The adoption of CRM as a core business strategy hinges on proving success - but first marketers must figure out how to define success.

A major packaged foods company was recently mulling over whether customer relationship marketing made sense for them. While their flagship brand had long enjoyed a huge category lead, the syndicated trend data was showing a worrisome slippage in the share of requirements, suggesting consumers were beginning to stray.

The circumstantial evidence pointed to aggressive promotional pricing by competitors - mainly, private label brands - as the crux of the problem, inciting value conscious consumers to trade off their brand allegiance for cost savings. In a category with flat to declining growth, the company could not afford to see any of its heavy users switch over to the competition. So in order to meet its fiscal targets, brand management was fighting back in the grocery aisles with point-of-sale discounting of its own. But all of that promotional skirmishing had created a circular loop: it was training consumers to buy on deal and heightening their price sensitivity, pushing them further away from the brand.

Yet in the market research department the outlook was decidedly less bleak. Skeptical of the syndicated data, the research staff trusted more in their own custom data which told a cheerier story: that, in fact, the brand's share of requirements had actually risen slightly over the previous year. So who was right? Or, more precisely, which set of data reflected what was actually going on? At stake: a major shift in budget dollars to solve what might be a spurious crisis.

Fuzzy data has always been the nemesis of marketers. No other area of business has so many ways to measure success - and so little success at measurement. Share of market, share of requirements, share of wallet, share of mind - the list goes on; yet most marketers will admit that figuring out return on investment is purely a speculative exercise for them. For one thing, intangible assets like brand equity defy monetization; on top of that, there is no single accepted measurement of marketing performance.

That's why a lot of the time marketing spending decisions can seem capricious. They get made out of habit ("It's always worked in the past"); on blind faith ("This is a breakthrough idea"); or out of boredom ("We have to try something different"); seldom is the decision the logical outcome of a bottom-up computation. And that's not because marketers have an aversion to calculus: it's simply because the measurement tools they have to work with are so blunt. No one has come up with a credible method of proving marketing payback; as a result, marketing strategy is usually a homemade mix of "educated guess" and gut feel.

The precise appeal of CRM, on the other hand, is its measurability: the idea that what truly matters is the growth velocity and asset value of the customer base. But those metrics can be difficult to fathom even for companies squatting on terabytes of transaction data (like grocers, telcos and general merchandisers). Many managers - particularly marketers - suffer these days from what Richard Saul Wurman labels "information anxiety", a phobia he describes as the "black hole between data and knowledge" - between the knowledge that is needed and the information actually provided. The irony is that after a generation of information deprivation, marketers are now faced with information overload. And so they lack the experience working with customer-level data to know the right questions to ask (or even the mere point of asking them).

To complicate matters, there are often competing measurements within a company. Should management worry more about product quality scores, service satisfaction ratings or brand equity strength? Yet asking the question "misses the point", according to the authors of a new book called "Improving Customer Satisfaction, Loyalty and Profit". The writers argue that, "These factors form a chain of cause and effect, building upon each other so they cannot be treated separately".

An integrated measurement system that shows an indisputable correlation between bottom-line results and customer loyalty is certainly the only way to prove the merits of CRM. But before that can be done a business must first take stock of its information assets; otherwise, it has no idea whether the data it needs to define success is even available. "Understanding the structure and organization of information permits you to extract value and significance from it," Wurman says.

Instead of treating information as a waste product, to be swept up and discarded upon use, it needs to be disassembled, re-structured and aggregated in such a way as to serve up the critical relationship measures that marketers yearn for. Those measures get stored in a data warehouse but the methods of value extraction are spelled out in a knowledge management plan. "Knowledge management is not a technology problem; it is a process problem", points out Amrit Tiwana in his book "The Knowledge Management Toolkit". Or, to paraphrase Pete Drucker, it is all about putting the 'I' ahead of the 'T' in IT.

A Customer Knowledge Management plan encompasses the methods by which customer information is collected, including their origination sources and refresh cycle; how that information needs to be structured according to its end use; and the processes by which it can be interpreted and digested by all stakeholders. It answers the following questions:

• What Information Do We Have?: How much do we know at what level of granularity about customers? How is current information being used to profile, segment and manage customers? What are the key metrics by which success can and should be measured?
  
  
• What Information Do We Need?: Are there any significant information gaps that compromise the ability to carry on a continuous relationship with customers? How can we obtain that information at what cost?
  
  
• How Do We Apply That Information?: How can we better use information to understand customers, make decisions and analyze results? What levels of abstraction are required to produce meaningful knowledge out of bottomless fact tables?

The tangible expression of a knowledge management strategy is a Relationship Scorecard that summarizes the behavioral attributes of the key customer groups; the main drivers of value, satisfaction and loyalty; and the holistic set of metrics that reveal incremental gain. By keeping more accurate score of the relationship measures that count, marketers will finally be in a position to know the exact meaning of success.