Put an End to Data Fatigue (1,274 words)
By Gina Valentino
Are you working harder than your database?
Perform this quick self-check:
- You're tired of running analysis.
- You're tired of linking spreadsheets.
- You're tired of marketing data but not matching merchandising data or financial data.
- You're tired.
If you've checked "yes" to any of the observations in our self-check, you're working much harder than your database.
Perhaps you don't mind. But try to imagine a world where the database works for you, providing timely and relevant information—not just data. You don't need Steven Spielberg or George Lucas for this; just a few simple changes and consistent metrics (a.k.a. ordinary but meaningful bits of information) and your database will be a workhorse. Here are five strategies to get you started.
Reduce the Number of Reports
Information overload is the beginning of analysis paralysis. More importantly, the more raw data you have, the more interpretations you'll have. So you need to reduce the number of reports you submit each reporting period.
A scenario you've probably experienced: A meeting room full of co-workers reviewing the same report will come up with an equal number of opinions about what the data exhibits. Take the leadership role and determine what summary reports are most beneficial to your organization. Omit the indiscriminate pages of minutiae and choose three reports that summarize performance particular to your area of responsibility. Since data are abundant, carefully selecting a few key performance indicators provides others in your organization with a clear understanding of the business.
If you're in charge of the customer file, you may want to show the overall health of the file—new customers, multi buyers, gross demand per customer, conversion and retention rates. If merchandising is part of your job, show inventory position, margin and net sales. The easiest way to decide what summary reports are important to your organization is to ascertain what key issues frequently are discussed in meetings. Determine what information would help the greatest number of people and keep everyone focused. Provide the same summary reports every reporting period and make comparisons to last year's budget, plan and/or forecast. Management would rather spend more time reviewing the information than trying to decipher a new format.
Database Reports Must Do All the Calculations
Too many of us look at the hard copies of database reports, input those data into a spreadsheet, create formulas to crunch the numbers, and format the report with lines, symbols, subtotals and page numbers. Technology supports the Olympics of mathematical gymnastics.
Ask your programmer or systems person to build the information on the summary page, or have an outside contractor develop the additional requirements into the design of the reports. The expense for a finite period of a contractor's time is cheaper than building the information infinitely yourself.
Another option is to build spreadsheet-driven formats and have a programmer or systems person schedule direct downloads. This way, on a daily, weekly or monthly basis, the database automatically will update the spreadsheets and store the calculations and comparisons. Then you'll have the luxury of reviewing the report and interpreting the information.
Reliable Data are From One Source—Even if the Data are for Different Uses
In a perfect world, marketing data are immediately captured from the order entry system—just like financial data. In reality, often the marketing data and customer-specific data reside in a separate relational database, whereas financial data may be sourced from a transactional flat file. While this appears to be dysfunctional, the purposes are genuine and do eventually marry.
Take the example of a first-time buyer purchasing four items at $25 each from the spring catalog with one item on back order. Marketing wants to capture the entire gross demand sales of $100; finance only captures the shipped sales of $75 plus S&H; and merchandising wants the gross demand sales of $100, the $25 back order and the $75 shipped sales and S&H (to calculate margin). And, of course, all three will want to know if the customer cancelled the back order of $25.
Who is right? All three departments are correct, but for different reasons. Each manager should report only the pertinent information for his/her area of responsibility. This way, data are consistent every reporting period.
Always indicate a data source on the summary report (e.g., data source: marketing system updated 11/02) to help the users understand the foundation of comparison. This will keep credibility high and tension low.
Interpret the Data Before Someone Else Does
As a manager, review the database reports to understand the information and interpret the findings. You have a responsibility to the organization to use your insight, experience and knowledge to dissect the information and draw conclusions. If the customer file counts are lower than those from the same period last year, you need to communicate the reason for the change. For example:
- Are you circulating fewer customers in order to reduce expenses?
- Are page counts lower this year?
- Did you mail later in the season?
- What is the complexion of the file—do you have more one-time buyers versus multi-buyers?
- Is the average order value higher, producing higher gross demand dollars than last year, but response is lower?
- Is the merchandise mix having an impact on items per order?
- Are back orders or cancels higher?
- Did the merchandise density change, or perhaps the creative presentation was revamped?
- Was a special offer promoted last year?
There are many reasons, often planned and expected, for differences. You need to be the person communicating the information and guiding the rest of the team through the data.
The greatest source of information is found in the database—but it's perceived as an esoteric discussion for the technologically advanced. This does not have to be the case.
Be assertive and request an information session with the programmer. If you feel apprehensive, ask to see the data model and dictionary. You'll receive an organizational chart of the data (data model) and a job description for each field (data dictionary). This is a visual representation of the data and the relationships of each field. The programmer can explain the flow, show you examples and answer questions. Many times, the call center uses one system to support the process of taking an order, verifying payment and relaying inventory information to a customer. This is good information, but not all-encompassing. The marketing system typically is a byproduct of the transactional database.
Reporting customer activity is specific to a single record with many transactions over a customer's lifetime. This activity starts when a customer is a catalog requester. A customer record is established, but purchasing has yet to take place. As a manager, you want to have visibility to the non-buyers. When they convert to buyers, what will happen when they use a gift certificate or the order contains back orders and the customer cancels it? Each circumstance can be shown in a data model and explained in a data dictionary.
If your firm doesn't keep a data dictionary, then it's time to start. Keep in mind that there's no universal standard. Each company will do this differently, but the components will be constant.
Over time, you'll be able to speak fluent database jargon and cite examples of a split order on suspense with the customer file tagged but not updated.
Implement these five strategies to help reverse the transformation of a tired database to an accomplished interpreter of key performance indicators.
Gina valentino is vice president and general manager of J. Schmid & Assoc., Shawnee Mission, KS. You can reach her by e-mail at firstname.lastname@example.org.