Customer Satisfaction 2.0
Three steps to achieving customer delight with automated predictive analytics.
As President of Mercedes-Benz, Steve Cannon, famously said, “Customer Experience is the new marketing”. Many would not disagree.
Whether you are using Net Promoter Score (NPS) or Customer Effort Score (CES), almost all businesses have customer satisfaction KPIs in place. But how do you know that these measures are useful predictors of behaviour, and more importantly, that they support sales and positive advocacy?
Welcome to CSat 2.0.
By introducing predictive analytics, and in particular automated predictive analytics, you can turn this concept on its head. You can now ask ‘What are the outcomes I am actually looking for?’, and then mine the data to generate the factors which drive those outcomes.
This means that a dynamic propensity model can be achieved for ‘nature’ (the attributes of the customers which predict greater sales and advocacy) and ‘nurture’ (how their customer experience affects these propensities).
In this way, practical indicators can be provided to sales and marketing teams to target the appropriate customers, and prioritise corrective or preventative action. This also has the secondary benefit of optimizing the cost of customer services, as a threshold of spend can be decided based on the desired service levels.
The traditional way of setting all this up would be to bring together predictive analytics software, a bunch of smart data scientists with time on their hands, and of course a hefty budget.
But as this isn’t possible for most businesses, this is where automated predictive analytics comes in. This is a new breed of predictive analytics which does not require pre-defining terms or feature extraction (i.e. the deciding and defining of the variables to use in the analysis). It can be of great assistance in this step, as well as automating the subsequent steps to get to a usable answer quickly, in a matter of minutes rather than weeks or even months (data scientists spend around 80% of their time manually transforming, cleansing and preparing data).
Three steps to success with automated predictive analytics
Step 1: Simplify
The first step is to cluster the Voice of Customer (“VoC”) data from social media, reviews, weblogs, loyalty data, CRM, POS and surveys. Similarly, customers can be clustered by attributes and behaviours. This provides value in itself in terms of insight, and will really help to provide the foundations for further analysis.
Step 2: Predict
Once the data is simplified and aggregated, there are a variety of statistical and machine learning techniques that can take the results and predict what will happen given the mix of events and segments.
Step 3: Recommend
The final step is to recommend prescriptive analytics which present insights which are actionable. This can take various forms such as a traffic light indicator. For example, if the analysis shows a customer who has complained more than twice is 50% more likely to leave, a set of business recommendations can be created to help avoid this, like a pre-emptive outbound call.
Understanding the interdependencies of customer services and sales can bring huge benefit. Predictive analytics, and in particular automated predictive analytics can assist marketers and customer services managers to do the right things at the right time for the right reasons. Welcome to CSat 2.0.
About Warwick Analytics
Warwick Analytics is a spin-out from The University of Warwick with proprietary algorithms for automating predictive analytics, even with heterogeneous data. It will be working with Capital One’s Growth Labs on the Capital One accelerator to refine its technology towards this sector and use cases and the teams are excited to develop cutting-edge applications.
The opinions expressed in this article are the author's own and do not reflect the view of Capital One.