|Sign-Up for Free Exclusive Services:||Portals|||||eNewsletters|||||Web Seminars|||||dataWarehouse.com|||||DM Review Magazine|
|Covering Business Intelligence, Integration & Analytics||Advanced Search|
The ROI of a Customer-Centric Database
As a marketing professional, you understand that customer information is critical. You also know that organizing information from many systems, external files and integrating campaign and response management into a single, accessible set of customer views will translate into profit for your organization. If you didn't believe this, you would probably not have the word marketing in your title.
Most of the time, however, the rules of well- run businesses require that this intuitive knowledge be translated into numbers before any capital is committed. Therein lies the rub. Often, marketing is more an art than a science, so translating what makes perfect sense into spreadsheets and graphs with lots of demonstrable improvement in revenue generation and cost allocation can be tricky. Add to that the dearth of readily applicable research, and you face a dilemma. With that in mind, this article provides a framework for identifying and quantifying the return on investment (ROI) for the construction of a customer-centric database. The theoretical foundation provided has been drawn from wide experience with customers and augmented outside research.
A customer- centric database provides one fundamental value over other database systems - a total customer view. This view is both vertical (multiple file sources) and horizontal (longitudinal history), which allows for better matching of a sales/marketing proposal to that customer. Stated slightly different, it allows for customers to be valued and ranked according to their match to a specific sales/marketing communication. That means getting closer to the maxim of delivering the right message about the right product to the right customer at the right time.
For example, if we told a network equipment sales specialist that we could give her a list of companies whose networks failed today due to over-capacity, she could visit them tomorrow. But, how confident would she be that she could close the sale? How valuable would that customer visit be versus any other use of her time? Now, of course, we don't have that exact information on the database. However, what if we did a query identifying all sites with "entry level" networking equipment whose sales revenue and number of employees has risen dramatically in the last year? Wouldn't that be close enough?
In a marketing environment, if we could identify those most likely to respond to a particular promotion based on past response history and current product and attribute profile, wouldn't we be better served by allocating scarce marketing dollars to those sites as opposed to any other random site?
Thus, given a total customer view, we can, for any given proposal, create a valuation formula to rank and prioritize our customer/prospect base. This formula requires that as much information as possible be organized into a customer view (hence the necessity for the type of system we are proposing). We can then optimize our budget for any given sales or marketing activity by selecting from the top and working down the list, always sure that no matter how far we go, we will always be selecting the customer with the best fit, and thus the highest propensity to generate revenue.
Figure 1 assumes a total population of 1,000 customer/prospect targets, and enough budget to mail to 195 of those targets.
The rational valuation prioritization scheme will generally outperform the random prioritization scheme, regardless of what the predicted propensities are, or even if they differ from the actual response rates. The reason for this is simple. As long as the prioritization formula results in a ranking where, relatively speaking, the highest propensities float to the top over the lower ones, counting down from the top will usually give a higher result.
A proper valuation formula requires two things: An intelligent application of expertise or data mining algorithms, and a comprehensive, accurate database organized by customer view to use in the calculation. In Figure 1, the lift between the random and rationally prioritized was 98.6 percent. Depending on the sophistication of the model, this lift can be anywhere from 25 to 100 percent. Maintaining a conservative posture, we will put this lift at 25 percent.
Assuming that the conversion rate of responses to sales is five percent, this would indicate that sales could be increased by 1.25 percent using rational valuation.
Figures 2, 3 and 4 attempt to quantify actual revenue increase using rational valuation in three areas of your enterprise: sales, cost of sales and bad debt. The assumption is that, given a current baseline and growth path (the natural rate of increase), and some assumptions about raising incremental response through better targeting of sales and marketing activity, an extrapolation of additional revenue generated is possible. Obviously this is an illustration and requires many other factors to remain constant. However, it does provide a framework to link better targeting to increases in revenue and decreases in cost.
