Customer experience is now a core strategic priority for most major organisations. NPS (and EMPS, as employee experience is also increasingly monitored), along with traditional operational metrics, such as revenue and employee numbers, takes a prominent place in investor reports (often page 1) as the role of experience in overall business performance is better understood. Search for customer experience-related terms reached an all-time high at the end of 2017¹ from the perspectives of both businesses and consumers. KAE, in a recent study, was able to calculate that almost a third of a customer’s decision to stay with their bank can be attributed to the impression consumers have of their experiences. It is therefore no surprise that as we reach budget season, CX is at the forefront of investment plans for many large companies.
The role of measurement and analytics in driving improved experiences is widely understood. However, in a recent survey, more than 75% of companies were still identified as being unable to locate specific problems that impact CX², and nearly as many are unable to track journeys that span multiple channels. It is clear there is some disparity between the hype around CX analytics and the widespread ability to harness data to build better experiences.
So what exactly is missing and what steps can companies take to build a comprehensive, accurate and genuinely actionable data set that supports the directives that boards are issuing to improve experience?
This is the first post of a blog series that will explore how professionals interested in CX can leverage their existing KPIs, make better use of data already available and consider how they can simultaneously achieve breadth, (simultaneously monitoring across channels, journeys and touchpoints) and depth, (exploring the impact of individual – or groups of – painpoints). By breaking experiences down into their most basic components and analysing all of them at the same time, KAE is taking the mass of information available in 2018 and turning it into meaningful insight.
A crude metaphor involves thinking of CX management like a watch. At its most basic, our core CX KPIs can tell us the time in the same way we can see how the hands on a watch move. Likewise, the clock face itself provides context for the hand positions just as we can identify the key drivers of CX KPIs by looking across touchpoints. However, if we are to own a genuinely accurate timepiece, we need to look beyond the clock face and understand the cogs and gears beneath.
Building from NPS and CSAT
Common KPIs such as Net Promoter Score (NPS) and Customer Satisfaction (CSAT) are, and should continue to be, the start and end of all CX monitoring programs. Investors now expect both reporting and targets for this kind of metric in standard business performance reporting.
Measuring and setting performance targets for NPS/CSAT is almost ubiquitous, but there is scope to use the results more broadly than simply as a performance tracker. As well as looking at how satisfied your customers are over time, we can understand how satisfied your customers are when they interact with other companies (incl. your competitors). Benchmarking is already common practice within and across industries for other KPIs, why should CX be any different?
Looking at the clock face, not just the hands
Sticking just with core CX KPIs, we can start to learn a lot about our business by measuring these KPIs across different parts of a business. Investigating how satisfaction varies across touchpoints allows the identification of problem areas within the customer journey, going beyond overall satisfaction. Going back to the watch metaphor, this is the clock face that provides greater context to the position of the hands.
Figure 1 – Satisfaction scores for different ’touchpoints’ for US retail banks. Touchpoints in this sense are groups of activities that stimulate contact between organisations and consumers, as opposed to channels – such as phone or email
Using data collected from KAE’s ‘CX in US Retail Banking’³ study, in which we analysed levels of satisfaction across multiple customer journeys, we can see the highest dissatisfaction is experienced when completing the application or while scheduling an appointment.
However, low satisfaction at a touchpoint doesn’t necessarily mean that it should be immediately targeted for improvement. Each touchpoint within a customer journey has a different level of importance, meaning it influences overall satisfaction to a different extent.
Overlaid with measures of satisfaction, importance becomes a very powerful metric for prioritising where we should focus our attention.
Figure 2 – Combined satisfaction and importance measures for ‘touchpoints’ in US retail banks. Touchpoints in this sense are groups of activities that stimulate contact between organisations and consumers, as opposed to channels – such as phone or email
Making Sense of the Data
Using key driver analysis allows us to see where things are and aren’t working at a structural level. Looking at the above scenario for a US retail bank, satisfaction in Customer Services is well below the average and its impact on overall satisfaction is high so this should be our highest priority touchpoint to identify reasons for under-performance.
Key driver analysis is not a new concept. In fact, it is widely used to break down the importance of components within complex systems. However, we believe the adoption and effectiveness of this technique has been inhibited by its need to be repeated each time a new subset of the sample is required and the number of variables you can include for consideration. For example, if you want to understand importance by customer segment, this would require a new run of the model for each segment. Whilst, on the surface, this should not be a big deal, there are serious implications for both the costs of running such a technique, and the timeliness of the insights it can develop.
Machine learning allows us to overcome this problem, modelling respondent’s satisfaction and the importance they place on different parts of their experience individually. This allows real-time segmentation of users without the need to re-run the model and is a potential game-changer for the ability to look deeper and in near-real time at granular experience data.
Looking deeper still
As part one in this series, we have reviewed the ever-growing importance of experience measurement for businesses and shown that machine learning may be applied to greatly improve established, but functionally limited techniques for placing the headline measures of NPS and CSAT in greater context.
In the next post we’ll look deeper and look at both why customers are satisfied, what’s causing the satisfaction and how much it is contributing. We expand on the machine learning methodology and show how we can conduct key driver analysis looking at specific events underneath each touchpoint and showing how they contribute to or detract from satisfaction. This layer of detail holds the key to understanding exactly what is causing customer experience issues, enabling a change from a scattergun approach to a data-driven strategy. In the third article we will look at how the numerical techniques inspired by the mutations and evolution of DNA can be used to combine remedial action and develop a coherent and optimised change programme.
- Results to this study are available, please contact CX@kae.com