David Raab Article Archive

The goal of customer relationship management is to take the most appropriate action in a given situation. Often this involves selecting the best marketing message. But such a selection involves a prior choice of whether to deliver a marketing message at all. The choice is explicit in outbound situations such as email or telemarketing, where there is a decision whether or not to contact a customer at all. But a similar decision is made during inbound activities such as a customer service call or even a Web site visit, where there is a choice whether to try to sell something during the course of the interaction. In both outbound and inbound situations, companies may wish to limit the number of marketing messages they send to their customers in order to preserve good will.

The easiest way to implement contact limits is by building them into decision rule inclusion or exclusion conditions. This approach is available in nearly any customer management system, so long as the contact limit itself can be expressed in a standard query and the past contact history is available. However, it is problematic in systems where queries cannot be shared, since the contact limit would have to be recreated each time a new decision rule is set up. This yields considerable danger that different users would implement the limit in different ways and gives no simple way to check for consistency. It also means that any global change requires modifying each existing rule separately to conform to the new policy.

The approach is much more practical when queries can be shared across multiple rules, or rules can be shared across multiple campaigns. This means the contact limit can be defined once and then deployed consistently. Changes can also be made in one place only. Of course, users must still include the necessary rule or query. This is something they have a certain motivation to forget, since it limits the audience for whatever rule or campaign they are creating–and the excluded customers, being the most active, are probably their best prospects. This temptation can be removed in systems that let administrators specify mandatory exclusion rules, a function designed primarily to enforce policies such as honoring do-not-call requests and excluding poor credit risks.

Some systems take a more direct approach, allowing users to define contact limits as independent business rules that are automatically applied to all contacts. This is the functional equivalent of making them part of a mandatory, shared exclusion condition. However, it also may make it easier to include sophisticated contact conditions that are not easily built into a standard rule or query. Specific functions that may be part of contact limits include:

-minimum contact interval. This is a minimum time between marketing contacts, such as three days. This sort of limit is most applicable to outbound media such as email, where each message may contribute to customer fatigue.

-maximum contacts per period. This specifies how many contacts a customer can receive in a fixed period of time, such as three contacts per month. Its function is similar to the minimum contact interval.

-limits by medium (or category). Marketers typically want to specify different contact limits for different media, such as mail, email and telephone. This accounts for differences in cost, effectiveness and customer fatigue. There may be multiple levels of such limits: for example, an overall limit for contacts of all types, plus specific limits for individual media. Marketers may also wish to set limits based on categories other than medium: for example, they may wish to ensure that offers are not concentrated in a single category.

-priorities by message type. Marketers often want to ensure that some messages are sent regardless of normal contact limits. Examples include reactivation efforts to newly closed accounts, contacts to resolve customer service problems, and steps in a multi-step contact sequence. This can be achieved by providing a function to exempt certain messages from normal contact limits. A more general solution is to define different message types and set limits for each type (similar to limits by medium). High priority messages would be assigned to a category with no limit.

-priorities by campaign. Usually contact limits are applied one message at a time, by judging whether the relevant contact limits have already been met. But sometimes a customer is simultaneously eligible for multiple messages–say, in a system that accumulates messages to send to a call center, field force or direct mail operation on a daily or weekly basis. In this case, the system may need to determine which among several messages is actually chosen for delivery. Some systems do this on a first-come, first-served basis; others base the selection on priorities assigned to rules or marketing campaigns. Usually such priorities are fixed at the rule or campaign level, although a better solution would be to calculate the value of each contact for the individual customer and choose on that basis.

-contingent contact limits. Certain types of contacts may trigger special contact limits: for example, a severe customer service problem may activate a rule that blocks all regular promotional messages until it is resolved. This sort of approach also requires that messages be classified as different types.

-limits by household. Contact limits are usually set for individual customers. However, some marketers choose to set limits at the household level. These may be in addition to individual limits, or instead of them. Such limits are more relevant to categories where the sender does not directly control which household member processes the initial contact, like direct mail and telephone, than to categories where the message goes directly to its intended recipient, such as email.

-customer budgets. Some systems permit marketers to attach a cost to each marketing contact, to specify a budget for each customer, and limit the customer’s contacts to this amount. The costs may reflect actual production expenses and be used primarily to limit the actual investment in marketing to individuals. Alternately, they may reflect a perceived cost in terms of customer fatigue or good will. The assigned cost may actually vary based on the perceived value of a given contact to a particular customer, and could even be adjusted after the fact based on how the customer responded. The budget itself is usually determined by the customer’s value or activity level, although it can also be modified based on factors such as segment, life stage, or marketing strategy. A customer may be assigned a fixed budget that expires at the end of a fixed period such as a month, or given a cumulative budget that increases based on activities and decreases as promotions are sent. Although customer budgets have the potential to be a very powerful marketing management tool, most systems offer few of these nuances.

-value threshold. Marketers may wish to exclude contacts that do not provide a minimum expected value. The value calculation may reflect the expected financial value to the company, the value perceived by the customer, or some mix of the two. Considerations involved in implementing such a policy are generally similar to those involved in customer budgets.

-future contacts. Most systems base their contact limits solely on the contacts already received by a customer. However, some systems let marketers preselect customers for future contacts and include these selections in contact limits. This approach can be used to give priority to certain contacts, such as later steps in a multi-step campaign. It can also be used in conjunction with a value calculation to ensure that customers are sent the highest value offers for which they are eligible in a given period, rather than being selected on a first-come, first-served basis. Preselection also permits marketers to reschedule high-value contacts so they are spaced far enough apart to meet minimum interval requirements.

-deferred contacts. Some systems are able to defer contacts that are not made when originally scheduled because of contact limits. For example, a contact that was not delivered in January because the limit of three messages per month had already been exceeded might be delivered in February. As with other types of future contacts, the system needs a mechanism to ensure the contacts delivered are actually those with the highest value, not simply those that were scheduled first.

* * *

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.