NPS\u00ae\/CSAT attitudinal measures (I use the terms interchangeably) are all very well but companies make a fatal error when they confuse Output with Input in the design of their touchpoint metrics. Too many times I have seen companies try to apply an NPS\u00ae\/CSAT measure to all and sundry items which (a) fails to action anything as a result (b) encourages a \u2018tie us in knots\u2019 mentality.<\/p>\n
Consider this. \u2018Ease of use\u2019 of say website is a great output you might want to improve. Note the importance of the output being written in \u2018attitudinal\/ emotional\u2019 language a customer might say to themselves. You may even relate NPS\u00ae\/CSAT on this measure as correlating highly with some behavioural measure.<\/p>\n
Company X now targets itself to improve its \u2018ease of use scores\u2019 by 10% on the year.<\/strong><\/p>\n But what does ease of use mean?\u00a0 Here we get to the nub of the issue.\u00a0 Ease of use like most Customer Psychology measures is a \u2018Formative\u2019 variable. This means that like the concept of \u2018a person\u2019 it comprises many parts (legs, arms, etc) but those parts are generally independent of each other (when you raise your arm, your leg doesn\u2019t rise as well!).<\/p>\n Why is this important?<\/strong><\/p>\n Well if \u2018ease of use\u2019 is formed of many parts – the look of the website, the font size, the download speed, the click through speed, the way the content is written as well as internal variables such as how one department speaks to another..<\/em>. \u2013 but only the output matters, you cannot then seek to isolate its individual components in an ROI formulation.<\/p>\n If you were now to take the logic of Output ROI into this Input ROI you would be committing a fatal error.\u00a0 Questions like, what is the ROI of the click through speed, the ROI of the font size or even the ROI of how one department speaks to another are \u2018impossible\u2019 to demonstrate in isolation even though they are critical to formulating a better \u2018ease of use\u2019 experience.<\/p>\n Hence, after you quantify your principles, you must move from Science mode to Art (creative) mode.<\/p>\n For a physical example consider a car. A customer could rate the \u2018speed and performance\u2019. You now might want to increase this output rating (or judgement) by say 15% year on year (i.e., quality performance improvements).\u00a0 So far so good, but now if you move Output ratings into Input you might start looking at developing an ROI on the quality of the wires going into the engine. Of course, even though it is fundamental to performance, it is unrateable, an ROI would fail at this level and you would do nothing.<\/p>\n Management Implications<\/strong><\/p>\n 1.\u00a0<\/strong>Be careful not to pull your Output NPS\u00ae\/CSAT measures into your Input NPS\u00ae\/CSAT measures.\u00a0 This means you must identify the principles (key moments) that count in the mind of the consumer and then moving into creative mode to formulate the inputs (any ROI barrier at this level will be like trying to put an ROI figure to those wires in the engine). Of course, you still need to demonstrate success, but you do that through piloting.<\/p>\n 2.\u00a0<\/strong>Always base your ROI on customer impact, best judged by both hard metrics (spend now) and soft attitudinal metrics (CSAT\/NPS\u00ae or Preference\/ NEV (net emotional value) \u2013 what the customer actually thinks and feels about you (not the same as how much they spend!). This is important to avoid the \u2018junk mail\u2019 effect i.e., acquisition rates rise, but your customers hate you and their tendency to habitually stick with you or look for an alternative supplier becomes eroded (the propensity towards \u2018search behaviour\u2019).<\/span><\/p>\n 3.\u00a0<\/strong>Consider that customers only rate based on current knowledge, embed and pilot innovation. Making it up is highly valuable in a re-design. Culturally this means openness to innovation and a reduced dependency on the failings of quantitative research (see: The <\/span>10 Things wrong with Quantitative Research<\/a>).<\/span><\/p>\n