Business Analytics & Insights
The Analytics Power Struggle: COE, Self Service…or Hybrid?
Analytics centers of excellence (COEs) began popping up in mid-market and larger organizations around a decade ago. Leveraging expensive, technically complex tools, harnessing advanced analytics skills, and coordinating cross-functional initiatives, analytics COEs helped organizations glean insights and visualizations from many sources of disparate data.
Now that contemporary self-service tools have emerged, business team members are eagerly embracing the ability to work with data themselves and in their own timeframes. Some leaders and organizational designers are questioning the scale of analytics COEs and considering how much of a role they still have in business-focused analytics development and maintenance.
It’s a complex organizational design question. In this article, we explore the advantages and disadvantages of organizing a central analytics COE, deploying a self-service analytics solution, or creating a hybrid of the two.
In the early 2000s, an analytics COE team was the gold standard of a data-driven organization. Using various visualization tools and harnessing coding and querying capabilities, team members assimilated data from digital and analog databases, then created analyses and visualizations.
More recently, self-service analytical tools in the market coupled with rising demand for information- and even prediction -driven decisions have changed the role and perceived value of analytics COEs. Business users are motivated to circumvent the logistical barriers and turnaround time of their analytics COE and create their own analytics solutions – or leverage consulting firms to do so on their behalf.
So should analytics COEs continue to exist? Or are they defunct; organizations should put data management/manipulation directly in the hands of business teams via self-service tools?
For most companies, the answer is a mix of the two. There will always be greater value in centralized ownership of certain responsibilities such as data management and governance. But restricting business teams can negatively impact the business: slowing time to solutions, missing key data-driven objectives, and restricting the depth of insights available. Let’s consider the two extremes, and where they might meet in the middle.
The benefits and pitfalls of centralized ownership of cross-functional analytics
An analytics COE with a formal strategy- and executive-sponsored responsibility for data-driven analysis can help organizations leverage data efficiently and effectively. A centralized analytics COE drives greater confidence and consistency in decision-making across business units, and ensures that siloed organizations adhere to a defined organizational data strategy.
But there are some pitfalls with a centralized approach to analytics operations. The business side can become frustrated with slow turnaround times to their solution requests. Business teams can face month-long cycles of queries, reports, and more queries to receive meaningful data to inform their business strategy, leaving operational and market opportunities in the ashes of time.
Analytics COEs can also struggle to align the organization’s strategic objectives and priorities with those of business teams. This is often due to the staffing construct of the analytics team, which tends to be made up of technical experts and data analysts. Additionally, the longer team members spend removed from the business, the more stale their business acumen becomes. A solid analytics COE should be comprised of a well-rounded and healthily cycling team of business, technical, and data experts. Working together, they can link the organizational and business units’ strategies to prioritize and deliver requests to deliver the most value with the biggest business impact.
Lack of buy-in among business teams is another risk to the analytics COE model. Reducing the impact of the aforementioned challenges will reduce this risk, but some business teams approach COEs with an inherent lack of trust and buy-in. COEs need to take this seriously and develop strategies to strengthen their relationships with business users. Just as in any cross-functional relationships, mutual trust and openness to collaboration vital to success.
Karmen Blue knows there are different ways to earn business buy-in. She sees that this buy-in is critical to her organization’s success and believes there are many levers to pull to ensure this buy-in is realized. When we sat with her, we discussed establishing an analytics council with membership from both the COE and the business. Today, the council gathers information on the reporting and tools in use, establishes processes and standards, creates a single point of access for specific analytics reports, enables the prioritization of efforts, and ensures alignment on the engagement of third parties. Possibly the most important role of the analytics council is that business teams understand and trust the COE’s prioritization of deliverables and turnaround times.
Business ownership: The benefits and pitfalls of self-service analytics
Business teams need rapid access to analytics-derived information, to apply those insights to daily, short-, and long-term decision-making. Leveraging data in real time to inform educated decisions can validate (or invalidate!) instincts, decide debates, assist with prioritization, draw a quantitative line in budget-making and project spend, and enable unbiased decisions.
But business teams running their own siloed analytics operations has its downsides. Most importantly, lack of coordination across the business reduces governance and leads to multiple versions of truth. Poor procurement processes can lead to missed cost-savings opportunities with software and data management vendor contracts. Failing to catalog cross-company analytical solutions can lead to duplicate solutions across the business. Confirmation bias can negatively influence the reports created with self-service tools: non-analytics experts typically stop their analyses once the data agrees with what they already believe to be true. Finally, long-term sustainability and maintenance of mature solutions can lead to higher costs and more effort to keep solutions running. Inherently, business teams do not place the same amount of rigor and structure around productizing solutions as COEs do.
A mixed approach
A hybrid approach to analytics data management and analysis combines the advantages of an analytics COE with business teams’ abilities to consistently and quickly make information-driven decisions. In this infographic, we step through an ideal scenario:
One takeaway from the scenario is that the analytics COE is engaged from the beginning of the process as the go-to solution adviser. The business approaches the group as a trusted resource with its best interests in mind. To make this successful, the COE needs to truly be that trusted adviser and work with business teams to engage third parties where it makes sense, instead of maintaining territorial ownership over analytics efforts across the organization. This ensures prudent use of company funds and streamlines third-party engagement, if needed.
Another takeaway is that centralizing analytic design and development fuels the analytics COE operating model to align with the organization’s needs. When the analytics COE is properly engaged, developing their team with the skill sets needed to enable the business will improve scalability. At the same time, it also enables a balance between the groups. Charlie Miller, senior manager, finance digital capabilities at the Coca-Cola Company explains that the biggest pitfall of giving a centralized COE is that they are often stronger in certain business areas and weaker in others. Allowing business groups to leverage self service tools means they can both benefit from its COE’s capabilities while augmenting its expertise, whether that’s in prioritizing, estimating, or even building out solutions.
To establish a successful partnership between the analytics COE and business teams, our recommended approach involves:
- Formalizing an intake process for business teams to submit requests without significant burden
- Adequate skillsets and resourcing of the COE to be involved early in an advisory role on key business initiatives
- Rotating COE team members through the business to ensure business acumen is maintained
- Establishing an analytics council within the COE so that efforts are adequately prioritized and prioritization is understood by the business
- Coordinated software and data procurement practices across teams
- COE involvement (in an advisory role) when business teams engage third parties
- Most importantly, trust between business teams and COE team members
Analytics COE responsibilities:
- Manage intake process and prioritize business requests
- Establish a clear operating model
- Create a balanced team with the right skill sets to drive efficiencies and capabilities
- Develop a strategic vision for the analytics COE’s role within the organization
- Create a steering committee for the COE, with seats allocated for members of the business
Business team responsibilities:
- Ensure alignment with the organizational strategic vision
- Provide requests early in the planning process
- Collaborate effectively during design discussions
- Participate in COE steering committee meetings
- Align on the definition and standardization of key metrics
- Catalog current solutions to avoid overlap
- Determine synthesis opportunities
Here’s an illustration of that model:
An effective partnership among an analytics COE and an organization’s business teams requires investment in process, relationships, and technology. The potential value in decision-making derived out of quantitative data that incorporates organizational strategy into each piece of information, however – is limitless.
Thought Logic’s Business Analytics & Insights team offers pragmatic, forward-thinking strategies and delivers proven, innovative results. We view technology and data as tools to solve business challenges and think of ourselves as being “trilingual” — understanding the business, technology, and quantitative aspects of the business analytics domain. Read more about our trilingual approach.