Cut Coordination Tax: Organize Around Outcomes

When we worked on the Gillette Exfoliating Razor, we tried something different. Instead of accepting the coordination nightmare, we asked a simple question: “If we had free reign, what’s the minimum number of people we’d need if they could really focus?”

The answer was about 30 people from across the R&D, Commercial, Marketing, Finance, Prep for Manufacturing teams.

Instead of feeding the work into the teams all these people belonged to, we created what you might call a virtual value center: 5 virtual teams that existed solely to figure out a winning proposition and get it to market.

Same amount of brainpower, spread over fewer people, concentrated to a shorter time, with less managerial overhead both because people were already working with the people they needed to coordinate with, as well as the shorter timespan.

It required some tough prioritization tradeoffs. It highlighted some islands of knowledge.

But it was worth it. The Exfoliating Razor became one of Gillette’s most successful products many years, developed and brought to market faster than anything they’d done before.

Here’s the thing. The breakthrough wasn’t about agile ceremonies or new tools.

It was about the coordination and collaboration architecture: designing your organization so that people can concentrate on what matters most.

Most organizations strive to optimize resource utilization, keeping everyone engaged across multiple initiatives. But innovation requires optimization for flow, getting the most valuable outcomes through the system as quickly as possible.

Now, let’s be honest about the constraints. Carving out focused teams isn’t trivial. You’ll hit knowledge islands where someone needs to split their time. You’ll face resistance from leaders who don’t want to “lose” their people to a cross-functional effort. And you’ll discover dependencies you didn’t know existed.

At Gillette, we had some advantages: executive commitment made the tradeoffs possible. Even then, certain specialists (like material designers) couldn’t focus 100% and had to play “floater” roles across initiatives.

And remember - we did this just for the #1 strategic priority, not across the board.

That’s an intervention pattern I’ve seen work well - Identify a strategic priority that requires intense cross-functional collaboration and faces significant feasibility/desirability risks. And experiment with this approach for that initiative - where there’s clear concern that the current ways won’t deliver the desired outcomes or would be too painful to manage.

Here are some even easier interventions to consider, that can help you identify and build the conviction to try this:

• Map your current people to your strategic initiatives and count how many teams and people are working on multiple efforts simultaneously. Set that as a KPI (your “Focus Factor”) and work to improve it over time.

• Use a “Portfolio”-level Kanban to bring more transparency to how many initiatives you’re working on at the same time. And to start conversations about managing the amount of initiatives in progress, especially those that cut across the organization.

Remember - The goal isn’t perfect focus (that’s rarely possible), but conscious focus. When you’re explicit about the tradeoffs, you can make better decisions about where to accept coordination costs and where to eliminate them.

Organize around outcomes rather than org charts. The label (value streams, empowered teams, or virtual pods) doesn’t matter. What matters is creating conditions where people can actually deliver on what you’ve asked them to deliver.

The Gillette experience taught us that sometimes the biggest constraint to innovation isn’t technology or market conditions: it’s how we organize ourselves to do the work. When you align structure with strategy, remarkable things become possible.

This came out of an episode of Scaling AI: From Activity to Impact.

Listen on the episode page or directly on Spotify.

Scaling Agile with SAFe requires more than certification: it takes context-sensitive implementation. Explore SAFe advisory or let’s talk.