From BI Review Magazine | December 2006 Issue
December 1, 2006
Data excellence is
proving to be the critical key to marketing success. With the advent of
desktop econometrics and market response modeling, companies with mega
budgets can finally determine which marketing activities are effective and
which are not. Econometrics examines the relationships over time between
marketing mix variables that are controlled, and performance measures, such
as sales or market share, that represent the outcomes of marketing plans.
The relationship among all the variables is teased out, so that companies
can determine the lift that each marketing activity can attain.
can optimize their budgets to determine the ideal overall dollars as well as
line item amounts that will accomplish goals such as maximizing customer
equity, generating new customers, gaining market share or improving cross
selling. Financial services companies with mountains of data are leading the
way in marketing resource management. However, it won't be long before all
industries are using econometrics and optimization in their analyses.
The cornerstone of the
analysis is data excellence - not just quality data, but the right data -
captured correctly and measured appropriately. We have found that a number
of the Fortune 500 companies we've worked with have many common data issues
that inhibit accurate analysis. Here is a selection of recommendations based
on the major problems that we've encountered.
Corporate Goals: A company may have
different strategic goals that are in conflict with each other. The goals of
market share for the marketing department and profitability for the
corporation oppose each other. The ramifications are broad and include, for
example, that different departments are compensated for achieving
conflicting goals. Case in point: A company may desire optimum budget
allocation yet achievement may be difficult with contradicting goals of
market share gains as opposed to profit increases. The data needs to reflect
corporate goals from the top down.
2. Define Required
Data: Oftentimes, the data that is captured does not measure the accomplishment of
the strategic goals. For example, a company whose goal is market share does
not capture or track any market share data on a consistent basis. They were
hard-pressed to locate market share data and couldn't relate the rate of
change in their share.
Necessary Data Only: Reams of accumulated data may be irrelevant in measuring the success of the
goals. One company captured over 50 different data sets, yet after defining
strategic goals and identifying the metrics needed to track accomplishments,
found that only 17 were necessary. Data should have a direct tie to
measuring results of the goals.
4. Keep Data
Consistent: Data inconsistency makes analysis difficult. Within the marketing
department, different regions may be capturing different data in different
formats. Corporate data protocol should be established on what data is
captured, how often and the measurement.
- Standardize the database. In one company, the
marketing data from various regions was gathered in Excel, Word and
PowerPoint. Data should be captured in a database and those databases
should be consistent across the organization.
- Standardize the frequency of data capture. In another
marketing department, one region tracked data weekly, another region
monthly and still another annually, making regional cross-sectional
analysis impossible. Another company gathered weekly data and then
aggregated it to monthly data every quarter. In econometric analysis the
more data points there are, the more detailed the analysis. Fifty-two
weekly data points are immensely more valuable than 12 monthly data
- Standardize the metrics. Tracking the number of event
attendees versus the amount of money spent on the event is not a
consistent metric for events. Metrics should be standardized, so the
analysis is comparing apples to apples.
5. Develop a
Corporate Wide Data Policy: This should include department guidelines and a data
dictionary. All too often, we're given undefined data sets and managers
can't explain definitions because the database developer has left the
company. Institutional data memory should be documented. Explanations should
include the rationale for capturing each data set, so the employees aren't
operating in the dark. They know what and why they are tracking specific
6. Develop Data
Responsibility and Transfer Procedures: If an employee is responsible for data gathering, tracking
and monitoring, build it into the job description. Define the transfer
protocol from one employee to the next. In a number of instances, missing
data was blamed on the previous employee, long-gone from the company. When
an employee leaves the company, the exit process should include review of
the database for which the employee was responsible.
Tracking Timelines: Expense tracking should be recorded when applied, not accrued or when the
bill is paid. Part of econometric analysis is linking marketing spend to
results, for example, the number of new customers acquired. If magazine
advertising was purchased in January, but the ad didn't run until July, the
expense should be noted in July, so it ties to the results.
8. Vary the
econometrics, if the expense is exactly the same month after month with no
variation, lift cannot be determined. Expense amounts need to change to
ensure that there is variation in the data, so the analysis can identify the
change in the relationship with other variables.
scrub their addressable data to ensure accuracy, most are not committed to
data excellence. While data mining has permeated corporate America, data
excellence is struggling and lagging behind. Data excellence can drive
budget decisions for a competitive advantage. The winners will be companies
with strategic goals aligned with data capture, comprehensive data protocol
and standardized measurement, which will lead to superb analysis and
optimization of the budget.
Gone are the days when
CXOs couldn't figure out which of their marketing activities were working.
Econometrics and optimization takes the guesswork out of the equation and
data excellence drives the process.
Barbara Lewis and
Dan Otto are principals at MarQuant Analytics, which helps companies define,
capture, measure and analyze data to improve marketing productivity. They
can be reached through their Web site at www.MarQuantAnalytics.com.