A message can use accurate customer data and still arrive at the wrong moment. It can reflect a previous purchase while ignoring an unresolved problem. It can be carefully segmented while leaving the recipient unsure why the organisation knows so much. For vendors, enterprise personalisation needs a more useful promise than increasingly precise targeting.
The July and September US marketing roundtable summaries describe disconnected systems, limited customer views, questions about customer data platforms and the boundary between helpful communication and intrusion. They also describe practical progress, including more relevant segmentation and efforts to connect customer journeys. The evidence presents a range of experiences, rather than a uniform level of maturity.
The Leadership Board’s interpretation is that a credible personalisation proposition should begin with the customer’s task and the organisation’s ability to act appropriately. Vendors need to explain what information is required, how the buyer can use it within its own rules and how the resulting experience improves. More data and more variations are insufficient descriptions of that outcome.
Start with the customer purpose
July’s hyper-personalisation discussion explicitly challenged starting with data rather than the customer’s goals. It also noted that organisational and change management issues could limit progress even when AI created new possibilities. That is a useful warning for vendors whose discovery process begins with a list of available integrations.
Ask what the customer is trying to accomplish at the selected point in the relationship. The source discusses situations around post-purchase service, membership value and sensitive customer needs. Each implies a different purpose for communication. The buyer should be able to explain what a successful interaction would help that person do.
Only then should the proposal identify the information required. This is a recommended design approach, not a claim that all roundtable organisations followed it. It gives the buyer a way to distinguish necessary context from data that is simply available. It also narrows the scope of integration and review work.
A vendor demonstration should make the purpose visible. Show the customer’s situation, the information used and the action selected. Explain why that action is appropriate and how the workflow behaves when the information is incomplete. That gives the buyer something more concrete to assess than a set of messages with different names and offers.
A unified profile needs a usable operating context
July’s hyper-personalisation summary describes an organisation weighing whether to buy or build a customer data platform, or CDP, to address disconnected systems. July’s retention discussion illustrates the consequence of those gaps: marketing could send promotions without seeing existing customer problems. September’s journey discussion adds continuing difficulty coordinating data and messaging across teams.
These accounts give CDP vendors a clear discovery opportunity. Establish which gaps prevent the buyer from taking a useful action. The proposal should explain how the relevant records will connect and which team will use the resulting information. A broad ambition for a complete customer view needs to be translated into specific decisions.
The buyer should also understand what the profile will not resolve. Bringing information together does not decide who owns the next action or whether a message is appropriate. Those responsibilities need to be agreed. Vendors should distinguish the technical connection from the operating process built around it.
For an initial engagement, a bounded use case may be more informative than an attempt to consolidate every source at once. Demonstrate one meaningful connection and its effect on the customer journey. Record the dependencies and limitations. This is a recommended evaluation method derived from the source’s integration concerns, rather than an asserted standard implementation pattern.
Make the CDP build versus buy decision specific
The source records a build-versus-buy question but does not provide a conclusion that can be applied elsewhere. Vendors should resist turning that open question into an automatic argument for their preferred architecture. The buyer needs to compare what each option would require in its own environment.
A useful proposal should identify the customer problem, the systems involved and the responsibilities that would remain internal. Explain how the proposed solution would be maintained and how changes to the journey or data sources would be handled. The buyer should be able to examine the ongoing work as well as the initial implementation.
If existing capabilities can support part of the use case, make that visible. September’s retention discussion includes an organisation with technology already in place but still working on coordinated messaging and processes. That is a reminder to assess the operating gap before assuming an additional platform is the immediate answer.
The commercial case should relate the proposed investment to a usable improvement. A connected profile may support better service coordination or more relevant communication, but the proposal must explain the route between them. Avoid counting the presence of a platform as proof that personalisation has become more effective.
Treat preferences as part of the experience
September’s privacy discussion raised questions about preferences across channels and conflicts between different sources of preference data. It included an account of email opt-outs that did not translate directly to paid channels. The hyper-personalisation discussion also described difficulty encouraging customers to actively set their communication preferences.
For vendors, these are practical design issues to investigate. A preference centre needs to be understandable to the customer and usable by the organisation. Ask what choices the person is being offered, where those choices are recorded and how the relevant teams receive them. Do not assume that one channel’s settings explain the whole relationship.
A recommended demonstration would follow a preference change through the proposed workflow. Show which system records it, what action follows and how an exception becomes visible. If a connection is unavailable or requires customer development work, state that clearly. The buyer needs to assess the experience that can actually be delivered.
The source discussions raise regulatory concerns, but they are not a substitute for a legal assessment. The vendor’s role in this article is to make functionality and dependencies clear so the buyer’s privacy and legal teams can evaluate the design. Avoid promising that a feature, licence or default setting alone makes an implementation compliant.
Know when personalisation should stop
The July and September summaries both discuss the boundary between helpful personalisation and intrusive communication. July includes an account of limiting messages and relying on direct human relationships in a particular service model. September adds concern about accurate, contextually relevant data and excessive reliance on AI.
The vendor implication is to include restraint within the proposed capability. Ask when the organisation should avoid a message, use a less specific version or involve a person. The source does not prescribe universal thresholds, so these decisions should be based on the buyer’s audience, use case and approved rules.
