Jess Weaver Grey

By Jess Weaver • February 13, 2018

The names of the offices vary: Office of Innovation, Department of Performance and Accountability, Office of the Chief Technology Officer. But these citywide teams have plenty in common. Made up of data scientists, technologists, and engineers, the departments tasked with pioneering interventions in data analysis, predictive analytics, and other initiatives that fall under the innovation umbrella of “smart cities” are new, lean, and very, very busy.

 

As cities across the country have embraced the power of data analytics to optimize services and anticipate emerging problems, governments have recognized the need to invest in teams with the expertise to deliver the technically demanding projects. According to a group of chief data officers and other leading technical advisors from cities across the U.S. gathered at the Summit on Data-Smart Government, the greatest challenges, however, are rarely technical. With a small staff and a limited budget, the most significant challenges for these variously named offices are prioritizing, scoping, and managing projects in collaboration with other departments. Leaders from Chicago, New Orleans, San Diego, San Francisco, and Seattle shared their best practices for engaging with departments.

 

Phase I: Solicitation

Oliver Wise, former Director of Performance and Accountability for the City of New Orleans, articulated a somewhat counter-intuitive approach to soliciting potential data analytics projects from other departments: solicit more projects than you can deliver.

 

Without overpromising completion of any, broad project solicitation puts the power of prioritization into the hands of Wise’s team, enabling them to select the projects that are feasible, impactful, and strengthen relationships between departments or with key external stakeholders. With a lean and adaptive team, Wise’s office prioritizes projects with the most “spillover” effect, meaning that they generate outcomes that extend beyond the confines of the project itself – be it political will, technical capacity, or interdepartmental relationships.

 

According to Wise, identifying the key data science ambassadors is also paramount. For Joy Bonaguro, Chief Data Officer of San Francisco, that begins by articulating the team’s story or value add early on, and sharing examples of successful projects that launched other departments to new heights in service delivery and constituent satisfaction. Across the board, the team leads emphasized the importance of building trust with departments. Wise articulated his team’s role as “innovation matchmakers,” who recruit internal ambassadors by curating projects that carefully manage resources and offer critical impact in areas of need – key for exponential growth with minimal resources. Seattle’s strategic advisor Richard Todd noted that investing in trust-building with internal champions of data analytics projects builds the foundation for cross-agency collaboration – a notoriously tricky deliverable due to often conflicting timelines and priorities. Successful projects depend on more than a single evangelist, however, and more upon a department that exhibits genuine willingness to change its practices and means of service delivery.

 

While acknowledging the necessity of using data science as a buzzword to get internal stakeholders interested in solving problems, San Diego’s Chief Data Officer Maksim Pecherskiy discouraged falling into the trap of thinking that every analytics project has to be huge. Solving for specific problems can demonstrate value to hesitant departments without stretching the team’s resources or risking overpromising. Starting small, in other words, can have a big impact – in terms of both future collaboration and strategic problem-solving.

 

II. Scoping

The wide range of processes implemented by various city innovation teams illustrates that scoping is a spectrum. It also takes time. Whereas smaller cities engage in more informal exploration across departments, assessing the potential impact of each project and the data readiness of each department, others employ more structured scoping processes.

 

San Francisco, for example, employs a formal application process involving collaborative scoping. Departments have to make their own case for the scope of desired projects, with Bonaguro’s team providing feedback as the selection process narrows. Smaller cities like New Orleans and South Bend employ a more informal and collaborative process, working with strong ambassadors from departments to understand the problem they’re trying to solve, and designing the most feasible intervention. Key to identifying and scoping projects, particularly among lean teams, is the issue of scale, with data leaders constantly looking to how a particular project might be replicated for other departments or in a different issue area.

 

In both solicitation and scoping, the buy-in of executives is key, even if they are not involved in the more detailed negotiations. Without the support from leadership, departments are less likely to take risks and consider more experimental solutions.

 

The CDOs also warned of the temptation of overpromising, at both the mayoral and department level, since initial analysis and iteration can shift the trajectory of a project in both scoping and execution phases. Managing relationships and expectations is paramount, even with mayors who are highly supportive of data-driven solutions. Several in the group had experienced a zealous city leader brandishing a new technology, driven by good intentions but more awed by technical functionality than value for a particular use case. A foundation of trust and proven results enabled CDOs to push back, redirect, or engage in deeper inquiry around the key problems behind the desire to employ a novel technology.

 

III. Execution

The most effective strategy for avoiding surprises in project execution, Bonaguro noted, is normalizing feedback loops, which demands a culture of continuous and open communication – and the processes or structures necessary to reinforce it. Going into the mysterious “data cave,” admonished Bonaguro, is never advisable. Rather, teams should communicate their findings, decision-making, and strategies at every major juncture of a project in order to both educate and engage other departments.

 

In contrast to teams in New Orleans, San Diego and Chicago, which operate under a more traditional client/consultant relationship in scoping and implementing discrete projects, San Francisco uses a cohort model among departments. In addition to cultivating a sense of collective ownership and creating the container for interdepartmental relationship-building, this framework realizes the efficiencies of standardized timelines for project assessment and, when necessary, redirection.

 

San Francisco operates as a data literacy academy in project execution, in effect training data-smart leaders across departments. Even after scoping and strategy are set in place, ongoing user research and iteration were universal themes in ensuring the maximal impact of projects, and must be included in the scope and budget of each project. Of necessity, too, noted Bonaguro, are off-ramps when projects have derailed beyond the limits of feasibility or failed to deliver value to constituents.

 

When you’re as busy as these teams are, regardless of city size or the nature of the given project, your greatest asset is knowing when to jump in – and when to take a step back.