Michalis Michael, CEO of DMR, discusses whether the Net Promoter Score is set to be replaced with more relevant measures
More innovative means of measuring customer experience may be needed in today’s increasingly digital world.
How can businesses measure the success of their marketing efforts? How does their current and future performance benchmark against competitors? How can they work out, for example, the levels of satisfaction and loyalty felt by their customers? The rise of social media during the last decade has simultaneously made these questions easier and in many ways more difficult to answer.
On the one hand, the internet is bristling with all the necessary data required to determine how a given business is performing, as customers willingly – even eagerly – share thoughts and opinions which provide insights into such vital issues as customer satisfaction. On the other hand, the sheer volume of the data available can make it challenging to separate the essential from the non-essential.
With the amount of potential key performance indicators (KPIs) provided in a world of social media, all businesses – large or small alike – need to address the fact that some indicators are more key than others.
On top of this, some companies – as noted by McKinsey partners Frédéric Gascon, Raffaele Carpi, and John Douglas – “measure and manage performance through lagging indicators, such as compliance with monthly output or quality targets. By the time the results are known, it is too late to influence the consequences.”
Clearly, then, the stage is set for new ways to measure performance: methods which are up-to-the-minute, capable of leveraging AI and machine learning technology to sift through swathes of data, and able to articulate actionable KPIs in a simple and accessible format.
Putting people at the heart of digital marketing
Out with the old
This is not, by any means, a new idea. In 2003, the concept of the NPS – Net Promoter Score – was born, and its simplicity reflects a real and ongoing desire to condense KPIs into a straightforward, actionable score.
A company’s NPS can be worked out by asking a single question to customers, usually delivered via a survey: “on a scale from 0-10, how likely is it that you would recommend Brand X to friends and colleagues?”
Respondents can then be split into different categories. Those who provide a score of 0-6 are considered detractors, 7s and 8s are passives, while 9s and 10s can be understood as active promoters. The NPS score itself can then be calculated as follows: NPS = (promoters-detractors)/all respondents) X 100, with scores ranging from -100 to +100. The result is a single figure which, according to Fortune, is used by 60% of the Fortune 1000 to predict customer behaviour and, by extension, the prospects of a given company.
As appealing as a single figure is, however, the NPS has its flaws – in fact, according to one study from The TQM Journal, NPS has been found “to be a very poor predictor of customer loyalty and customer satisfaction.”
This is, perhaps, unsurprising – after all, not all survey responses are truthful, nor are they always a correct recollection of what really happened. And, in a digital world brimming with customer opinions, as brand Twitter accounts field complaints and Retweet praise, it is likely that better results are waiting to be found ‘in the wild’, where no one is asking any questions.
The goal, then, should be to combine the admirably concise reporting and sometimes predictive power of a single score-based figure with all the nuance provided by today’s vast quantities of available information.
The customer of 2021: how has COVID-19 changed consumer expectations?
In with the new
Through the use of machine learning models, it is absolutely possible to gather detailed and unsolicited customer opinion through the analysis of social media and other online posts. Social media listening processes allow us not only to measure ‘buzz’ – or volume of online posts about a brand – but to determine the sentiment expressed towards that brand, whether it be positive, negative, or neutral.
Semantic machine learning models can go even further, in fact, by recognising purchase intent and recommendations; actual behaviour is captured by engagement ratios for likes, comments, and shares of a brand’s social media posts, and even the reach of a brand’s public relations initiatives.
It goes without saying that this kind of data is far more reliable and multi-faceted than the survey responses on which the NPS is dependent. The good news is that the information described above can indeed be collated into a single KPI.
Our data scientists, for example, have been able to choose from all available social intelligence metrics – buzz, purchase intent, net sentiment scoreTM (which is, itself, our trade marked composite metric) – and condense them into one number between 0 and 1 that we call the Social Presence Score (SPS).
The process is a little more complex than the formula for calculating NPS – it involves annotating the data and weighting these metrics with our own in-house method – but there are several business advantages to exploring alternatives to NPS.
Using systems like SPS, companies will gain the ability to benchmark their brand and its overall performances on the market against competitors (and, indeed, against their own performance in previous months or years), identify specific metrics that require improvement, and predict future performance – including sales.
Not only brands but also individuals are marketing in a world that is too complex to be measured with surveys alone, and it is therefore more important than ever to embrace KPIs based on sources such as social intelligence – alongside the exciting technology capable of turning those KPIs into actionable insights.