To calculate sales increase, enter a baseline sales amount assuming current random valuation efforts for "Period 1" and "Period 0." From this, the model will extrapolate for "Period 1" and "Period 2." Then, enter an assumption on the incremental lift derived from utilizing rational valuation over random valuation, as well as an assumed conversion rate of activity to sales. This will generate a multiplier to add to the natural rate of increase and thus provide a comparison of additional revenue generated.
Note that Figure 2 emphasizes the benefits of incremental revenue generation. That is, we assume a constant investment in activity and simply enhance the response rates to that activity. Another way to frame this analysis is to reduce the investment made, while still maintaining the same response rates: a cost reduction framework. In summary, then, it would only be necessary to mail 92 pieces to get the same 14.75 respondents. Thus, by not mailing the remaining 103 pieces, money is saved by more efficient use of marketing budgets.
Of course, there is plenty of middle ground here, and the actual ROI will include both cost reductions through more efficient use of funds, as well as incremental revenue from additional responses.
Another area where a customer-centric database provides ROI is cost of sales. By increasing the quality of leads and allowing better prioritization, the cost of sales can be reduced. Figure 3 quantifies these annual savings, which can be significant even at an incremental level.
Next, most companies write off a significant amount of bad debt as a result of customers unable or unwilling to pay. Since a customer-centric database can capture and store this activity, and these customers can be modeled, future targeting and prioritization can reduce this debt by avoiding a more careful engagement with customers who either have exhibited this behavior in the past or fit the model of customers who may fall into this category in the future.
Additionally, with better targeting, a higher sensitivity and awareness of customer needs is attained. This allows for a better match between customer and product, reducing the unwillingness of customers to pay for products and services. Experience has shown that this effect can amount to three percent annually.
Using these revenue increases and cost savings, generate the cost of constructing and maintaining a customer-centric database for the defined time frame. Utilize depreciation of capitalized cost and include all expenses. Compare this number to the sales increases and cost decreases in Figure 4 to determine ROI and break even points.
The "Halo Effect"
Incremental ROI can be recognized through a very specific, quantifiable effect called the "halo effect." Recognition is the operative word, as no additional responses are actually generated. Rather, through the mechanism of constant customer keys and complex customer definition, hallmarks of fully functional customer-centric databases, responses that do not appear to be attributable to promotion activity can be linked and measured through a closed loop.
Here is how it works. Typically, some element of response to a promotion cannot be linked to the original promotion, resulting in background noise that cannot be used to justify the ROI of a promotion. This occurs for a few reasons:
In a worst case, actual attributed response rates have been seen to increase by 400 percent over original from .25 percent to two percent (that's right; actual measurable response is four times greater than original attributed response). This occurs through a more intelligent, complete process for identifying customers, generating promotions and defining and capturing response.
Figure 5 shows how the numbers look, assuming we only achieve a 100 percent increase from partially to fully attributed response rates. Assuming a cost per piece of $2.50 leading to a total cost of $250,000, a previously unprofitable promotion becomes profitable when the additional responses are recognized and attributed. This is not due to any manipulation of results. Rather, through the intelligent design of a customer-centric customer identification and maintenance process, we are able to better capture the complex realities of promotion response, reducing the amount of seemingly random background noise and better understanding the return on our marketing dollars.
Understanding the ROI of a customer-centric database provides several benefits. First, it allows for the business to set proper expectations regarding where the savings will come from and where the incremental revenue will be generated. It can also help drive the important business, product and technical requirements, thereby making a successful implementation very likely. In addition, understanding the ROI will better enable the evaluation of both internal and external costs. Probably the most significant benefit to this knowledge will take the form of sleep - those involved in the project will have more of it.
For more information on related topics visit the following related portals...
Business Intelligence and Strategic Intelligence.
David Cameron is vice president of marketing and product integration at AptSoft Corporation (A HREF="www.aptsoft.com">www.aptsoft.com and can be e-mailed at email@example.com. He manages the marketing and product integration efforts, which include working directly with prospects and clients to better understand the applications and value of AptSoft's enterprise software. In this role, he draws on more than 13 years of experience in building customer-centric data and process integration applications for large corporations and applying them to strategic sales, marketing and service solutions. Cameron was previously the vice president of sales at Wheelhouse Corporation.