By the end of this lesson, participants will be able to:
- Understand what Next Best Action means in an editorial or audience context
- Recognize how activation evolves from fixed rules to more contextual decisions
- Comprehend the role of advanced personalization in newsletters, paywalls, recommendations, and retention messages
- Identify opportunities and limits of more anticipatory and automated activation
From Structured Activation to Contextual Decision
In module 3, the focus was on moving from behavioral segmentation to more structured activation supported by CRM and CDP. At this level, the additional leap is for the organization to stop thinking only in terms of fixed segments and campaigns, and move toward a more contextual logic: which action makes the most sense for this user, at this moment, on this channel, and with this goal?
This is the domain of Next Best Action.
What is Next Best Action
Next Best Action does not mean that a system knows everything or that it can automate every decision without supervision. It means that, based on signals, context, and rules or models, it attempts to identify the most reasonable action to advance the relationship with a user.
In a media context, that action could be:
- Recommending a specific piece of content
- Inviting to register
- Offering a thematic newsletter
- Modifying the paywall message
- Launching a retention action
- Avoiding showing a proposal if the timing is not right
From General Campaigns to Finer Personalization
With greater maturity, the organization stops working solely with generic activations and starts to better combine:
- Thematic affinity
- Stage of the journey
- Consumption intensity
- Preferred channel
- Conversion or abandonment probability
- Response history to previous actions
Personalization goes from “showing similar content” to a more sophisticated way of deciding what experience, message, or offer fits best.
Use Cases
At this level, use cases may appear such as:
- Paywalls with differentiated logic based on behavior and propensity
- Content recommendations aimed at habit rather than immediate clicks
- Newsletters or messages depending on the stage of the journey
- Win-back actions based on early signals of abandonment
- Suppression of messages when the system detects low relevance
What distinguishes these use cases from simpler activation strategies is the role of context. A basic campaign sends the same message to a defined segment at a scheduled time. A Next Best Action system determines whether to send anything at all — and if so, what, when, and through which channel — based on signals specific to that user at that moment. The suppression case is perhaps the most counterintuitive: sometimes the data suggests that the right action is no action, because a user is in a fragile moment, has already been contacted recently, or is showing signals of conversion readiness that would be disrupted by an unsolicited prompt.
Limits and Responsibility
The finer the activation, the more important it is to maintain criteria regarding:
- Consent and privacy
- Explainability of certain automated decisions
- Risk of overwhelming or manipulating the user
- Need for human supervision and ongoing evaluation
The risk of user overload deserves particular attention. A system technically capable of sending personalised messages across ten touchpoints a day is not necessarily a system that should do so. Users who feel tracked or over-contacted typically disengage faster than users who receive no personalisation at all. The threshold between relevant and intrusive is not defined by the system — it is defined by user behaviour, and monitoring that boundary is a human responsibility, not an automated one. Any organisation operating at this level of personalisation should have explicit criteria for when automation is paused and human review is required.
An Insights-Driven organization not only automates better but also knows when not to automate or when to introduce brakes, exceptions, and reviews.
Try it yourself
Your personalisation system has flagged five users for potential action:
User A — Subscriber for 14 months · Visit frequency dropped from daily to twice weekly over the last 3 weeks · Last newsletter opened 18 days ago · Billing renewal in 9 days
User B — Anonymous · 11 visits in the last 7 days · Reads exclusively Climate and Science · Has never seen a registration prompt · Consistently above 5 minutes per session
User C — Registered, not subscribed · Received 4 personalised emails in the last 10 days · Opened none · Visited twice this week via direct URL · Hit the paywall twice today
User D — New subscriber, day 3 · 12 articles read · Strong affinity for Investigations and Long reads · Has not engaged with any newsletter yet
User E — Subscriber for 6 months · Visits daily · Has opened and clicked every newsletter for 3 months · Yesterday clicked “cancel subscription” but did not complete the cancellation
Propose a Next Best Action for each user. In at least one case, the right answer is to do nothing — or to actively suppress a message the system might otherwise send.
Consider:
- What is the NBA for each user — and which single signal most strongly justifies your choice?
- For which user would sending an automated message right now be counterproductive or potentially harmful — and why?
- User C has been contacted 4 times in 10 days with no response. At what point does frequency become friction — and what governance rule would you put in place to prevent it happening automatically?
- User E has shown a cancellation intent signal. What is the NBA — and what would you need to know before automating any response to that signal?
The most sophisticated personalisation systems know when to act. The best ones also know when not to.