Hire the leader who will own operations before the design is finished, not at handover, and add the rest of the operations team as the project moves from construction to commissioning. By commissioning, the shift supervisors and technicians should be in seat and training on the systems they will run. Uptime Institute has argued for years that operations should take part in a data center project from the start, because late changes to a finished facility are costly. The hiring market makes that harder: in Uptime's 2026 survey, more than half of respondents reported difficulty finding qualified candidates for open positions.
This guide ties each hire to a project phase. It combines Uptime Institute's published roadmap of operations work with an editorial recommendation on which leadership role to have in seat at each stage. Uptime does not name job titles or hiring dates, so the role mapping below is our recommendation, not an industry standard.
In an Uptime Institute Journal article from September 2014, the argument is that operations is easily treated as the last group to the table, when it should be an active participant from the beginning. The article notes that many data centers have to make significant changes in their first year after substantial completion, and that value engineering done without operations input can raise lifetime costs by more than it saves at build.
A March 2015 Uptime Institute Journal article on staffing splits the problem into three parts: the number of staff, their qualifications (licences, experience and technical training), and the organization (the reporting chain, with roles and responsibilities defined). Across the sites Uptime had assessed, staffing numbers were inadequate in only 7% of data centers. Qualifications were inadequate in about twice as many, and organization was ineffective more often still. The same article reports that, in Uptime's incident database, the root cause of 39% of data center incidents fell in the operational area. These figures are from Uptime's assessments at the time of publication and describe the sites it reviewed, not the whole industry today.
The practical reading: a plan that only counts heads misses the two problems Uptime found more often. The leaders you hire first decide the qualifications you ask for and the organization you build.
Uptime's roadmap uses five design-build phases: pre-construction, design, construction, commissioning and turnover. The second column of Table 1 summarizes the operations work Uptime places in each phase. The third column is our editorial recommendation for the leadership role to have in seat by then.
Table 1. Operations work and leadership hires by project phase
Phase | Operations work in Uptime's roadmap | Leader to have in seat (editorial recommendation) |
|---|---|---|
Pre-construction | Define measures of success, service levels and KPIs; set the concept of operations and of maintenance; establish vendor requirements; define roles and an organization chart covering Operations, Facilities, IT and Security. | Head of data center operations (a VP, director or site general manager, depending on scale). This person owns the concept of operations and decides what is in-house and what is contracted. |
Design | Review the design for failure modes and ease of maintenance; define staffing levels, shift strategy and staff qualifications; set vendor service contracts and maintenance plans; choose the maintenance management system. | Lead critical facilities engineer, plus a maintenance and reliability lead. They review drawings with the head of operations and turn the staffing plan into role descriptions. |
Construction | Write standard, method and emergency operating procedures; define the training program; set up the maintenance management system and document library; set shift and inspection protocols; establish critical spares and change management. | Operational readiness or transition lead, a controls and building management system lead, and the shift supervisors who will run the site. |
Commissioning | Commission the operating procedures; train staff on infrastructure systems; run emergency drills and recovery training; deliver safety and vendor training; populate the maintenance system. | Technicians and operators, hired early enough to take part in testing and training. The site leader signs off that the team is ready. |
Turnover | Review commissioning results and the punch list; put the operations management program into effect; refine and exercise the procedures. | The same site leadership team now accountable for production. Plan for continuity, because the people who were present at testing hold the knowledge. |
Source for the second column: Uptime Institute Journal, “Start with the End in Mind” (September 18, 2014), summarized. The roadmap predates today's high-density and liquid-cooled builds, so treat it as a process frame and add the qualifications your own design requires.
Commissioning is where the operations team and the contractors meet. Data center projects commonly organize testing into five levels, ending with integrated systems testing, such as a simulated utility failure to confirm that backup power picks up the load (CxPlanner, a commissioning software vendor; numbering varies by project). A 2024 JLL article on commissioning strategy describes a staged cutover with written procedures, predefined go and no-go criteria for each stage, and the client's operations team involved in planning. In our view, operators who join after these tests miss the chance to see how the facility behaves when systems fail, before customers depend on it.
Titles differ by company. Use the table to separate the work from the label, and ask each candidate for evidence of the work. The roles and evidence prompts are editorial recommendations built from the phases above.
