"Our customers only like to receive emails, they would never open a direct mail piece, our customers do not respond to communications unless there is at least a 20% discount and our customers will only open emails on Wednesday."
A very concise and straightforward description of customer behavior. The problem? The sample size is only 10 people and half of this is anecdotal and not based on survey data. In our very well-intentioned desire to understand our customers, it is easy to take a little bit of intelligence and make sweeping generalizations that become ingrained in organizational thinking. I have watched, sometimes in horror, as teams institutionalize what is really a good guess or an opinion of who the customer is and what they need and want.
The challenge is to gauge how big a sample size is big enough to ground the data in reality? It certainly depends on the overall size of your customer base. I have worked with very small data sets of customer segments and needed to make generalizations from a small amount of data. The smaller the data set, the more you need to monitor, dig, re-survey and not assume that there are not variances that can change your thinking over time. Include the context around the data you are analyzing. "This is 10% of our customer file, this is 50% of our most recent purchase segment."
When analyzing and presenting customer data, always indicate when the data is directional and not absolute. Trends are important, but need to be looked at with a certain amount of skepticism. Let's try and keep OPINION out of our customer definitions. Let's try and keep GUESSES out of our customer definitions. I have found it to be tremendously difficult to change the team thinking on the customer once it has been established. Tread carefully in defining what you see happening with your customers. And if all else fails, you can always survey the customers directly and ask them who they are and what they want! And if you can, please make it more than 10 people!