Customer Analytics – An interview with David Slansky, Advisory Partner KPMG
In this era of big data, there are increasing opportunities for collection, processing and analysis of vast and complex data sets about customers such as demography, behavior, transaction trends, economics etc. The ability to mine and utilize insights from analyzing these huge and complex data to create business value is generating significant interest to business leaders globally. In this question and answers session with David Slansky, we review customer analytics, its benefits to businesses, and how organisations can begin to leverage data from internal and external sources for better customer insights and business value.
What is customer analytics?
Customer analytics helps you to understand your customers’ (people or companies) behaviour and drivers of this behaviour in order to improve your relationship with the customers to ultimately gain market share and drive revenue. The understanding is derived through continuous collation and analysis of customer information as well as environment information to generate insights into customers’ past, current and future behaviour. We see rapid shift from generic customer segmentation into more individual approach when you treat each of the customers separately in the model output. This enables you to treat each and every customer as an individual and deliver exceptional customer service and experience.
Why should customer analytics matter to today’s business organisations?
The typical effects of properly designed, implemented and embedded customer analytics are customer satisfaction improvement, increased number of customers and/or their profitability and decreased customer attrition. On the other hand it also helps with decreased cases of misunderstanding and fraud.
Apart from these direct effects there are also other benefits, such as gaining competitive advantage (through the customers’ empathy and behaviour prediction), improving brand reputation but also help with new offerings development.
What are the ways in which organisations are using/ can use customer analytics? And what business benefits are being derived?
Customer analytics can be used in any phase of customer lifecycle. Customer life-cycle is the key differentiator of customer needs. Customer analytics can identify in which phase of the cycle is each customers and suggest appropriate next-best-action for each customer for example during the
- Acquisition phase: Targeting groups of customers for acquisition or dormant customers activation including specification of the communication channel and strategy
- The Retention phase: Early identification of moments leading to dissatisfaction and attrition and implementation of appropriate reaction for each of the moments.
There is endless number of specific benefits each market sector can get from customer analytics – not only for retail customers, but also for corporate/wholesale segment. Just to name a few:
- Financial services – automated generation of customer lead lists for each sales agent, where each customer has specific needs and reasons for servicing / cross-sell / up-sell. Increase in sales revenue and improvements in customer relationship management (no. of calls, meetings) and experience through personalized product offerings.
- Energy –forecasting of electricity energy supply and energy demand across different geographical areas for optimization of distribution and grid balance. Decrease in forecasting errors, resulting in yearly savings and reduction in transmission losses.
- Retail/ Consumer Markets – knowledge of customer demography vis-à-vis purchase patterns and trends can be utilized for planning and executing more relevant and targeted customer communications and promotions as well as improvement of supply chain effectiveness.
Typically organisations only have limited data and even where this exist it’s usually in different systems and requires cleansing to improve the integrity. Are there any experiences of how to start to derive value from data analytics while the data issues are being addressed?
There is a notion that you need best-in-class BI and best data quality before embarking on customer analytics. It is not true. Customer analytics should be driven by business challenges and the earlier you start the better, even if you have data silos with suspicious data quality to begin with. Piloting customer analytics activities (e.g. segmentation, activation, cross-sell / up-sell / retention) will not only generate business value, but also help identify, which data sources / attributes are: i) predictive of customer behaviour, ii) predictive if better data quality , iii) missing, and iv) unusable / useless. Based on such insights you can make better decisions regarding your Information Technology/ Business Intelligence/ Datawarehouse strategy, focusing on areas with the greatest effect on customer analytics results within your organization. These pilots should be both business-wise and time-wise limited. Our experience is that 8 – 12 weeks pilot delivering tangible business results with up to 5 – 6 weeks of data gathering and preparation is the most successful approach.
Some good practices in data preparation for analytics are described below:
- Create a data wish-list based on goals (what you want to achieve, e.g. for supporting cross/up-sell customer behaviour data are required, such as current products, transaction, level of aggregation, frequency etc.)
- Populate all available metadata describing available internal data/sources. If little or no documentation is available, execute data discovery process specifically aimed at selecting appropriate data sources based on desired analytics objectives:
- Optionally, a data gap analysis of required internal data and available data which will help the decision on whether there is sufficient data to carry on or not.
Always bear in mind that the customer analytics can be successful even without 100% available and high-quality data. The data quality and availability affects the precision of the predictive models (as well as the ability to contact customers in case of contact details), but the precise data models are not the only ingredient for successful achievement of your business goals. Successful business embedment (meaning integration of analytical solution outputs into business as usual operations, including education, motivation and follow-up support of business people) is what makes or breaks overall success. Getting back to data quality issues – in order to avoid missing or problematic data, find proxies – you may consider using other data sources, either within your company or outside (external data). From experience, our very successful behaviour prediction models use “signals” from more than 60% of external data.
For more information, please contact:
Ladi Asuni, email@example.com
David Slansky, firstname.lastname@example.org
Bisi Lamikanra, email@example.com