Wednesday, March 26, 2014

NPS and Text Analytics

Net Promoter Score (NPS) measures the loyalty that exists between a provider and a consumer.

NPS is based on a direct question: How likely are you to recommend the company/product/service to your friends and colleagues? The scoring for this answer is most often based on a 0 to 10 scale. The point is what can NPS measure, and what is its limitation?

First off, the NPS is not an individual-level score. It is calculated as a percentage that indicates the preponderance of consumer sentiment at some aggregate level regarding a brand or a company. That said, it doesn't correlate with key financial metrics such as top-line revenue, profit, stock price or EBITDA. This is to its credit since, unlike many other consumer metrics, this indicates that it's not correlated with size, market share or valuation. In general, it does track the direction and drift of consumer sentiment towards the thing being measured and, as such may do a better job of differentiating that sentiment from other widely used metrics like customer satisfaction, which for the most part is a measure of the potential for customer attrition. One of the most important pieces of NPS is the comments. Not only those of deflectors, but those of the promoter are very valuable.

Applying NPS is helpful when utilizing additional analysis. From an analytics perspective, it's fairly arbitrary in how it reduces to a single number the percentages of customers who responded with the different scores. NPS is widely used by larger consumer businesses. It's hard to ignore and often shows up in analytics of one kind or another. So a serious analysis needs to define what's being predicted or measured, and then determine whether NPS does a better job of that than some other function of the input percentages. One helpful approach is to apply text analytics to the comments received as part of the NPS survey. Very often the "Promoter" customer provides comments that can provide valuable input on areas that need to be addressed. Very often the Promoter's comments are the same as the "defector's" comment, the only difference is the numerical score.

Focus on key business drivers. There are things about the NPS that make it attractive as a measure of consumer sentiment, just not for its intended use as a measure of revenue growth. As noted, it's to its credit that it doesn't correlate with market size, share or valuation. The number by itself doesn't mean anything. The key to this is to get additional information from the customer to understand key drivers (internal - processes, people, etc. or external - product, messaging, customer education, etc.). While you manage to an NPS goal, you manage it by focusing on those drivers. Without that, it may as well be a random number based on how the customer's day is going!

Generally, the consistency in NPS results from a company measuring its product line or brand against itself: It is historical NPS last year vs. this year to see improvement, it is arguable upon how you can use it to compare as an industry benchmark or vs other products and companies since the inputs are not consistent (different market, geography, product etc..). Analysis that leverage its original raw rating -- not the aggregate score -- can also be informative and insightful for the purpose of targeting communications to specific consumer groups. And if the aggregate score is developed and used in the context of continuous consumer tracking surveys, then it can also aid in refining marketing strategy, spend and content.

Therefore, as most other measures, NPS is best when utilized with other tools. The biggest mistake in using it is to consider it the "sole" voice of the customer. It only makes sense as one more metric among many. Combining text Analytics with NPS provides more of a complete view of the customer concerns. NPS complements other efficiency metrics within the organization really well and such set of metrics can be a good way to compare against other organizations in the industry or even other industries that you are trying to learn about customer engagement from. Last but not least, any metrics, including NPS is means to the end, not the end.


Post a Comment