Table 2. Operations leadership roles for a new data center (editorial recommendation)
Role | Owns | Evidence to ask for |
|---|---|---|
Head of data center operations | Concept of operations, service levels, staffing model, in-house versus contracted scope, readiness sign-off. | A facility they took from testing into production; the procedures and staffing model they wrote; how they handled a serious incident. |
Lead critical facilities engineer | Design review for maintainability and failure analysis; technical standard for the site; escalation on electrical and mechanical issues. | Design changes they made because of an operations review; experience with the redundancy level and cooling approach of this site. |
Maintenance and reliability lead | Maintenance plans, vendor service contracts, critical spares, the maintenance management system. | A maintenance plan they built from scratch; how they set vendor response times; what they track. |
Operational readiness or transition lead | Procedures, training program, drills, documentation hand-off, readiness checklist. | A readiness plan from a previous opening; how they confirmed staff competence before handover. |
Controls and building management system lead | Monitoring and alarm configuration, reports, integration with the maintenance system. | Alarm rationalization work; how they tested alarms during commissioning. |
Shift supervisors | Rounds, alarm response, change approval on shift, handover between shifts. | Time as an authorized operator on comparable systems; how they run a shift handover. |
For power and construction leadership, which sit beside this team on a campus build, see our guides to data center power leadership and data center construction leadership.
The 2015 Uptime article gives a simple method. First list the hours needed each year for maintenance by trade (preventive and corrective work, vendor support, project support and tenant work orders). Then decide the shift presence required, since presence requirements rise as uptime objectives rise: whether people are on site around the clock, and which trades are needed on each shift. Divide the total annual hours for each trade by the productive hours one person can work in a year. The result is rarely a whole number; Uptime says fractions can be covered by overtime (it advises keeping overtime below 10%), by contractors, or by rounding up.
The example below is illustrative arithmetic, not a staffing recommendation. It shows why shift presence alone drives the count.
Table 3. Illustrative example: people needed to cover shift presence only
Step | Assumption or result |
|---|---|
Coverage required | Two qualified people on site at all times (assumption) |
Hours in a year | 8,760 (365 days of 24 hours) |
Annual shift-presence hours | 2 × 8,760 = 17,520 |
Productive hours per person per year | 1,800 (assumption; depends on leave, training and local rules) |
FTE required for shift presence alone | 17,520 ÷ 1,800 = about 9.7 FTE |
This equals about 9.7 full-time equivalents for continuous shift coverage under the stated assumptions. It is not a final headcount: round up and test the roster against qualification, leave, rest and break requirements. Add maintenance and project hours only to the extent they cannot be covered during those shifts, so the same work is not counted twice.
Hiring operations leadership after the design is fixed. Maintainability and failure analysis are cheaper to act on during design. Uptime places both in the design phase.
Counting staff instead of qualified staff. One of the sites in Uptime's 2015 article had two to three people on every shift, yet only two people on the whole staff were qualified to operate and maintain the critical equipment.
Using overtime as the plan. Another site in the same article had qualified people but too few of them, and overtime well above the advised 10% limit, with signs of fatigue.
Leaving training to the last weeks. Uptime places the training program in the construction phase and the drills and system training in commissioning. Roles hired late cannot take part in either.
Assuming the market will supply people on demand. Uptime's 2026 survey, based on responses from more than 800 data center owners and operators, found that more than half reported difficulty finding qualified candidates and that staff turnover remains a persistent problem, with staff often lured away by other data center companies. Plan retention and development, not only recruitment.
The phase the project is in, and which roles in Table 2 are already filled, in-house or contracted.
The availability target and the shift model, since both determine the staffing count and the qualification level.
Which functions will be performed by in-house staff and which by contractors.
Start dates tied to project gates (design review, start of testing, handover), not to calendar months.
The evidence you will ask for in each role, using Table 2 as a starting point.
The location, relocation support and the pay you can offer, set against what competing operators in the same market are offering.
With these decisions written down, a recruiter can price the search and test candidates against the work rather than the title. To compare search partners, read how to choose a data center executive recruiter, and for the market picture behind these hires see the 2026 data center talent shortage.
Start with the head of operations while the concept of operations is still open, add engineering and maintenance leadership during design, put procedures and training in the hands of a readiness lead during construction, and bring technicians into commissioning. For senior operations, construction or power searches on a data center build, companies can look at Christian & Timbers' data center executive search practice.
Sources: Uptime Institute, 16th Annual Global Data Center Survey press release (July 28, 2026); Uptime Institute Journal, “Start with the End in Mind” (September 18, 2014); Uptime Institute Journal, “Proper Data Center Staffing is Key to Reliable Operations” (March 12, 2015); JLL, commissioning strategies (November 22, 2024); CxPlanner, the five levels of commissioning (August 6, 2026).
