those answers require reconstructing the original builder’s memory, every revision becomes more expensive.
Governed Personalization treats that maintenance problem as part of communication design. David draws together five conference sessions into a practical standard: personalization should run through documented, reusable and testable rules. The message is the visible result of decisions about data, timing, channels, content and review.
Choose one campaign family to examine. An application-completion reminder or event follow-up gives the team a concrete audience and next step. Write those down before adding variants. The purpose provides a way to decide which differences deserve attention and which only add branches.
In the recording · 06:05 ↗Give timing a definition that survives interruptions
The episode highlights population timestamp filters as a way to make campaign timing more tolerant of operating conditions. A rigid day count can become awkward when a record enters late, leaves and returns, or misses an expected processing window.
Document the event that starts the clock, the window in which a message is useful and the condition that ends eligibility. Then test interruptions deliberately. A reminder should not become appropriate again simply because someone re-entered a population after completing the requested action.
The transferable lesson is the timing review, rather than a universal cadence. Check the local implementation with sample records, including a missed run and a completed action. Make the resulting rule visible to the people who plan the communication calendar.
In the recording · 07:42 ↗Choose a structure the team can safely change
The technical division is clear: queries shape the data, while Liquid markup and content blocks present it. Well-named exports let reviewers inspect the inputs before asking why a paragraph appeared. Reusable blocks keep approved language in a manageable place.
Consolidation still needs judgment. One dynamic message can work when the audience shares a purpose, schedule, owner and structure, with only a few sections changing. Several messages may be easier to maintain when those responsibilities or requirements differ.
Count the combinations the team must test, not just the number of templates. Replacing many straightforward messages with one deeply nested mailing can move the complexity without reducing it. Choose the version another staff member can explain and review.
In the recording · 12:07 ↗Decide what each source is allowed to influence
A student’s explicit academic-interest selection can support relevant academic information. A vague referral or uncertain list classification supports a much weaker conclusion. Keep the source, its age and its meaning attached to the decision about how it may be used.
The episode also separates parent behavior from student intent. Activity by one person should not quietly become a claim about another person’s preferences. Where evidence is uncertain, a general invitation to identify interests may be more useful than a confident personalized statement.
A simple review question helps: could a staff member explain the outreach to its recipient in an ordinary sentence? If the explanation depends on an inference the recipient would not recognize, reconsider the input or the proposed action before polishing the copy.
In the recording · 15:35 ↗Map the whole communication family
The Communications Governance Map brings the decisions together. Record eligibility, purpose, channels, personalization inputs, reused assets, suppressions, fallback behavior, test cases, success measures and maintenance ownership. Include a review date so the map remains part of the work.
Channel coordination belongs here too. An email, portal and staff conversation can share useful context while doing different jobs. Decide what each channel contributes and how a completed action changes the next touchpoint. Otherwise several individually reasonable messages can arrive as repetitive or contradictory instructions.
Give blank, stale and conflicting data their own branches. Test a completed task, an opted-out channel and a record whose information changed after planning. A fallback should preserve accuracy, and a suppression should remain visible to the next editor.
In the recording · 17:45 ↗Keep review and measurement tied to the purpose
Review should reflect the consequences of the message. Routine operational reminders and sensitive or policy-related communications require different levels of care. AI can help refine approved material or propose alternatives, but institutional claims still need verified sources, and generated code still needs testing.
Choose the outcome before launch. For an application-completion campaign, review completion. For an event campaign, examine registration and attendance. Opens and clicks can help diagnose a problem, but they do not settle whether the campaign helped people take its intended next step.
At the next review, consider both performance and maintenance. Did the campaign support its purpose? Can the team still explain its rules? Revise or retire an asset when its audience, assumptions or purpose no longer holds. That makes personalization a shared working tool rather than a collection of exceptions.
In the recording · 20:20 ↗