Designing Your Org for the Next 18 Months
Most reorgs are what happens when nobody designed the org in advance. Structure accumulates through a year of individually reasonable decisions, the friction becomes obvious, and then a restructure is required — which costs a quarter of disruption and a real amount of trust. The alternative is unglamorous: once a year, spend half a day designing the structure you will need eighteen months out, then hire and change toward it deliberately. Companies that do this restructure rarely, because they are always adjusting slightly rather than correcting late.
What’s inside
Why eighteen months is the right horizon
The four-step exercise
What to decide now and what to defer
The hiring sequence that follows
Designing for the org you will have
The review cadence
Where Blomma fits
Why eighteen months is the right horizon
Shorter horizons do not change any decisions. If you plan six months out, the structure you need is roughly the structure you have, so the exercise produces nothing actionable.
Longer horizons are fiction. At startup growth rates, a three-year org plan is a guess about a company whose strategy will have changed twice. Designing to it wastes effort and, worse, produces premature structure — layers and specialists you build for a company that never arrives in that shape.
Eighteen months is where the two curves cross. It is long enough that the structure genuinely differs from today’s — you will have added a management layer, or split a function, or built an executive team — and short enough that your strategy is probably still recognisable. It also matches hiring lead times: a senior hire you need in eighteen months is a search you start in twelve, which is a decision you make now.
One practical note: tie this to whatever planning cycle you already have. An org plan produced alongside the annual plan gets used. A standalone org exercise gets filed.
The four-step exercise
Half a day, once a year, with your cofounder or one or two executives. The output is two pages.
Step one: name the constraint at the end of the period.Not today’s constraint — the one you expect in eighteen months, given the plan. If the plan is to move from mid-market to enterprise, the constraint will be an enterprise go-to-market motion you do not have. If the plan is a second product, the constraint will be that everything currently routes through one product team. Get this wrong and the rest of the exercise optimises for the wrong thing, so spend real time here.
Step two: list the decisions that have to be made well and fast by then.Concretely. Enterprise pricing and contracting. Roadmap sequencing across two products. Which segment gets engineering capacity. Then note, for each, where it would get decided under today’s structure and whether that works.
Step three: design the smallest structure that puts those decisions in the right place.Not an aspirational chart — the minimum change from today. Which functions split, which layer gets added, which outcome needs a single owner it does not have. Resist the urge to draw the org of a company three times your size.
Step four: sequence it.What changes in the next quarter, what in six months, what in twelve. With named triggers rather than dates where possible: “we add the second engineering manager when the team passes twelve,” not “in Q3.” Triggers survive plan changes; dates do not.
What to decide now and what to defer
The discipline of this exercise is deciding as little as possible while still being useful.
Decide now:the senior hires with long lead times, because a twelve-month search is a decision you are already making by not starting it. The outcome owners for anything that will matter in a year, because those people need runway to grow into it. And any structural change that is blocking something today — defer that and you are choosing to pay the friction for another year.
Defer:exact team boundaries below the top layer, because they depend on who you hire and how the work evolves. Specific headcount numbers per team, which will be wrong. Titles, which cost nothing to decide later and constrain you if decided early. And any layer you are not yet forced to add — premature layers slow decisions and are hard to remove.
Write down explicitly what you are deferring and what would trigger revisiting it.This is the step that makes the plan honest rather than vague. A deferred decision with a named trigger is planning. A deferred decision with no trigger is avoidance, and it is how companies arrive at the friction they were trying to avoid.
The hiring sequence that follows
An org plan’s main output is a hiring sequence, and it should be ordered by lead time rather than by urgency.
Senior roles first, because they take longest and because they determine the structure below them. Hiring a VP after you have hired their four managers means the VP inherits a team they did not build, which works considerably less often than the reverse.
Then the roles that unblock other hires — a recruiting lead before a hiring surge, a finance lead before a fundraise, a first manager in a function before that function doubles.
Then capacity hires, which are the most flexible and should absorb the variance when the plan changes.
One principle worth applying: hire the person who can run the function at the end of the period, not the one who fits it today. Slightly over-hiring on judgment at the senior level is usually the right error, provided there is genuine scope for them within a year. Under-hiring means a second search in twelve months, plus the cost of managing someone out.
Designing for the org you will have
The most common failure in this exercise is designing for an idealised company rather than the one you will realistically have.
The idealised version assumes you will hire every role on time, at the level you specified, and that they will all work out. The realistic version accounts for the fact that a senior search takes six to nine months, some hires do not work out, and you will not fill everything.
So design with slack. Ask which parts of the structure depend on a hire you have not made, and what happens if that hire is four months late or does not happen. A structure that only functions if all five senior hires land is a plan with five single points of failure.
And be realistic about your existing people. It is tempting to plan a structure that requires three of your current managers to have grown substantially. Some will; not all. The honest version names who needs to grow into what, and what support they will get — because a plan that quietly assumes growth without funding it is the most common way org plans fail, and the failure gets attributed to the individuals.
The review cadence
Annual design, quarterly check, with a specific question at each.
Annually:the full four-step exercise, from the constraint up.
Quarterly:has the constraint changed, and have any triggers fired? Thirty minutes. This is what stops the plan from becoming a document you wrote in January and abandoned in March.
After any significant strategy change:re-run step one. A new market, a new product, or a large round changes the eighteen-month constraint, and a structure designed for the old one will actively work against the new plan.
And keep the previous year’s plan. Comparing what you predicted with what happened is the fastest way to get better at this — most founders discover they consistently under-plan senior hires and over-plan structure.
Where Blomma fits
The hard parts of this exercise are the two honest assessments inside it: which of your current leaders will genuinely be able to run their area at three times the size, and which decisions you are going to have to stop making yourself. Both are difficult to think through with the people involved, and your executives cannot advise you neutrally on a plan that determines their own scope.
Blomma is an always-on AI career coach with no stake in your structure. Use it to pressure-test step one, because an org plan built on the wrong constraint is worse than no plan. Use it to work through which of your existing leaders needs to grow into what, and whether you are funding that growth or assuming it. Use it to check whether the structure you have drawn still functions if two of the senior hires slip. And use it to prepare the conversations the plan implies — particularly with a leader whose area is about to be split or who will have someone hired above them. Where the situation warrants a human who has planned at your stage, bring one in.
The second application follows directly from the honest assessment above: where the plan requires people to grow into bigger roles, coaching is what makes that a plan rather than a hope..
A reorg is a correction. An org plan is a design. The difference between the companies that do one every year and the ones that do neither is half a day and a willingness to name the constraint before it arrives.