How to Organize for Data-Driven Government

How to Organize for Data-Driven Government

by Jake Bittner

A few weeks ago, I wrote about how the existing procurement process is holding back state, local, and federal agencies from fully realizing the results of becoming data-driven organizations. My main point was this: To become data-driven, government agencies must uncouple BI/analytics work from contracts to build transactional system contracts. Instead of building transactional systems and then plopping analytics on top of them, they must create a separate Business Intelligence functional unit that analyzes data across departments and enforces standards of analytics, so that any internal analysis is consistent, accurate, worthwhile, and valuable.

But this suggestion begs the question—how would those BI Units, often called BI Centers of Excellence (BI COE) or BI Competency Centers (BI CC), operate in public sector organizations? How can they help drive more consistent, accurate, valuable and objective analytics?

First and foremost, as it would within any commercial organization, the BI COE would be responsible for establishing standard business rules and processes for any BI/analytics programs across the organization, no matter which system, program, or process they analyze. This is the common function of a BI COE, and much has been written on best practices of BI COE’s in for-profit enterprises that are equally applicable to public sector organizations. While public sector BI COEs should certainly draw from these commercial best practices, there are some nuances that government agencies need to consider as they head down this path. How to successfully establish a high functioning BI COE within government will be a subject of a future post.

Second, the BI COE must find a way to exert control or strong influence over how the organization’s BI/Analytics funds are spent. There are several ways this can be accomplished, but if it doesn’t, it will be very difficult for a BI COE inside a public sector organization to succeed, as one-off solutions will continue to proliferate silos of information.

One way of empowering the BI COE with this responsibility is to work with Contracts to create a multiple or single award ID/IQ or BPA contract with qualified vendors that specialize in delivering BI/analytics. This contract vehicle would allow the BI COE to easily access a pre-qualified firm (or firms) that can do more than just provide technical resources – they’ll actually help the organization strategically apply analytics with measurable business impact. The firm(s) selected should be focused on providing both thought leadership on BI/analytics and technical solutions. Specifically, these firms should be experts at analyzing a business situation, determining how to apply analytics to it, and delivering the business results. If the BI COE does not do this, and instead defaults to the using a staffing firm, or any of the large IT generalist firms with existing contracts, they will not achieve the business impact they need to sustain themselves going forward.

Having a pre-established contract also helps prevent another major source of project failure in government BI/analytics projects. Often the desire to avoid the hassle of a new contract results in the government just asking their current IT contractor to pick up any new IT work. While this may be effective for other types of IT work, with BI, this is usually a first step toward failure. Agencies that choose this path often waste tremendous amounts of money and time while their existing IT contractor struggles to learn how to do specialized analytics work, usually unsuccessfully.

Another way to allow the BI COE to have influence over expenditures would be for them to create and manage a certification process and require that any firms bidding on BI/analytics work that falls outside of the BI COE contract be certified on the agency’s BI standards before they can bid. This would allow the BI COE to ensure consistency and uniformity of best practices, even when it has been decided that a different contract vehicle should be used.

Finally, there is another critical nuance government agencies must deal with to complete the important step of separating analytics work from building transactional systems and truly become data-driven. There is an inherent conflict of interest in having the same firms that implement the agency’s transactional systems also implement the systems that analyze their performance. When the creators of the system are responsible for measuring its success, there’s a clear incentive to use the data they control to tell the story they want heard. In order to truly become a data-driven organization, with consistent, accurate, valuable and objective analytics measuring and driving their performance, government agencies must preclude the firms that implement their transactional systems from also providing the agency’s analytics work, including any BI COE contract.

Separating the analytics from the systems contract and creating a properly implemented BI COE is a significant—but necessary—departure from the status quo. When the contractor in charge of implementing the transaction system is also in charge of analytics, their primary goal will be to make sure their big system-building contract looks good and continues to provide them with huge profits at the cost of objective performance metrics. If the organization is truly committed to being data-driven, separating transaction system work from analytics work only makes sense.