Net Promoter is a trendy way to measure how well a company is doing in its customers' eyes. It uses a single question, "How likely are you to recommend the company to your friends or colleagues" on a zero-to-ten scale, and subtracts the 6 and below from the 9's and 10's. Voila, the net promoter score.
The Net Promoter Score has some good things going for it: it's highly standardized, so benchmarking is easy. It's easy to understand and gives you a single number to focus on. And it's proponents claim it correlates well with useful things like customer loyalty and word-of-mouth.
But Net Promoter only really measures one thing, the customer's overall relationship with the company. And that's where things start to go off the rails. Many companies actually want to measure a lot of different things: how did that specific customer service call go, how was the salesperson, did the automated system work, and so on.
Rather than develop survey questions to measure those specific things, some companies try to adapt the Net Promoter question to fit. The results are not always pretty.
For example: How likely are you to recommend this customer service representative to your friends or colleagues? You can see the logic for the company trying to standardize on Net Promoter and measure the CSR's performance. The problem is that at most companies the customer has no choice in who they speak to when they call customer service. So there's no point in recommending the agent, and a certain percentage of customers will give a zero on the question for exactly that stated reason. That's not a problem with the CSR, that's a problem with the survey question.
Or worse: How likely are you to recommend this customer service representative's knowledge of products and services to your friends and colleagues? This question, as written, is meaningless. The intent is to understand the CSR on different qualities (knowledge, friendliness, eagerness, etc.), but you can't really recommend a person's particular skill in a vacuum. You can't say to your friend or colleague, "I recommend you talk to Sally's knowledge and Bob's friendliness, but Sarah's efficiency."
Fortunately most people catch on to the fact that these questions should not be taken literally, and that prevents the data from being completely useless.
But if you want customers to interpret your survey question in a way which has them answering a different question than the one you asked, why not just ask the question you want them to answer?
It's much easier to interpret the answer to a straightforward question like, "Please rate the customer service representative's knowledge of products and services."
So while Net Promoter has its place, don't try to fit that round peg into every square, octagonal, or star-shaped hole you encounter. Just ask the question you want customers to answer.