This can be demonstrated through exceptions. What happens if the customer context is uncertain? How does the system respond when a service problem is active? Can an authorised user override an automated recommendation? The answers help the buyer evaluate control, rather than only the sophistication of targeting.
Restraint also needs an owner. Someone must decide which situations require different treatment and review whether the rules remain appropriate. A vendor should make that responsibility visible in the implementation plan. Otherwise, the organisation may acquire a capability for detailed targeting without the process needed to use it judiciously.
Use accurate data with an honest account of its limits
September’s enterprise transformation discussion emphasised data consistency, completeness and freshness. It also described the need to map information carefully and remove unnecessary data. Those operational details belong in a personalisation proposal because they affect whether a customer is being understood correctly.
Vendors should explain how the proposed use case handles missing or uncertain information. A profile can contain many fields without all of them being suitable for the intended decision. The buyer should know which sources are being used, how conflicts are surfaced and what review is needed. These are recommended questions for evaluating the offer.
The same care applies to inferred interests. An observed action may suggest a useful question without settling the customer’s intention. The source’s emphasis on context and the limits of intrusive targeting supports a cautious approach. Show how the organisation can investigate or offer a helpful next step without treating an inference as a confirmed personal fact.
This is also an opportunity to make the integration scope more credible. Identify which data is necessary now and which could be considered later. A narrower, well-understood use case can provide a clearer evaluation than a demonstration built on a fictional complete record that the real organisation cannot reproduce.
Personalise across the hand-offs that customers experience
September’s customer journey discussion described attempts to coordinate marketing, sales and support, alongside continuing gaps between their systems. July’s retention discussion showed how those gaps could produce an inappropriate promotion. The combined evidence points to the value of continuity across a relationship.
A vendor should map the selected journey from the customer’s perspective. What has the person already told the organisation? Which team responds next? What information can that team legitimately access? Where could the customer be asked to repeat the same explanation? These questions are a recommended way to locate friction.
The proposal should then show how the solution helps with a specific hand-off. It might make approved service context available to a marketing workflow or help a customer-facing team understand previous interactions. The exact design must follow the buyer’s permissions and responsibilities. The source does not support unrestricted sharing across every function.
Consider the human role within that continuity. September’s enterprise transformation discussion includes using AI to support customer service while retaining human interaction. This provides a useful alternative to assuming that personalisation requires full automation. A tool may create value by helping a person respond with better context.
Match the approach to the data the buyer actually has
September’s hyper-personalisation session contrasted experience with detailed consumer personalisation and the difficulty of achieving similar depth in B2B settings. That distinction matters when a vendor demonstrates a capability using rich example data. The prospect may not possess comparable information or permission to use it.
Discovery should establish the available evidence before promising individual-level experiences. The source includes both more basic segmentation and ambitions for finer personalisation. A vendor can help the buyer decide what is useful at its current stage and what additional work would be required to progress.
Avoid making the most granular option the default definition of success. A relevant message to a well-understood group may address the buyer’s problem more credibly than a highly specific message based on uncertain inference. That is an editorial recommendation from the source’s concerns about context and intrusion, not a comparative performance claim.
The evaluation should use representative information from the intended environment, subject to the buyer’s controls. Explain which parts of a demonstration depend on idealised data. This gives the sponsor a realistic understanding of readiness and reduces the risk that implementation begins with an undisclosed data collection project.
Measure the experience as well as the response
The September summaries include reported improvements in communication responses and interest in understanding customer experience. July links personalisation with post-purchase service and demonstrating membership value. These accounts suggest that evaluation should be connected to the purpose of the interaction, rather than a single campaign measure applied everywhere.
A recommended pilot should define what the customer should be able to do more easily and what the organisation hopes to learn. Observe the relevant response alongside the work required to deliver it. Include signs that the contact is inappropriate or unwanted, using the evidence the buyer is authorised and able to collect.
Do not treat an engagement increase as proof that every aspect of trust has improved. The source does not establish such a relationship. Explain what was observed and what remains uncertain. If the interaction is meant to support retention, the buyer may need a different review period from one used to examine an immediate response.
The business case should also include the effort of maintaining the experience. Data connections, rules and content can require continued ownership. Make those responsibilities part of the proposed operating model so the buyer can judge whether the capability is sustainable within its team.
Sell a useful interaction the buyer can deliver
The strongest response to these roundtable themes is a specific proposal. Identify a customer need, the minimum useful context and the action the organisation can responsibly take. Explain the data connections and internal decisions required, then agree how the buyer will evaluate the result.
The July and September discussions show interest in richer personalisation alongside unresolved operational constraints. Vendors should address those constraints directly. A credible proposition helps the enterprise make communication more relevant while retaining control over when, how and whether it happens.
Related reading
Across The Leadership Board: The ownership gap quietly driving enterprise data security investment (Data).
Speak to The Leadership Board
Selling CDPs, personalisation platforms or customer journey services? Speak to The Leadership Board about the data, trust and implementation questions shaping enterprise marketing conversations. Use Marketing Buyer Intelligence to build a proposition around the customer experience your buyer can realistically deliver.