Getting the most out of your data: practical steps and case studies

by Paula Miranda

Data has become all the rage. Just about every organisation is investing in analytics, but few are deriving tangible business outcomes. By 2022, Gartner predicts that a staggering $104.1 billion will be spent worldwide on software alone. While companies who execute well may expect a 40-60% uplift in revenue, around 60% of analytics projects fail, and a similar proportion of organisations currently have a low level of data maturity.

This underwhelming value capture from customer analytics typically stems from one of three factors.

  • There may be poor understanding or even scepticism around the value of analytics investments amongst senior executives and the frontline. To counter these negative mindsets, you must start with clear business use cases and aim for a pragmatic, believable level of maturity.
  • There are often opaque and expensive technology contracts, designed to benefit the service provider more than the client. To avoid this, it is best to start with a minimum technology investment (ideally centred around an open PaaS architecture solution) so that you can demonstrate value early, and earn the right to do more.
  • Finally, hyped-up analytics projects often never leave the laboratory, and end up as theoretical one-off exercise, with limited operationalisation. Investing in the right human capabilities and operating model ahead of, and in parallel with technology, can avoid this pitfall.

To maximise value from your customer data, there are three broad sets of capabilities that need to be improved:

Customer analytics – the ability to create and surface actionable insights from structured and unstructured data

Data architecture and integration – a single coherent, consistent, and easy to access enterprise view of data collated from disparate sources

Data quality and governance – access to high-quality accurate data and the operating model to be able to make improvements along the way

While organisations often purely focus on one of these dimensions at the expense of others (usually analytics capability), effort and investment are required across all three areas to increase your chances of success.
 
Three Modes of Execution

Blackdot_960x1290_Operationalising-customer-data-3.jpg

Between Servian and Blackdot, we have seen numerous efforts to uplift the customer experience based on data-driven insights, within multiple industries and domains – each with varying degrees of success. What often distinguishes those that achieve real outcomes, is coherent endeavours to improve capability across a number of dimensions, to a level that is sustainable with BAU activities. Whilst it can be tempting to jump to greatness in one big bang, the business value is maximised by prioritising logical groups of capabilities to a pragmatic-level, based on current maturity. Below we dive into three practical examples:
 
1. For organisations with a low-level of data maturity, the greatest immediate value comes from having single view of the customer, along with quality proprietary data.
 
A Superannuation fund created a set of common data models to automatically consolidate, cleanse and unify multiple data sources, to provide a complete and rich view of customer and marketing performance. This organisation suffered from overloaded marketers, spending a lot of time manually creating one-size-fits-all campaigns to service their customers. Additionally, the company lacked the capability to measure and act on campaign performance.
 
They advanced their customer engagement and reporting capability, through a set of models that consolidated their proprietary data and integrated it into tools that enabled a single view of customer - measuring marketing performance against targets. The transformation included data cleaning and migration to a cloud warehouse, as well as setting the organisation up with the governance to ensure the maintenance of the data.
 
Consolidation enabled the company to work from rich, data-driven customer insights and improve marketing effectiveness. Overall, they experienced an increase in speed to market, as they had the  information needed to base automated targeting decisioning models on – freeing marketing’s time for innovation on strategic priorities.
 
2. Once the base data quality is remediated, there is significant value in building future-fit data architecture, and integrating further data sources to inform the development of predictive models.
 
An FMCG company augmented their proprietary dataset with additional data points, to create segmentation models that predicted the propensity to take up tailored offers. The company’s legacy segmentation was based on an analysis of simplistic data from surveys, and was not scalable to other regions. Consequently, digital transformation was required to augment their existing dataset by acquiring further customer data, to enable more sophisticated and scalable segmentation.
 
This organisation assessed their customer journey to identify opportunities for more sustainable data acquisition, and reconfigured digital assets where possible to collect insights relevant to segmentation. The uplift in richness of the available data, set the stage for a more sophisticated segmentation model – using machine learning to predict the appropriate customer segment, and the importance of data attributes in the segmentation process.
 
The evolved segmentation strategy was able to make predictions about additional customers outside of the legacy model, and also enabled more effective marketing through personalisation.
 
3. Organisations who have overall data maturity, can then create real-time predictive models to maximise campaign ROI and customer value.
 
A Sports Entertainment company wanted to create an immersive experience for subscribers engaged in their platform, transforming their app into the ultimate game guide. To achieve this goal, they needed to improve the user experience of their AI assistance, to innovate the customer experience and ultimately uplift advocacy.
 
Enhancement of the AI assistant included migration to a cloud, serverless infrastructure, and leveraged deep learning algorithms to generate predictions relevant to their customers. The 65 features which existed within the AI assistant were rebuilt and expanded to 86 – improving the performance of predictions, and allowed coverage of more game scenarios.
 
The revamped AI assistant delivered an improved customer experience and offered compelling options for customer engagement. This saw an 150% increase in subscribers – with consistent marketing spend, and 140% increase in app engagement compared to others in the category.
 
How To Get Started
 
This might all feel a bit too complicated, requiring navigation of multiple decision-making dimensions. However, the next best action (if you can excuse the misuse of the term) for most organisations, is fairly well-defined depending on the business context and current state of data maturity. We have summarised this into a simple 7 step recipe:
 
  1. Set the future state ambition and identify data-driven use cases
  2. Define the desired customer experience for each use case
  3. Design the high-level target data and technology architecture
  4. Understand current capability maturity relative to the ambition
  5. Prioritise and sequence capability gaps into an actionable roadmap
  6. Make platform choices and capability sourcing decisions as required
  7. Embark on agile test and learn execution to frontload value and move to the desired future state

Co-authored by Paula Miranda, Senior Manager – Blackdot and Antonio Pugliano, Associate Partner – Servian

Register for our upcoming webinar for a practical walk through of the insights explored above.