Many  organizations are hesitant to launch themselves into needed  organizational transformation. Not only is it disruptive and  uncomfortable, but experience taught them that change is never really  over. Thus, they prefer to avoid it until they have absolutely no choice  anymore.

Let’s  be honest here. The world (and your niche market) is changing whether  you adapt to it or not. Refusing change just because more of it will  come later is akin to not paying your taxes because you paid them last  year. By avoiding or delaying adaptation, you only make it harder on  yourself.

You  can alleviate the discomfort of change. The trick isn’t to delay it,  but to make it easier to happen naturally and gradually. This way, you  avoid disruptive and downright frightening organizational transformation  events. Make flexibility and adaptation part of your organizational structure.

The power of a sandbox-driven decentralized model

If the term sandbox-driven decentralized model is new to you, don’t worry. I wrote a piece on the decentralization scale that explains everything. But here’s the kicker:

  • A sandbox-driven decentralized model has a central authority who set the organization’s goal and intent,
  • it  uses sandboxes to manage work, combining a specific mission with a  clear purpose, resources, expectations, and constraints in one neat  package.
  • it  uses autonomous teams who are dependent on one of the sandboxes, and  who must figure how to deliver the mission goals within the boundaries  set by the sandbox while maintaining profitability and still minding the  organization’s goal and intent.
  • Each teams build and maintain relationships with other teams as needed, without any red tape.
  • The  whole system measure success through the ROI of the delivered value.  Not process. Not management. If the results are there and everybody’s  happy, what more is there to ask for?
A simple example of the model.

For those who have read Dave Gray’s The Connected Company, the link with the model he described, podularity,  should be obvious. After a few years of implementing this approach, I  tend to put a bit more focus on the sandboxes, though, as they are key  for the system to work.

The  power of the model lays in its flexibility. While the steering of the  company (so to speak) is centralized, neither the work nor the teams  are.

Central authority

Large,  centralized models are hard and slow to change. But in this lighter  model, this is perfect for overarching organizational goals and intent  that should stay stable for at least a year. While goals and intents  benefit from slow changes, the missions driving the sandboxes can still  move much faster.

The central authority has a different relationship toward decision power than traditional management. Their goal is to manage the system rather than the people or the work. They confine their power to  strategy: They own the overall objectives and intent of the company. All  they have to do then is to set up the sandboxes needed to reach those  goals.


Sandboxes  are temporary in nature to allow the organization to adapt quickly to  an ever-evolving context. Their scope is more medium-term (usually, at  least), with the possibility of change or cancellation at each  milestone. Sandboxes manage work, not people.

Sandboxes are great to deal with subject matters that concern the organization right now.  Any given sandbox isn’t meant to last long into the future. When  context changes, close the sandbox and open a new one with a more  relevant budget, and constraints. Always set short milestones, and task  each to deliver value that can be used right away. By doing this, you  get immediately consumable value at each step. You can close a sandbox at any time before its mission is completed without losing your investment.

Both the central authority and the teams contribute to the sandbox.

Sandboxes  are living things. Both the central authority and the teams keep a  dialog open to tweak and evolve the sandbox as needed. The goal is to  keep it relevant and makes it a platform for delivering the expected  value.

Autonomous teams

Autonomous  teams are where people belong. Those teams are stable (with maybe some  specialists temporarily required for their unique expertise), usually  composed of a mixture of expertise, and have the authority to make their  own decisions as long as they stay within the boundaries set by the  sandbox they are currently attached to and as long as they deliver a  positive ROI.

Autonomous  teams are trained and organized so they can easily be switched between  sandboxes as needed. The team thus have to focus on one mission at a  time only but expect to move on to a new mission eventually. The event  won’t be a surprise either: teams are always very aware of the ROI  delivered through the sandbox and the milestone-based progress, and if  those aren’t satisfactory, the teams will know it well before central  authority make a decision. In fact, teams themselves might make the case  to end a mission because it doesn’t bring the expected value.

Autonomous teams also have the authority to deal with what we call local challenges,  or context challenges that affect them but not necessarily other teams  within the organization. Those could be related to specific clients,  local laws, quirks in their niche, and such. The dealing with those  issues is part of the team’s continuous improvement efforts, and  shouldn’t concern the rest of the organization.

Making uncertainty irrelevant

Traditional  organizations rely on predictability. They must predict context and  outcome a long time in advance to assign proper resources, plan  long-term strategy, and measure success. If you’ve been working in those  companies in the past decade or so, then you know how well this  approach turns out.

Between market shifts, rapid technological advances, emerging competitors, and disruptive innovations, prediction start to sounds more and more like divination,  only using Gantt charts instead of tarot cards. Plans go awry all the  time. However, since they are usually locked in because of budget and  resources allocation cycles, doomed projects can trudge on for a long  time before the organization reacts.

With a sandbox-driven decentralized model,  a change of plan, of context or need can be addressed quickly at a low  cost. Suddenly, it doesn’t matter if the change is the result of  upheaval in markets, a disruptive competitor, or a realignment of the  entire organization. Employees expect change, regardless of the cause.  They don’t have to face it since it’s part of their daily lives.

With  such a model, change becomes an opportunity to explore new things,  experiment, and find innovative solutions. Since engaged employees  thrive on those challenges, you end up transforming fear of the unknown  into excitement. Now that’s a change!