Creating an Analytics Function

4 Essential Functions to Provide

Throughout industries analytics has taken on an almost ubiquitous nature of late. There are increasing demands for data scientists, open source developers and an in-house capability for providing actionable insights through data. But, can an analytics capability be provided by a data scientist or developers alone? At Version 1, we advise customers that we partner with on how to design and build an analytics function, and the key question centres around what an analytics function should be expected to deliver.

The answer is not just models and dashboards, big data databases or propensity scores. In practice, how you structure your analytics function should be determined based on responsibilities, who should provide the major elements that an analytics function needs to achieve.

Let’s take a look at 4 functions that are essential to consider, especially in organisations where analytics can be used for a variety of applications across multiple lines of business.

Analytics Strategic Planning – Setting the Vision

This includes development of an analytical roadmap, selection and prioritisation of analytical initiatives across the organisation. Should this responsibility rest within each line of business for their own analytics capability, or should there be a central committee with a direct line to those setting the strategic vision for the organisation as a whole?

Analytics Modelling – Producing the Insights

This is the nuts and bolts of what many would consider the purpose of an analytics function. The grunt work or the clever stuff, so to speak. It’s  here where your statisticians or data scientists are essential. An important aspect to consider is the criticality of domain expertise. How close to the lines of business do the analysts need to be?

Analytics Infrastructure – Enabling the Insights

The Analytics Infrastructure should focus on providing technology and data governance including the assessment of GDPR requirements. When organisations speak about “one source” or “one version of the truth”, it is the analytics infrastructure that has responsibility for this. The analytics infrastructure should be optimised so that it is not a blocker, but an enabler for the analytics modelling function.

Analytics Operations – Realising Business Benefit

The more analytically mature an organisation is, and the greater the analytical footprint, the greater the need for common standards, change management, facilitation of collaboration and integration of results into operational front ends.
Where analytics is deployed across multiple lines of business how can you ensure that the same work is not being replicated in other parts of the organisation? How can you be sure that best practices are maintained throughout?

When you consider these 4 key areas in relation to your analytics function, where they should reside and who should have responsibility, then the question around how the analytics function is structured or designed becomes clearer. Should you opt for a hub and spoke model? Or perhaps you would achieve better results with a Centre of Excellence? Understanding the 4 key connected areas should help you produce an analytics charter that will go a long way to answering your questions and helping you to decide the best design for your situation.

On one level, it can be easy to think that is it a good idea to place the responsibility for each different functional area with separate teams, such that each are playing to their individual strengths, rather than a singular analytics function. It sounds practical. In reality though, this makes it easier to lose focus, easier to head off in different directions and become siloed and essentially more conflicts will arise. It may become more difficult to get things done. It is not a straightforward dilemma, which is precisely why it needs your attention at the outset! One size is never going to fit all. Assess the most appropriate balance given your organisation, your team and your culture. Make it work for you, however, be aware that the availability of each of the 4 functions is a key factor for success.

Tools Covered

IBM SPSS Statistics

Related Solutions


Tech Tips

Tagged As IBM SPSS Statictics For Advanced, IBM SPSS Statistics for Beginners

Need some help?

Three women working on laptops at a table for Version 1 SPSS Training

Learn how to use SPSS from the experts

With more than 20 years of delivering highly successful training programs, Version 1 offers a wide range of training options to best suit your requirements, enabling you to optimise your IBM SPSS Software, achieve your analytical goals and continually improve your results.

Related Articles

Take a look through our SPSS Articles covering a broad range of SPSS product and data analytics topics.

Arrange a free consultation to discuss your analytical needs, and identify the best solution for you.