Within many organizations, data has begun to fuel many aspects of daily life, from business decisions to products functionality. In the rush to become data driven, many organizations have correctly begun by investing in their data platforms and strategies for hiring and upscaling hard to find data talent. Even with the right people and infrastructure in place, many organizations are still struggling to glean value from their hefty investment in data. This talk will address a key and often missing component of a successful data strategy: the ability to build a winning data project portfolio.
This talk will lay out a framework to help leaders decide which data projects to invest in and when. We will walk the audience through this framework and how we’ve used it. Different projects carry different risks and potential benefits. Too often data teams are forced to focus primarily on “quick wins” or dedicate too much of their time and energy to exciting sounding but impractical “moon shots” that fail to deliver business value. We advocate for an intentional approach to data project portfolio creation that focuses on project that have the right mix of risk/return and that enable future advances.
We’re prescriptive in terms of mechanisms for making trade offs on data projects, but believe that project idea generation should include as many people from different backgrounds and parts of the organization as possible. Once a large set of ideas is generated, take those ideas and plot their predicted cost and impact. Impact can be a broad net so we recommend thinking of it in terms of both impact to clients (does it solve a known customer problem?) and impact for the organization (might this project enable many future projects?). Where you see projects that are related or can be built on the same underlying components, draw lines to connect them.
Once clusters of projects begin to form, go back through and mark the feasibility of each project. There’s no need to throw out projects that aren’t feasible today, rather this exercise allows for the visualization of what investment is likely needed in the future. Think critically about chains of projects where one project enables one or many others. Those clusters of projects will have a higher impact.
In an ideal cluster, there is a healthy mix of “quick wins” (typically process improvements), incremental gains (making existing products better or automating workflows) and innovation (using data to create new product directions). The right blend of these different types of projects varies by company and stage. Return on investments should be revisited often and projects must be linked to what they’ve enabled.
Last but not least, it’s important to make sure that the projects chosen align (or effectively challenge) the overall goals of the organization. Finally, if the feasibility of the ideal mix of projects chosen doesn’t reflect what a team in house is capable of, this is a great time to cycle back and update the data strategy around people, data, or tools.