One objective of customer relationship management is to continually refine an organization's insights into its customer base.
These insights drive how an organization communicates with customers, so that each contact is more intelligent and meaningful than the last.
Often, necessary data is not in place to draw meaningful conclusions from previous customer experiences. In order to drive towards information based, data driven business decisions, organizations must have a process and tools in place to complete all four components of campaign management.
Campaign Management is defined as the process of designing, executing, and measuring marketing campaigns through the use of applications that:
- Select and segment customers
- Track the contacts made with customers
- Measure the results of those contacts
- Model those results to more efficiently target customers in the future
The Common Mistake
The first step in the Campaign Management Process tends to be the primary focus for most marketing departments. Organizations spend a large percentage of their efforts identifying customers, defining the offer or message, applying suppressions, defining segments, and delivering the custom message, whether it is email, direct mail, or telemarketing.
Unfortunately, tracking a customer's reaction to a custom message is usually not given the same level of attention and planning as the execution component. This leaves marketing analysts and marketing departments salvaging bits and pieces of incomplete data information, from which they are asked to draw conclusions and define business insights.
Issues around tracking, sourcing and matching a customer's behavior must be evaluated and resolved during the planning phase of a campaign. In this article we will talk about clearly defining the customer-centric business questions that are to be answered, and some of the data sourcing issues that must be addressed in order to successfully learn from campaign efforts.
Avoiding the Pitfalls
As you consider the business objectives of campaign execution you have to ask yourself the following questions:
Define Your Objective: What am I trying to learn from this customer contact? What do I need to know to make the next communication more rewarding for the individual and the organization?
By prioritizing the questions that need answers, it is much easier to identify the data relevant for response measurement and analysis. Once you identify the data that does not reside in a source you currently have access to, such as through a corporate data warehouse or a marketing data mart, you are better prepared to approach other sources for assistance. This allows for clear, concise conversations with administrators of these source systems.
Remember these individuals know their data well, and share with them the business question you are trying to answer. They may contribute to your learning process by bringing their system-specific data knowledge to the discussion.
Identify Channel Data Dependencies: How are my customers going to respond to me? Which channels will they use?
Often the primary information captured at a channel level is very specific to a channel. For example, call centers capture the length of a call, or the activities that took place during a call. Log files on a web site are used to determine the time on a page or number of web page hits. These systems are often not designed with marketing efforts in mind, so there will be a gap between current data captured and your data needs. You must seriously weigh the effort and time required to fill this data gap by modify these systems against the importance of the business question being answered.
Be aware these systems you will be considering are often operational systems, which need to be running at all times. Typically, managers of these systems are very protective and hesitant to introduce change. Prepare to discuss the value of the data and whether you will be using it on an ongoing basis. Also, if this data is critical, begin these discussions far enough in advance so there is sufficient time to make necessary system modifications.
Consider Data Quality: How reliable is the data?
As with any analysis, data quality should always be evaluated. You don't want to expend a great deal of effort tracking and gathering data if it is incomplete. Leverage those individuals within your organization who have intimate knowledge of the source systems. They often can tell you if the data looks reasonable. One simple example of information that may seem telling, but may not be populated is a "Reason for Closing an Account" code on a call center system.
From a marketing perspective, it would be interesting to know why an individual closed an account. From a customer service perspective, however, a service representative may not ask the question, depending on the disposition of the caller, or in an effort to shorten the call time. Call center representatives may fill this field with a meaningless code during the call, which provides no insight during analysis.
Define Response Stages: Do I need to capture “degrees of response?” If so, what are the degrees?
Degrees of response are stages that a customer may pass through before displaying your preferred behavior. For example, for a credit card solicitation, response stages might include 1) calling the call center for additional information, 2) completing an application, 3) receiving credit approval, or 4) activating the line of credit. You must define these stages during the planning phase of your campaign, and determine if this level of detail is important.
This information might be useful in uncovering unknown customer contact issues. In the credit card example above, if you find that every customer who called for additional information never completed an application, this may be the result of insufficient call center training, or difficulty fulfilling one-off information requests. Keeping in mind your key business objectives, if you still believe this type of data is critical to your learnings, make sure to thoroughly define the “degrees” of interest and how you will track them.
Gain a Holistic View: Can I get a view of the customer's response behavior across all channels?
You must determine the importance of knowing how a customer reacts, and if they contact your organization across multiple channels. Perhaps they request more information through the call center, but place their order through the web site. From a marketing perspective, this may be important to you. However, in order to get an “all channel” view of the customer's behavior, there must be a means for identifying an individual across systems.
The disparate source systems that you are considering may or may not identify customers at an individual level. If they do, they may use very different means to identify an individual customer.
For example, the web site may use an individual's email address, while the call center or legacy systems might use a social security number. Very rarely will you find the email address captured in the call center data source. And due to security concerns, web sites rarely use social security numbers to identify an individual. If the web site and the call center do not use the same identifying information, greater data manipulation effort may be required to identify one person across systems, such as a match on name and address.
Again, this is extra effort, which may give you incomplete results, and should be evaluated during the planning phase. Consider tailoring your message to encourage consumers to use a single channel for contact and using that source as your primary source for response analysis.
Accumulate Data & Knowledge: Where do I want to store my data once I have completed my response analysis?
Once you have gathered and reviewed the captured data after a campaign, it is critical that the data is archived in a location for future use. You may find subsequent analysis is needed on the campaign, such as development of a predictive model or answering additional business questions.
Failure to archive the response data will, depending on the structure of the source systems, result in the need to re-pull and massage the necessary data fields, which can be time consuming, or worse, result in the loss of the key data you worked so hard to capture.
Bottom Line:
In order to conduct meaningful response measurement and analysis, it is vital that you have a thorough, well thought-out plan for capturing and storing customer behavior data.
The execution of campaigns, and all of the effort that goes into that process, is wholly dependant on response tracking in order to be of value to the organization. Depending on the number of channels an organization uses to communicate with customers, implementation of a holistic tracking plan can be daunting. Remember your objectives, and work in a phased approach to reach your goals.
