Your strongest engineering manager resigns on a Thursday. By Friday you're running their three teams yourself, and the only person who could plausibly step in has never run a planning cycle. You knew this was a risk. But succession planning tends to live in an HR template that gets filled in once a year and never read, so when the gap opens there's nothing behind it. This post shows how to keep a short, honest succession plan for the leadership roles under you, and how to turn it into development work that happens this quarter.
Why the annual template doesn't work
Most succession plans fail in one of three predictable ways:
- Names without evidence. A box says "ready in 12 months" because that's what the candidate's manager hoped. Nobody can say what they've done that suggests it.
- No development attached. The plan identifies a successor but changes nothing about their work. Twelve months later they're exactly as ready as before, and the box still says "12 months".
- Secrecy. The plan sits with you and HR. The candidate doesn't know they're on it, their manager isn't building them up, and when you leave, the plan leaves with you.
The fix is a plan small enough to review every quarter, where every name is tied to a specific gap and a specific stretch assignment.
Pick the roles that actually matter
You don't need a successor for everyone. Start with the roles that would hurt most if they were empty for three months. For a director or head of engineering, that's usually:
- Your direct reports, especially managers of managers.
- Technical leaders whose judgment the organization leans on: the principal engineer everyone asks about the data model, the architect who owns the payments design.
- Roles that are hard to hire externally because they need a lot of context: the person who knows how the legacy billing system really works, or who holds the relationship with your largest customer's engineering team.
- Your own role. If you'd struggle to name who could cover for you during a three-month leave, that's the most important row on the page.
A quick filter: if this person left tomorrow, would a decision of consequence stall within a month? If yes, the role goes on the plan.
The one-page plan
Keep it as a single table. Here's a hypothetical example for a head of engineering with four direct reports and one critical technical role.
| Role | Incumbent risk | Candidates and readiness | Biggest gap | Stretch assignment this quarter | Owner |
|---|---|---|---|---|---|
| Director, Product Engineering | Medium (recruiter interest) | Priya: 12 months. External hire: 4–6 months | Has never owned a budget or a headcount plan | Builds next year's headcount plan for their two teams, reviewed with finance | Director + you |
| EM, Payments | Low | Sam: ready now (covered a 6-week leave) | None blocking; wants more exec exposure | Presents the quarterly payments review to the CFO | EM |
| EM, Platform | High (burnout signals) | None internal | — | Hire a senior EM into Platform with explicit succession intent; split on-call ownership now | You |
| Principal engineer, data | Medium | Lee: 2 years | Design authority beyond one team | Leads the cross-team schema review; writes the decision records | Principal |
| Head of Engineering (you) | — | Priya: 2+ years. Sam: 2+ years | Neither has run a multi-team planning cycle | Each runs one quarterly planning session for their area, start to finish | You |
Notes on the columns:
- Incumbent risk is your honest guess at the chance the role is vacated in the next year: resignation, promotion, burnout, a long leave. The Platform row, with high risk and no internal candidate, is the most urgent on the page.
- Readiness uses three buckets: ready now, about 12 months, about 2 years. Anything further out is a hope, not a plan. "External hire" is a legitimate entry; it means you'll have a job description ready before you need it.
- Biggest gap is one thing, phrased as something the person hasn't done yet, not as a personality trait. "Has never owned a budget" is actionable. "Not strategic enough" isn't.
- Stretch assignment is the point of the whole exercise. Every candidate row should change what someone works on in the next three months.
Judging readiness on evidence
Readiness is where wishful thinking creeps in. Two habits keep it honest.
First, ask what the candidate has already done at the next level, not what they could do. Covering a peer's team for six weeks during a leave, running a planning cycle, handling a performance problem in someone else's team: these are evidence. Strong delivery in the current role is not, on its own.
Second, use a model of what changes between levels. Ram Charan, Stephen Drotter and James Noel's The Leadership Pipeline describes a series of "passages" between leadership levels, for example from managing others to managing managers, and argues that each one requires a change in three things: skills, how you spend your time, and what you value in your work. That's a useful checklist when you assess a candidate. A strong engineering manager stepping up to director may have every skill and still not be ready, because they value being close to the work and spend their week in their teams' design reviews. Ask the three questions for each candidate:
- Skills: Can they do the new job's core tasks, such as building a headcount plan, coaching a manager, or negotiating priorities with a peer function?
- Time: Do they already spend some of their week on work at the next level, or would they need to drop everything they currently do?
- Values: Do they get satisfaction from other people's results, or mainly from their own?
Skills can be taught. Time and values shift only through doing pieces of the job and getting feedback on it.
Turn gaps into stretch assignments
A good stretch assignment is real work at the next level, with enough support that failure is survivable. Some that work well for engineering leadership roles:
| Gap | Stretch assignment |
|---|---|
| Budget and headcount | Builds the next-year plan for their area; defends it in the review with finance |
| Managing managers | Takes on a first-time manager as a coaching relationship for a quarter |
| Executive communication | Presents their area's quarterly review to the leadership team, with you in the room but silent |
| Cross-team influence | Owns a dependency or migration that needs two other teams to change their plans |
| Running a planning cycle | Runs one quarterly planning session from intake to final commitments |
| Breadth | Covers a peer's area during their leave or sabbatical |
Covering a leave is both the best readiness test and the one most often wasted. When a manager goes on leave, the default is that you absorb their work. Hand it to your succession candidate instead and coach them through it: you get evidence, they get experience, and you get your week back. If letting go is the hard part, see delegating without hovering.
Talk about it openly
Keeping the plan secret feels safer, but it removes most of its value. A few rules make openness workable:
- Tell candidates they're being developed, not promised. "I see you as a possible future director, and here's what I think you'd need to show. This isn't a guarantee, and the role may never open." That's honest and motivating. Avoid anything that sounds like a commitment to a specific job on a specific date.
- Ask what they want. Some of the people you'd pick as successors don't want the role. Some would rather grow as technical leaders than as managers. Finding that out now is far better than finding it out when you offer them the job.
- Share the plan with your own manager and HR. If you're the single point of knowledge, the plan disappears the day you do. Your manager should be able to read the table and act on it without you.
- Keep incumbent risk private. The "incumbent risk" column is your judgment about real people's circumstances. It belongs with you, your manager and HR, not in a shared document.
When a candidate does step up, they'll often be managing former peers; point them to managing the people who were your peers last week before their first day.
The quarterly review
Thirty minutes a quarter is enough. Put it in your calendar now, alongside your planning cycle, and walk the table row by row:
- Did the stretch assignment happen? If not, why not? A plan where assignments keep slipping is telling you development isn't really a priority.
- What did we learn? Move readiness up or down based on what the candidate actually did, not on how the quarter felt.
- Has incumbent risk changed? New signs of burnout, a promotion on the horizon, an upcoming leave.
- Any empty rows? A critical role with no internal candidate needs a decision: develop someone from further away, hire with succession in mind, or restructure so the role isn't a single point of failure.
- Set next quarter's assignment for every candidate row.
Building leadership bench strength is one of the areas covered in the PTME study guide, alongside organizational design and executive communication.
Start this week
Write down the five roles in your organization whose departure would stall a decision within a month, including your own. For each, write one name and one thing that person has never done that the role requires. If you can't fill in a name for two or more of them, that's your most important leadership problem this quarter, and it's a lot cheaper to work on now than on the Friday after someone resigns.