A blitz is a leadership tactic for getting results fast. However, any sort of blitz, whether it’s an American football defense going for a quick sack or a team of data scientists trying to quickly solve a problem, has a time and a place. The right time to employ a blitz for a data science project is when you’re trying to solve a problem that’s specific and limited in scope.

Rally the team

Most teams do well when they’re allowed to focus on one specific problem; a data science team does extremely well in this situation. If you do your upfront work correctly and tightly define your purpose and outcome, you’ll experience great synergy from your data science team.

Data scientists operate with a precision mentality that requires focus; in most circumstances, this translates to leaving them alone. A blitz is a special circumstance that allows the group to multiply their individual powers into something pretty awesome.

Plus, running a blitz is fruitful and fun. Although data scientists should be doing most of their work in the solitary comfort of their own cubicles, blitzes are usually a welcome excuse to network with other great minds.

Run a modified Kaizen Blitz play

What is a Kaizen Blitz?

A Kaizen Blitz is a management technique whereby a concentration of resources is organized for a short period of time to attack a very specific issue. The facilitation on a Kaizen Blitz is extremely tight, and a successful one can accomplish in one week what would otherwise take months on a typical Lean Six Sigma project.

I don’t recommend running Kaizen Blitzes in your organization unless your team is already familiar with Lean Sigma. I like the Kaizen Blitz concept though, so we’ll use the idea to construct our own blitz format.

The benefits of a one-week blitz

A Kaizen Blitz squeezes an entire Lean Six Sigma project into the span of a week. A typical Lean Six Sigma project will traverse five well-defined phases in about six to eight months. With the right problem, focus, and leadership, this entire process is tightly condensed into the span of only five days. In a similar fashion, you can compress a data science process like your innovation funnel, as long as all the right pieces are in place.

Blitzes can be used effectively for short-term and long-term results. A short-term result can accelerate the value of your efforts and provide critical information on how to steer your organization to the long-term result. For instance, stringing together four back-to-back Kaizen Blitzes through a tight series of improvement efforts will give you incredible insights on how well your assumptions are aligned with your long-term prospects.

Requirements: the right problem and leadership support

Above all, you must have the right problem — it should be specific and limited in scope. I’ve often consulted with leaders who like the idea of quick results, but they attempt a Kaizen Blitz with a problem that’s too big. If every Six Sigma project could be run in a Kaizen Blitz, you wouldn’t need Six Sigma projects, right?

Furthermore, you must have the leadership to set priorities and clear calendars. Blitzes involve more than just your data science team; these intense efforts will also include people from marketing, product development, operations, and other departments. It seems like a great idea to have all these people at your disposal for a week to solve your most pressing problems, though the reality is it’s tough to pull people away from their regular assignments for any extended period of time. You must exercise your leadership and allow your team to focus intently for one week without too much disruption in their normal jobs.


Blitzes are an engaging and effective tactic that can bring quick results from your data science team when used properly. First, make sure you have a good problem or opportunity that deserves concentrated attention from your best resources for a short period of time, and then muster the leadership to make the call. Finally, make sure you’re in tune with your bigger picture, or else your team’s winning efforts might be wasted.