Build the Operating Loop and Transfer Judgment

Late 2016. KKR and TowerBrook had just invested €250 million for a minority stake in OVH, with the US as a priority market. The US business had one employee in San Francisco and no customers. I came in as president and CEO of OVH US to build a $100 million subsidiary. I hired the executive team, started two US data centers, and acquired VMware's vCloud Air business. Then the US business began to gain traction, and Octave Klaba, OVH's founder, decided to control the two data center builds himself. The infrastructure was in the US. The authority over building it now sat in France. My mistake was treating the authority we started with as something that would hold. I never secured a boundary around it.

The decision log made the problem visible. Across eight OVH US standups, it captured 15 decisions. Nine were still pending. Seven of those nine were waiting for action from teams in France. The US team needed authority over the teams and resources required to act on its judgment.

I traveled to France every month for two years, a week at a time. It worked while I was there. I remember a dinner with Octave. I walked him through what we needed. He agreed completely. The US was the largest market in the world. The next morning I flew home. By Monday morning, the engineers in France were working on other things. Taking the issues to the founder was the right route. It still was not enough, because a conversation that ends in agreement is not the same as a commitment that holds after you leave. The priorities held only while I was in the room, which made me the hero the plan depended on.

This is where the operating system comes together. VOOCS means Vision, Outcomes, Ownership, Cadence, Systems. Vision defines the direction. Outcomes make success measurable. Ownership gives one person the authority and resources to act. Cadence puts results, commitments, and unresolved decisions into a regular review that ends with a decision. Systems preserve the instructions, records, access, and practiced coverage needed to repeat the work and improve it. Define it. Measure it. Own it. Close it. Scale it.

When all five hold together, the team can use its judgment without returning to the same person. At OVH US, most of the loop was running. The direction was clear, the migrations had dates, the standups and the decision log were in place. Ownership was the weak element. The team owned the commitments and the outcomes without controlling the work that determined whether we could deliver them. That decision log is what the Ownership gap looks like on paper.

Media Temple showed me the other side. The company had run on the founders' judgment and hard work rather than written operating plans, and when no one does the planning work, the founder fills the vacuum. We built a plan around the core mission and a detailed plan for each division that its manager could execute. At first Demian Sellfors, the founder and CEO, was dismissive. Then he realized what delegation required: giving up control. I remember him saying one day, "But that's what I do." My answer was some version of "Not anymore," and the harder part was helping him see what he would do instead. Customer support's NPS went up almost immediately once the managers had the authority to change how reps were evaluated, putting more weight on resolution and customer satisfaction. Media Temple connected the customer promise to the work. The mission called for premium support. The measures needed to show whether customers experienced it. Division managers needed authority to resolve the conflicts beneath those measures, and Demian needed to let them use it.

The Spurs make the same point from sports. When Popovich stepped down as head coach in May 2025 to serve as president of basketball operations, Mitch Johnson, 38, had been inside the system for years. He inherited a foundation, not a void, and in his first full season San Antonio won 62 games and reached the Finals. That run does not prove that culture alone produces wins; the roster had rare young talent, and no system wins without it. What continuity made possible was the rest: a successor leading from an inherited foundation, and an organization that never had to start over.

Chapter 5 puts VOOCS to work on one recurring problem: a support team handling routine locked accounts, one of the problems from Module 4. The book says the roles, targets, and review details are illustrative, not reported results from Media Temple or XTIUM. Read it as a model of a filled record, not as a case.

Vision is the customer promise: restore authorized access so the customer can get back to work. Outcomes: if the median time to restore access is 60 minutes, the 90-day target is 30, tracked beside repeat lockouts, reopened tickets, and failures to follow the approved identity checks. Ownership sits with the support manager. Trained reps handle eligible accounts under the approved procedure; failed identity checks, privileged accounts, and suspected security incidents go to the designated security owner. The manager assigns a backup, provides the required access, and reserves two hours a week for the expert to teach and review cases. The next commitment is concrete: by Friday, the backup handles ten eligible cases with the expert observing and completes a supervised exception exercise.

Cadence: at Monday's review, if only eight of the ten cases were completed, the team finds out why, and the manager protects the coaching time or changes the commitment with its consequences visible. A new deadline without a changed condition gives the team little reason to expect a different result. Systems: the approved procedure, the shared case record, access that belongs to the role, and a backup who has practiced ordinary cases and exceptions. The expert then steps out of the routine path for an agreed trial while staying available for defined exceptions. A quiet week is useful evidence only if the work and exceptions the system must handle have been tested. One record, one owner, and an existing team meeting are enough to begin.

One page: the VOOCS Operating Record. Five prompts, header fields, a decision check. The worked example above shows one filled in. Where a recurring decision keeps returning to one expert, the Hero-to-Multiplier Map sorts why (missing knowledge, access, authority, capacity, or practice) and names what the expert does instead.

If you run the company, open the record for one initiative and sign the Ownership line yourself: which decisions the owner controls, what is reserved, and how fast you will answer. If you manage in the middle and the decision is parked above you, take it to whoever controls the resource as a specific request: which work comes first, what capacity is committed, who can change that order. Get the answer in writing with a review date. If the commitment will not hold, change the customer commitment rather than carry a date you cannot deliver. If you are the expert everyone calls, your next contribution is teaching the judgment, not approving every step. Make the transfer an explicit part of your role, with time on the calendar, a named backup, coaching, recognition in the review and the promotion case, and an honest conversation about what you move to next. Put both sides in writing.

Think about the engineer who always gets called in, works until sunrise, and solves the crisis through heroics. Everyone calls them the firefighter, and everyone means it as praise. The company taught them that every time it thanked the rescue and never funded the prevention. Help them carry that expertise into prevention, recovery procedures, and teaching the team to respond. Reward people for building systems that reduce repeated failures and help others handle the next one.

The expert who teaches is not giving the job away. They are earning the next one.