A pragmatic approach to Big Data and analytics
To use the Retail industry as an example of the world of Big Data, there are a growing number of digital touchpoints that enable the collection of shopper data.
These include mobile devices, web interactions, retailer-supplied apps, in-store data collection (footfall, P.O.S, staffing, customer tracking, queue management, loyalty card etc), CRM system, social media etc. Consequently, retailers have access to more mission critical data than ever before and this is growing exponentially. If retailers could fully utilise this data they would be able to,
- Drive sales
- Increase profits
- Optimise performance
- Control costs
- Boost customer loyalty
BUT, too often companies struggle to convert these large volumes of data into actionable insight.
Most organisations are capable of analysing historical data but the real leap is to shift from the historical perspective alone and instead to use it as a building block for a more predictive approach.
In other words, this involves the development of an analytics roadmap which in a retail context might help to predict market trends and shopper behaviour and enable retailers to adapt their strategies accordingly.
What does an analytics roadmap look like?
Descriptive analytics is commonly in use in most organisations. It uses traditional Business Intelligence capabilities, bringing together data from various sources to report on business performance through reports, dashboards and visualisations. It can show trends and patterns hidden in your historical and current data and alert business users when certain outcomes occur. In other words, you know what has happened but you need to look to diagnostic analytics to dig deeper and find out why.
Diagnostic analytics uses advanced Business Intelligence capabilities, data discovery, exploration and visualisations to look at correlations and relationships within your data to identify what drives business results. Insights gained during diagnostic analytics can provide a deeper understanding of why a business performs as it does, and help decision makers affect performance. This knowledge is vital in the development of predictive models.
Predictive analytics helps you to make better informed decisions based on a combination of what has happened and what is most likely to happen in the future. It uses predictive models, statistical analysis, data mining, real-time scoring and a range of algorithms and techniques, all contained within the top Business Intelligence and Data Management toolkits.
Prescriptive analytics builds on predictive analytics to optimise decision making by evaluating a number of actions and their likely outcomes through techniques including what-if simulations, rules and decision logic.
So why does this article talk about a pragmatic approach to Big Data and Analytics? It is simply to offer advice to any organisation embarking on an analytics journey, and it is – “Beware high-budget projects that take long time periods poring over Big Data yet deliver little to benefit the business”. Remember that “Right Data” always trumps “Big Data”. Focus on how you will use analytics to achieve critical business outcomes, preferably without impacting on the business whilst doing so.
Here is a useful fact to remember. 2% of any Big Data source is relevant to any particular business outcome so the trick is to find the 2% you need and apply it, solve the problem then move on to the next business outcome (and it’s 2% of data) and so on. Make Big Data work for you or you will drown in the Big Data Ocean.
To discuss your data and analytics requirements please contact Paul Cooper at firstname.lastname@example.org