Menu

CEB Blogs

Topics

Corporate IT

Managing the Growth of Business-Led Analytics

More and more parts of a big company are using data and analysis as a core part of their every day jobs, and this has meant an explosion in the number of teams outside IT who are now dedicated to data

As data and the analysis of it become more and more important to managers in all parts of a company, and as firms continue to digitalize more and more of their operations, analytics projects are sprouting all over most large companies.

The IT function has traditionally been responsible for business intelligence – which is usually a fancy name that denotes the collection, storage, and analysis of data – and reporting now have to work with functional analytics teams across the whole organization. And CIOs have to decide if these business-led analytics teams are the way forward, or another type of “shadow IT” that needs to be reined in.

In the past, analytics was often centralized in IT close to the systems that created, stored and transferred data. But as the demand for “insight” skyrockets and as analytical acumen is employed all over the company, IT may no longer have the necessary resources or business knowledge to maintain this arrangement.

IT generally still retains staff who own foundational data management and governance activities, but many analytics efforts take place in other areas of the business. Analytics teams typically organize data and run analyses, but do not perform traditional IT activities like building or running software and systems.

Three Models

Through a number of recent conversations with CIOs in CEB’s networks, there are three common organizational models for analytics that provide an alternative to centralizing it all in IT.

These models are not mutually exclusive and, regardless of which a company chooses, IT functions considering how to respond to the growth of analytics teams outside IT should realize that business-led analytics is here to stay and that they should adapt their engagement model to support it.

  1. Centralized outside IT under a chief data officer:

    Benefits: Although similar in structure to the traditional model, analytics efforts are centralized in a team that is separate from the corporate IT function. The head of this group is often the chief data officer (CDO) and reports directly to the COO, CAO, or CEO. This not only signals the importance of information and analytics to company strategy, it provides specialization for areas that are not core IT competencies such as data science and predictive modeling.

    Risks or limitations: This set up could create even more coordination challenges, specifically between IT and the CDO or centralized analytics team, adding an extra layer of handoffs.

  2. The “hub and spoke” model:

    Benefits: In this hybrid approach, IT acts as the hub in which some analytics initiatives are centralized and functional business units serve as the spokes. Enabling these teams to perform their own analytics, can make IT and others more efficient. There is often collaboration and knowledge sharing between BU analytics teams which helps to break down functional silos.

    Risks or limitations: Although IT plays the role of advisor or coach in this new model, it must be careful to not over control or corral decentralized analytics. Dictating which tools teams can use and adopting an inflexible posture will only lead to greater resistance and duplicative effort.

  3. Use an external service provider:

    Benefits: By farming out analytics to third-parties, companies can capitalize on the technical expertise of external providers and quickly improve their own expertise at analytics modeling. It also helps companies to integrate internal data with external information sources that would be otherwise unavailable. For example, a telecommunications company in CEB’s networks uses a third party who marries social media data from Facebook with its existing customer information to make predictions, such as when a customer may be planning to switch services.

    Risks or limitations: Exporting analytics leads to the risk of domain expertise loss. And, depending on the current structure of a company’s data, it may not be possible to work with third-parties for analytics (e.g. operating and storing data in the cloud facilitates this partnership more easily than using on-premise infrastructure).

 

More On…

Leave a Reply

*

 

Recommended For You

IT’s Tricky Relationship with Business Analytics Teams (and How to Fix It)

As data becomes increasingly important to companies, business analytics teams are popping-up across the org...

Close