"There are no facts, there is no truth, just data to be manipulated." - Don Henley, The Garden Of Allah

Data quality is one of those essential issues that everybody wishes they could ignore, akin to cleaning out the garage or completing your tax forms. Yet tiresome as it may be, the issue of data quality is paramount to success for many customer-facing enterprises. Bad data kills - and the victim may be your customer experience management or Net Promoter program.

Suspect data colors everything--including the attitudes of your stakeholders. If managers lack confidence in your customer data, they will be hesitant to embrace your program. And you can hardly blame them: If they don't believe the data is good, why should they take action based on it?

No company today could operate without a basic reliance on financial and operational data to make business decisions (although a few seem determined to try). Why don't managers afford the same level of trust to customer data as a basis for major decisions? Perhaps if they did we would we see more winners and fewer sinners on our web site.

A quick comparison among these various types of data is revealing. One of the reasons we rely on financial metrics is because we believe they are valid and can be precisely interpreted. A combination of well-defined principles and years of shared learning have yielded a language of numbers that is universally accepted. Hard operational data has a similar status, such as Days Sales of Inventory and WIP in the factory, or wait times and abandon rates in the contact center. We trust these measures and we know how to act on what they reveal.

Why can't we count on customer data in the same way?

To be sure, it's not because managers lack interest in the information. Senior managers consistently rate insight into customers as one of the most desirable types of information. However, they are frustrated by the two issues that make financial data so valuable: quality and ease of interpretation.

Let's focus more closely on data quality. Companies that market to consumers build confidence in their data by applying rules of statistical significance. Get the right number and profile of people in the sample and you can bet you have insight into the entire group's behavior. However, companies who make their money in complex, high value business-to-business deals cannot play the numbers games so easily. In this instance, sampling doesn't really work. Worse, it falls over in plain sight, with everyone watching, usually with a metaphorical thud.

Sales managers often wield the killer blow. Reviewing the data at the customer level, programs rarely supply the right respondents that beg to be taken seriously. Instead, reviews must contend with an endless spiral of debate as to whether the "right people" were consulted, made colorful at the margin by personal anecdotes from the sales force--"I was there last week," etc. This quandary is unlikely to be solved by reference to "significant response rates." Sound familiar?

The fact is, it is hard to recruit and maintain the right level of respondents to gauge customer experience, but acknowledging that basic challenge is at least a start.

And now the good news: B2B companies have a natural advantage here since their customers have a greater vested interest in their success. The big difference between doing business with consumers and with businesses is that the latter should respect your right to make a profit. Yet this advantage is often squandered by lengthy questionnaires, a flawed approach to sampling, and poor follow up.

How do you turn the tides? The starting point is to take a census rather than a sample.  Aim to get every quality respondent to respond, and assume that a non-response has meaning (not usually positive meaning either). In other words, shoot for a response rate of 100%. You won't get there (although best-in-class companies often reach 70%), but adopting the census approach is key. Companies that break out of the "sampling mindset" inevitably build more meaningful customer experience and Net Promoter programs.

At the risk of leaving you with a shopping list, here is a shopping list:

• Conduct short surveys. Long surveys tend to reduce response rates and scare off those precious executive respondents. You are better off sacrificing length for response rate and frequency.
• Stratify respondents. Segment your respondents into groups and develop distinct strategies to recruit and survey them. You'll get better results if you tailor your surveys to each audience.
• Sell the program. Make it clear to customers that they will ultimately benefit from responding to the surveys by being instigators of change. Engage the team that owns the relationship (often sales) and keep the customer experience management program at the forefront of all communications with the customer.
• Close the loop. Execute closed loop follow up processes, flawlessly. For every response that doesn't hit your loyalty threshold, quickly get back to the respondent. Prove to your customers that you are listening and taking action. If you ignore their responses, they will infer that you don't really care and will not bother to respond the next time. Smart companies follow up with everyone who falls short of a promoter.

Once you start treating the information-gathering process as a census, rather than a mass survey, you will adopt a different attitude about how you manage respondents, recruit participants and interpret data. This is the start of truly managing customer experience, and ultimately driving profitable growth.

Okay, you've heard enough theoretical rambling from me. How about sharing some of your practical experiences? If you've had some success with a census-based approach, post a comment below.