NEW: The Hackett Group Launches AI XPLR: A Pioneering Gen AI Readiness and Opportunity Assessment

Enterprise Analytics: Decentralized, Centralized, Decentralized?

By Nilly Essaides
October 23, 2018

When viewed from 30,000 feet, our research and ongoing conversations with clients reveal a vacillating trend in how companies deliver analytics to the enterprise. After a period of decentralization, more companies are consolidating their analytics and decision-support activities in centers of excellence (COEs). Yet the most advanced organizations are once again decentralizing the capability through new data management and self-service analytics tools.

Phase 1: Decentralization

Before the advent of digital transformation and the reliance on sophisticated solutions to enhance process-performance insight and future-looking decision support, many companies maintained a decentralized analytics model. FP&A did financial analysis. The business units and functions did their own. Different parts of the organization used different tools (often Excel) and relied on their own data to understand their performance and the needs of their customers. Analytics models proliferated, and there was zero coordination among the different groups at finance and the business level.

Phase 2: Centralization

In today’s environment, we see more companies consolidating their analytics activities in centers of excellence (COEs). According to The Hackett Group research, 77% of finance executives report that they either already have operational analytics/decision-support COEs or expect to have them in the next two to three years.


This expectation is supported by The Hackett Group’s 2018 Key Issues Study. Finance executives expect future analytics activities at the functional level to be nearly half what they are currently, with commensurate increases in the use of either finance-owned or independent COEs.


The trend toward greater centralization is a product of several factors.

  • Digital transformation has introduced a plethora of sophisticated analytics solutions that make economic sense to adopt in a single location run by experts.
  • There’s also a current scarcity of finance talent with analytics acumen. According to our 2018 Digital Skills Poll, the greatest gap between existing and required finance-talent capabilities is analytics acumen. It’s also the number-one training initiative for finance today. It will be a while before most organizations evolve into the third phase.

Phase 3: Decentralization 2.0

Ironically, there are already signs that going forward, companies will democratize analytics and push a lot of the work back to the functions and business units. One reason is that demand for advanced analyses and reporting is growing too rapidly for a single “provider” to keep pace without creating bottlenecks in the delivery of essential insight. Another is that the business units and functions are closer to the market and know best what kinds of analyses they need to drive better performance.

Again, technology has a critical role in enabling this “switch back.” New self-service analytics tools are growing in sophistication as well as user-friendliness. So analytics consumers get answers faster. Our research reveals that there’s already an 18% gap between enterprise performance management (EPM) top performers and peers in the use of self-service capabilities.

Two other important factors determine just how much analytics can be distributed to the users. First, the company must maintain a single source of the truth, so everyone accesses the same data. Second, there’s currently not enough analytics expertise in the field to optimize the use of the tools and benefits.

Choosing the right model

The big question for CFOs and other senior leaders is how to decide which activities should be handled by the centralized entity and which should be retained by the analytics hub. To determine the split, finance executives should ask the following:

  1. Does the organization deploy advanced data-management platforms, like data lakes and data marts, to host varied data? Does it have a robust data-management and governance infrastructure? To successfully democratize analytics, a company must have a single source of approved data that everyone can access and use.
  2. Does the organization have the necessary widespread skills to use self-service technology?
  3. Finally, does the COE have enough capacity to handle standardized and specialized analytics queries?

For now, finance and other senior leaders are exploring different approaches, and there’s not enough data to determine which one is “best.” Some organizations will choose to continue to centralize standardized analytics and reporting while empowering business users to run ad hoc analyses. Others do the opposite: They let the business handle standardized activities while retaining sophisticated, high-level queries in the COE. Then there are those that maintain only data and system governance and a coordinating function at a centralized entity.

Taking the lead

We expect the centralized model to remain the pervasive approach at least for the near term. One reason is that most organizations have yet to develop enterprise-wide analytics skills.

Another important reason to maintain a centralized model is that most organizations are still in the exploration phase of deploying advanced analytics. While we expect an eight-fold increase in mainstream use, current adoption is low. Often, advanced analytics projects are launched on a discrete basis to respond to specific business needs. It’s best to have a coordinating entity to govern and assess these pilots and help evangelize the value and ROI in order to scale up.

Because analytics excellence often resides within the finance group, it’s incumbent upon senior finance leaders to build credibility for the new technologies and spearhead a more cohesive and expansive enterprise analytics delivery model.

For more on digital transformation in finance, see How CFOs Can Accelerate Group Transformation Programs With A New Finance Platform.