Over the last decade or so, I've regularly written whitepapers and given presentations explaining some of the key concepts behind BICCs and how to go about setting one up. And studies show they are indeed an important part of analytics best practice: organizations with a BICC see big benefits, including increased use of analytics, increased user satisfaction, and lower costs.
The Analytics Winds Are Changing
However, there are some big changes happening in the world of analytics and even long-running, successful BICCs have been struggling to adapt to some big new trends:
- Information has become a profit center. Analytics is becoming an increasingly important part of customer-facing processes, so business people want a greater say in how analytics are provisioned.
- Expectations are higher. Analytics maturity has steadily improved and tolerance for slow data processes has declined. Today, 31% of business people have to wait days or weeks for an average BI request.
- Most data isn't in the system. On average, 45% of the data business people use comes from outside of enterprise BI environments.
- Shadow analytics are more viable. There has been a sharp growth in "shadow BI" as self-service data discovery tools, cloud analytics platforms, and open-source platforms have make it easier than ever for business people to do meaningful analytics without the help of central IT.
The result is that today, 40% of business people report using an equal amount of homegrown BI applications, instead of those provided by central IT.
Since many BICCs were created to support "top down" BI development, they lack the framework to support and govern these types of "bottom-up" initiatives. They are struggling to find a workable balance between business user empowerment and governance with self-service data discovery, and can end up with a "worst of both worlds" strategy that ends up delivering neither.
Faced with decreasing business confidence – studies show that BICCs are now the function considered least likely to be driving BI around the world – BI leaders are beginning to explore more self-service analytics as an alternative or augmentation to traditional platforms, but are worried by the consequences of surrendering centralized control.
One of the ironies of analytics today is that while the technology is more powerful and easier to use than ever, the fast-changing organizational and technology landscapes have lead to increased discontent and confusion for IT and business users alike.
The challenge is to combine the benefits of self-service: more flexibility & speed, greater adoption, and better alignment with business needs; but avoid the potential problems, including needless duplication of data and tools, lack of overall visibility of the business, and diluted responsibility.
After discussions with several organizations going through this transition, I've compiled a shortlist of top tips for BICCs looking to adapt to the new world of self-service analytics:
1. It's About The Community
Building a community that brings people together across traditional silos is an essential part of any strategy to introduce a more flexible, self-service approach.
There has been a fundamental shift in the information power balance inside organizations. As technology budgets have been increasingly decentralized, IT no longer has a veto over business use of data.
Instead of a traditional scenario in which Business and IT play separate, provider-versus-user roles, everybody has to combine efforts to jointly explore and learn. This requires compromises on both sides. An active analytics community — with regular, face-to-face meetings between IT, analysts, and business leaders — is the only way to build the trust and working relationships required to make the right tradeoffs.
Successful transitions typically involve a large kickoff meeting with all the "information stakeholders," designed to break down the existing silos and discuss new organizational structures and processes going forward.
This is then augmented with social tools to build a virtual community that shares best practices online. Regular meetings are organized where teams can present their solutions and get feedback, and outside experts are invited to talk about best practices at other companies and in other industries. (I've spoken at several of these types of events — let me know if you need a speaker to get people thinking about the new analytics and big data possibilities).
With the decline in more traditional veto powers, this community helps IT leverage "soft power" to ensure that the organization is making the best use of information. "Social norms" can help encourage and support analytics experimentation while ensuring compliance with guidelines that help the community as a whole.
2. Agile Methods
Traditional BI is scalable but not agile while shadow BI is agile but not scalable. Agile BI is "an approach that combines processes, methodologies, organizational structure, tools, and technologies that enable strategic, tactical, and operational decision-makers to be more flexible and more responsive to the fast pace of customer, business, and regulatory requirements changes."
First and foremost, business-driven agile enterprise BI is about flexible organizational structures, and a willingness to embrace BI as an ever-changing, iterative environment. Agile methods are widely used in software development, and you can read more about these ideas at the Agile Manifesto site, in analyst reports, and books on agile BI.
3. Reorganize and Refocus
The need to support agile BI and align closer to the business typically requires organizational change. This often involves moving centralized IT staff and roles closer to the business units, and co-opting local experts to build a looser, more federated BICC team. This helps bring back "shadow BI" under the umbrella of the BICC while retaining the local links to business leaders that are essential for alignment.
