The Organizational Structure Design Problem
Most companies that set out to "fix the org" start by redrawing the org chart. New boxes, new reporting lines, a few titles changed, a reorg memo on Monday. Ninety days later the same things are still falling through the cracks: decisions stall, two people own the same number, three people own none, and the founder is still the bottleneck for anything that matters.
That's because the org chart isn't the problem. The org chart is a picture of who reports to whom. It says almost nothing about who owns what, how you'll know if they're doing it, or what happens when they're not. Reporting lines are the least interesting part of organizational design and the part most reorgs spend all their time on.
Structure is a design problem, not a drawing problem. And like any design problem, you don't start with the artifact — you start with the work that has to get done.
Start with the work, not the boxes
Before you decide who reports to whom, answer a harder question: what are the handful of outcomes this business actually depends on? Not activities — outcomes. "Run the AP process" is an activity. "Cash conversion stays under 45 days" is an outcome. Strip each function down to the small set of results it exists to produce, and you'll find it's a shorter list than anyone expects. Most middle-market functions live or die on three to six core accountabilities.
This is where the design discipline pays off. When you define structure around the work, two things happen that never happen when you start with the chart:
Every accountability gets exactly one owner. Not a committee, not a "shared" responsibility, not "the team." One name. The fastest way to diagnose a broken org is to point at a critical outcome and watch how many people raise their hand — or how nobody does. Shared accountability is unaccountability. The discipline of forcing a single owner per outcome surfaces the gaps and the overlaps immediately, and it does it on paper, before they cost you a quarter.
The structure follows the accountabilities, not the other way around. Once you know the work and who owns each piece, the reporting lines mostly draw themselves. You group owners by the natural seams in the work, not by org-chart aesthetics or who's been there longest. The chart becomes an output of the design, which is exactly what it should be.
This matters double in a merger. When you combine two companies you have two of everything — two heads of operations, two pricing processes, two definitions of "done." Redraw the chart first and you're just deciding whose box wins. Map the accountabilities first and you force the real conversation: in the combined business, who owns this outcome, and what does the other person do now? That's the conversation integration plans are supposed to have and usually avoid until Day 100, when it's already a retention problem.
Then attach KPIs that someone can actually move
Once every core outcome has an owner, each owner needs a way to know whether they're winning. That's what a KPI is for — and it's where most companies quietly sabotage the whole design.
A few rules that separate KPIs that drive behavior from metrics that just decorate a dashboard:
Measure the outcome, not the activity. "Number of sales calls" is activity. "Pipeline conversion rate" is the outcome the calls are supposed to produce. Activity metrics reward motion; outcome metrics reward results.
The owner has to be able to move it. If you hold someone accountable to a number they can't materially influence, you haven't built accountability — you've built an excuse generator. The KPI and the accountability have to belong to the same person.
Three to five per role, not twenty. A role with twenty KPIs has zero priorities. The discipline of picking the vital few is the whole point. If everything is measured, nothing is managed.
Mix leading and lagging. Lagging indicators (EBITDA, churn, conversion) tell you what already happened. Leading indicators (pipeline coverage, time-in-stage, defect rate at first inspection) tell you what's about to happen while you can still do something about it. Owners need both — one to be judged on, one to steer with.
Tie them to the value-creation thesis. In a PE-backed business, the metrics that matter are the ones that move the numbers the sponsor underwrote the deal on. If a KPI doesn't ladder up to the investment thesis, ask why it's on the board at all.
Then make it live with scorecards
Here's the part that determines whether any of this survives contact with the real world. A redesign that lives in a slide deck decays back into the old informal structure within a quarter. The thing that keeps it alive is a management cadence — and the instrument of that cadence is the scorecard.
A scorecard is simple on purpose. For each role or function: its core accountabilities, the three-to-five KPIs attached to them, the target for each, the actual, the trend, and the owner's name at the top. One page. Reviewed on a fixed rhythm — weekly for operating metrics, monthly for the lagging financial ones — in a standing meeting that doesn't get cancelled.
The scorecard does three jobs the org chart never could:
It converts a static design into an operating system. The accountabilities you defined on paper now get inspected on a cadence. The structure stops being a one-time event and becomes how the company runs.
It makes accountability gaps visible in days, not quarters. A red cell with a name next to it is a conversation that happens this week. The owner explains, commits to a corrective action, and you check it next cycle. Problems that used to surface at the board meeting now surface at the Tuesday standup.
It protects the design from drift. When a new outcome emerges or an owner leaves, you don't redraw the chart — you update the scorecard. The system absorbs change without a reorg every time the business shifts.
Color-code it — green, yellow, red — so the conversation goes straight to where attention is needed instead of marching through every line. The meeting isn't a status report; it's a management decision forum. Red gets a countermeasure and an owner. Green gets left alone.
The design loop
Put the three together and organizational design stops being an annual fire drill and becomes a discipline:
Define the work → assign single owners → attach KPIs they can move → manage it on a scorecard cadence → and iterate as the business changes.
The scorecard is the prototype you never stop refining. That's the design-thinking heart of it: you don't try to get the structure perfect on the whiteboard and then defend it for three years. You build a working version, put it under load, watch where it breaks, and adjust. The org chart is a snapshot. The scorecard is the living system — and the living system is what actually runs the company.
If your reorg produced a new chart but the same problems, you solved the wrong problem. Go back to the work.
