Why Companies Need Analytics? - Facets of Getting to Useful Analytics
Why Companies Need CRM Analytics?
An Overview of Facets of Getting to Useful Analytics.
Nethra Sambamoorthi, Ph.D
11 Bartram Road, Englishtown,
1. Introduction - Key Three Questions.
Any time data is collected, the first question which one should have asked even before the data is collected is, should we need to collect data on a certain initiative and if it leads to a resounding yes (typically it is yes, as we all agree that any thing that is not measured can not be managed), what data elements need to be collected (how frequently, from whom, and when) and equally important are the what metrics need to be looked into (specific details of how we are going to utilize these collected data points) that we are expecting the system to answer as we keep running the operation; a CRM operation. We took on the question mentioned in the title, as this seems to be one of the common questions with in the CRM community. Typically, there seems to be the feeling that we can talk to customers to find out the answers for the complex analytical problems that companies face in implementing the CRM solutions. We know that while it is possible to get directed and simple answers by talking to the consumers, often their behavior and actual transaction patterns might reveal totally different insights into the CRM process. (Remember ultimately what matters is the transaction – namely selling) Some times, it is important to know what the customer would need/like/prefer, even before they can articulate that to leverage the market opportunity. This and other issues discussed above gives rise to the needs of CRM analytics. In the following, I bring out the real time data exchange and associated reasoning for CRM analytics, and a combination of online/offline analytics for meta CRM activities, metrics which talk about various CRM activities, over and above CRM data exchange and its metrics for customer interactions.
2. Current Status:
Often we see companies end up collecting typically name and address and all the data that are part of the transaction, because of their needs for financial auditing purposes, but start thinking about analysis, not knowing what to get out of the data cavalierly slipping into the approach: “let us mine this data where we have hundreds of variables with transaction data”. The whole exercise is either too wide a hunt to start with and in case it is not a wide hunt the whole process of data collection does not yield meaningful calculations or interpretations. The valuable transaction data provides whatever the remaining power of the understanding of customer behavior, self prophesied by its recency, frequency, and monetary type analysis.
3. Value metrics should be designed into the CRM system, and should not be an after thought.
Like quality, valuable metrics about companies operations are designed into the process; in CRM into the process of marketing a product or a service. It cannot be an after thought. In other words, you cannot get valuable metrics after the (CRM) process is designed. Part of the reason the second generation of CRM applications are having difficulty in justifying their activities as they are typically designed to answer questions like the following with ease. Value metrics has to be defined to be part of the CRM design. Otherwise, we will end up redesigning CRM systems and data interconnections to execute excellence in CRM.
So the best practices for CRM analytics demands to fore see the following questions and hence all the relevant set of data elements associated with the answers of these questions.
4. What are the typical CRM related questions that management asks once the CRM operations start kicking?
Typically the following are the questions are asked during the execution phase of a CRM operation. Many of them seem to be simple common sense database practices and however, data quality and data collection process is always an after thought in 80% of the companies.
How do we make sure we exclude the consumers/customers who should not be sent as declared by business rules?
How to figure out the right offer (Analysis of right offer to the right person - Personalization of offers)
What do we do to make sure we can increase the response, as much as double, just by selecting the right timing to the members - calendar vs. file related timing in creating a compelling offer?
How to increase the value of the message/offer to the member? How to estimate the value of our messaging/offer to the consumer?
How to track your responses so that you can better target next time? Utilize your powerful member information.
How do we find new prospects with in the internal database and outside the database to increase total responses and the revenue?
- Who are those 20% of your current members who bring in 80% of your revenue.
- How do we specifically increase that base?
- How to create offer which are specifically designed to a member in terms of specific offer and at the same time to increase the total value of potential sales more than the previous visit's receipt value?
- How do we campaign the marketing activities to the members, for systematic month after month revenue stream to the company during 2003? - Optimizing marketing campaigns. A good laundry list of how many different types of campaigns we will be able to with the types of data collected.
- How do we evaluate the effectiveness of these communications or offers? - ROI, ROI, ROI
With out feeling remorse, get resounding yes to above questions that will be asked with any campaign that will be fielded and the availability of data to understand answers to those questions.
A resounding -yes should be shared with every level of the management from the analysts to the president of the company.
5. Who and When one needs CRM analytics?
Any company who are spending money in CRM systems, CRM software, call centers, interaction centers, and direct marketing activities such as sending offers/messages in US Mail/email, have to raise the above simple list of questions in section 4, within the management, to make sure that the CRM implementers with in the organization are keeping these as part of the standards to fulfill as we embark on the operation phase of a CRM initiative.
6. CRM Analytics Means Real Time Systems:
While the direct marketing and database marketing have perfected a large amount of the analytics to serve them, the special aspect of the CRM analytics is that the results/scores/statistical recommendations should be available for the rep or for the web to provide the right communications to the customers. Thus, traditional analytical services typically off line analysis, while will add power to the whole CRM operation, especially when we are talking about more systematic analysis of customers/consumers, the system has to be built around the concept of real time delivery of decisions /scores /statistical recommendations. The types of questions for which answers will be very relevant for a customer service representative are the following:
What are the optimal companion products for a customer?
When is the right time for up-selling?
What are the cross products that could be sold to the customer?
Note that though the answers to these questions are opportunities to sell products and services to customers, the CRM system should also provide answers to all the FAQs of customers on their current products. The CRM system should be updating the latest answers to the FAQs, which becomes a knowledgebase for the customer representatives to provide right answers to the right question from the right customer.Unless we provide an excellent service on their questions, we will not have the opportunity of selling new products or upgrading the current product.
7. Some standards to follow in creating value metrics:
- Always justify ROI calculations based on transaction data, not based on coupons and rebates. In some industries the fraud in usage of coupons and rebates is as much as 40%. Now figure how good your CRM ROI could be.
- Keep well defined privacy standards and hence make sure that the data collection and data utilization process does not interfere with these business rules. This is really a complex issue in pharmaceutical industry, but even non-pharmaceutical industry can have very subtle issues that need to be addressed carefully not to violate these privacy and confidentiality.
- Remember results, which can have the interpretation of other things remaining the same, is more valuable than methods which can not provide such basis, as we always need to make incremental or comparative statements in metrics to the tune of apples to apples comparison. This alone is a good reason, why we need sophisticated statistical relationships and not just two-dimensional tabulations.
- Of course, if we use sophisticated statistical relationships, make sure such relationships are not affected by extremities in data points, a common trap interpreters fall into if analysts are not careful in the analytics.
- Understand and challenge where and how the line is drawn between patterns vs. noise in the data; this can be as philosophical as this can be the tool to understand patterns in data.
- Be clear and demand from the analysts that the signal to noise ratio in terms of dispersion around the pattern is more than the dispersion in the noise.
- Know metrics that need to be looked for trends vs metrics that need to be used for understanding causal effects that would explain incremental changes.
- Always test ideas using right test and control samples and have enough samples to conclude meaningfully. Controls are what proactively sets the stage for apples to apples comparison of metrics between two groups of experimental units.
Without the analytics, companies will never be able to discern where to put their limited resources and to provide differential services to their customers, better and higher levels for most important customers and tiered services for others. The concept of “fail forward fast” to converge on successful initiatives will never happen, and before the companies reach profitable equilibrium after many trials, they might go out of business. Analytics provide systematic and scientific ways of analyzing and understanding the customers, processes and initiatives.
Use analytics to fast forward to success.