Sudoku starts very much like marketing analysis: We have some but by no means all the data to solve the problem.
The popular puzzle game sudoku is an exercise in analysis and logic; contrary to appearances it is not a data or numeric problem. The objective is to arrange the characters 1-9 in a large matrix so that each row, column, and 3×3 square contains only those digits. The recent “baseball sudoku” variation proves this point: The objective is to arrange each of the nine baseball positions—left field, center field, catcher, etc.–in the matrix according to the same rules.
This leads to two conjectures regarding marketing analysis, particularly as they relate to strategic questions:
1) The information used in an analysis is often incomplete, thus raising the uncertainty of the conclusions drawn. Numbers on the page or in the spreadsheet suffer from two potential limitations with respect to strategic questions:
• Because data usually come from transactional systems, they reflect a current set of business rules and processes aimed primarily at improving efficiency—in other words, they capture well-defined events. Therefore, marketing effectiveness must often be inferred or teased out of the behavior embedded in the transactions.
• The “garbage in gospel out” syndrome–people have a tendency to believe what they see, particularly if comes from a computer and is “quantified.” This point was made famous by Benjamin Disraeli: “There are three kinds of lies–lies, damn lies, and statistics.” So, as when using the Internet, you need to be very sure of the source and particularly the context of the data you use in decision making.
2) The absence of numbers, or data, may lead to the conclusion that the analysis is flawed or not even possible. Very often the lack of numbers, in and of itself, is associated with a perceived lack of value. Consider:
• Strategic questions focus on figuring out what we should do, not on assessing how well we have done. As a result, the answers aren’t likely to be rooted in transactions but rather in a variety of sources and estimates. Strategic analysis often results in building a consensus among the experts based on their collective experience.
• Strategic questions have high risk/reward profiles and thus an increased pressure to prove the point with data. The unfortunate result is often a serious dilemma for analysts: “We can’t answer the question you asked, but we can answer this one (read: we have data), and it might be a good substitute.”
A key point to remember is that while data are critically important to understanding what one should do, they are not synonymous with the term “analysis.” In fact, the definition of analysis doesn’t include the concept of “number” in it until you get to a branch of mathematics dealing with calculus and limits.
A quick Internet search yields the following definition: “Analysis is an investigation of the component parts of a whole and their relations in making up the whole.”
In short, analysis is answering the question “Why are things the way they are?” Once we understand that, then we can address questions around what the world might look like if we did something different. Whether it is situation analysis, SWOT analysis, consensus planning, or other strategic tools, the process of breaking down a topic into its constituent pieces to better understand the marketplace is analysis in its purest form. In the language of the chief financial officer it means reducing business risk by quantifying uncertainty.
Understanding customer needs, desires, and motivations as well as understanding their behavior are also key marketing analytic challenges. Data certainly play a role in both identifying component parts as well as confirming any conclusions about their relationships drawn in this manner–but the focus should be one of understanding and assimilating; not on reporting or explaining the status quo.
So what happens between the strategic planning sessions, where analysis is supported by data that are often incomplete, and the day-to-day operations, where analysis without data is deemed unacceptable? There appears to be a shift from thinking analytics to numeric analytics. Articles are full of references to “drilling down,” “data mining,” “scorecards,” “dashboards,” and other data-centric phrases as representative of analysis. It has come to the point that “we don’t have the data to do analysis” has become a crutch if not an outright excuse for not analyzing. This is a mistake, because analysis will help you identify the priorities for which data are critical to the conclusion.
The big important questions, such as whether you should double your marketing budget, are rarely phrased in ways that are easy to answer with data. Because transactional data, the common denominator in most analytic systems, result from operational processes they rarely reflect what might happen if you changed all the rules. The experts on teasing out what happens when you change the rules slightly use data to estimate the effect on sales when various levers are pulled, tweaked, and moved around. But as they will all put in a footnote, there are significant assumptions associated with any historical-data-driven approach.
Simply put, data-driven analysis can’t explain anything that isn’t represented in the data in the first place. So to get the data, marketing must test. Advertising weight tests, simulated test markets, and split-cell campaigns are among the ways of generating data in order to read the results.
As the questions become more strategic there is a greater emphasis on insight and guesstimates. In his book “Marketing by the Dashboard Light,” Pat LaPoint points out that we all need to become better guessers. The good news is that this is a skill–one rooted in analysis and collective experience.
This leads to some conclusions about analysis in general and marketing analysis in particular:
First, start by deciding where you want to be, rather than with where you are. Goals and objectives, including ways of measuring success, should be routine discussions. From a systems perspective, resist the temptation of implementing low-hanging fruit now and figuring out the rest later.
Second, marketing’s creative mandate should be built on a foundation of analysis for the express purpose of learning how best to satisfy customer needs. It is not about testing the creative; it is about assessing the effect of creative on performance.
As mentioned at the beginning, sudoku resembles a marketing problem. In fact, in an article for ChiefMarketer.com (a sister Website of MultichannelMerchant.com), marketing strategist Lane Michel recently described it similarly: “your customer database probably looks like a sudoku puzzle with a lot of blank cells in the grid [and]…. the most insightful results come from figuring out what data gaps are important to fill.” So the guidelines for sudoku seem to be appropriate for marketing analysis as well:
1) “When you’ve eliminated the impossible, whatever remains … must be the truth.” Explaining the effect of marketing programs on customer behavior is often an exercise in inductive logic whereby general rules are based on limited observations.
2) “A place for everything and everything in its place.” Segmentation schemes categorize customers into traceable groups so that the whole equals the sum of the parts. Moving customers between segments is often a key strategy for improving customer equity.
3) “Conspicuous by its absence.” Opportunities often exist where there isn’t activity today, and the core focus on innovation and growth suggests that marketing will be constantly searching out the white space and determining if it is a profitable target.
In the end, marketing analysis must be accountable, transparent, and defensible. That does not necessarily mean it comes in the form of a report. The best analyses are delivered in Word (or possibly PowerPoint), not in Excel. Why? The process of distilling the information into conclusions, providing relevant context, and making recommendations regarding a course of action is the job we as marketers were all hired to do.
Anthony Power is a principal with Power Consulting, where he is responsible for helping clients create strategies, positioning, and solutions for innovative marketing products and services. He can be reached at email@example.com.