The focus of analytics moves away from "reporting" towards "data exploration" tools that give business users more flexibility to find the information they need themselves. More of the responsibility and work is delegated to local teams, who use more iterative processes based on quick prototypes and face-to-face discussions rather than formal requests.
The BICC must move to an "agency" approach in recognition that the business units will only work with them if they are the best option, not because they "have to." Ensuring this may require investment in new tools and training, and leveraging the team's unique knowledge of cross-functional opportunities that would not be available through outside providers.
The BICC must have a service-oriented approach, but be aware that "customer is not always right" and help the business "get what they need rather than what they ask for." The role of the BIC changes from that of "gatekeeper" to "air traffic controller" overseeing the smooth running of the different analytics initiatives.
4. Offer Key Services
The "agency" services offered by the BICC should include:
Innovation Services. The BICC must be responsible for helping find solutions to business problems rather than being responsible for a technology infrastructure. If the best way to answer to a business problem is a one-off, manual mash-up of data in Excel that doesn't touch the central data warehouse, it should still be part of the BICC's scope to help organize and optimize that process. The BICC should help run innovation workshops (e.g. Design Thinking) to help uncover new opportunities and understand the underlying analytics needs of the organization.
Training. The BICC must offer regular training on tools, data, and best-practice techniques – not just when new BI products are being rolled out internally. And the BICC team may require new training, such as learning about how new technologies such as in-memory or Hadoop are disrupting existing best practice, or how to coach local teams on best-practice visualization techniques.
Data Bureau. The BICC should provide a one-stop shop for data, internal or external, structured (transactions) or 'unstructured' (text, social, documents, etc.). This requires a deep cross-functional knowledge about what data is being used across the company, not just what is stored in central IT systems. It also requires "data wrangling" skills to bring disparate data sources together for prototypes or one-off analytic needs.
Tools Bureau. The BICC helps provide expert recommendations of which technologies to use. Compared to earlier BICCs, there is typically a longer list of supported (or tolerated) tools, including more advanced analytics and data-mining. The use of standards is still strongly encouraged, to provide as much scope for cross-functional compatibility and expertise sharing as possible.
Sandbox Environments. The BICC should provide sandbox environments that let businesses create their own large-scale data stores. The use of a centralized infrastructure helps lower costs and lets the BICC more easily spot overlaps and opportunities (see "support the BI lifecycle" below).
Marketing and Community Building. The BICC is responsible for communicating BI success and encouraging active participation in the BI community. This may require skills that are not currently part of technology-focused BICC teams.
Support for BI Collaboration. Although not well-developed today, there is tremendous scope for using social collaboration tools within the lifecycle of BI provision itself. Currently the process of proposing new opportunities and iteratively getting to the right solution is typically very manual, carried out through email and phone calls. Modern collaboration platforms allow such projects to be carried out in a more effective way.
Analytics on Analytics. The provision of analytics is an increasingly important business process inside organizations. The BICC should be responsible for using analytics on the analytics process itself, providing information to business leaders about the use of analytics, and helping optimize the system as a whole.
4. Provide Training Wheels
It's a common mistake to assume that all users inherently possess the skills required to correctly integrate and analyze data — not every business organization is ready or willing to be empowered. It's important to ensure that self-service tools are not made prematurely available to business users and used without an awareness of the potential risks. The BICC should put in place some system of training courses and certification levels that let users progress to greater functionality and power at their own pace.
In additional, a system should be in place to ensure as much as possible that the local analytics that are created and used for business decision-making can be validated to ensure that they are not using data incorrectly or jumping to the wrong conclusions. This is an area where "community policing," through online sharing and review, can help.
5. Support the BI Lifecycle
The BICC has the role of keeping track of the different analysis projects going on around the company, and helping identify and areas where there are overlaps or redundancies. Over time, new data sources and analyses need to be shared more widely and become a standard part of the information infrastructure. The BICC should put in place systems that allow analyses to be "promoted" to departmental or enterprise level.
Full Speed Ahead with Agile Analytics Organizations
Analytics is one of the hottest trends in technology, but the old organizational structures for supporting analytics are not longer adapted to current needs.
New self-service analytics approaches provide an opportunity to dramatically improve analytics business agility and IT/business alignment. But this requires new cultures and new ways of working, and both business and IT organizations have to make compromises. Communications and community-building are an essential part of making the transition successfully.
Are you in charge of a BICC? Give me you comments and let me know what other best practices I should be sharing!