There is a moment, somewhere between forty and sixty people, when the way your company used to work stops working. It doesn't taper off. It snaps. One quarter you're running the place out of three group chats, two recurring meetings and the CEO's head — and the next quarter, the CEO can no longer hold the mental model, the all-hands has stopped being sufficient, and senior people are starting to do important things that other senior people don't know about. The founding team feels like they're managing the company by phone instead of building it.
This is the moment most companies reach for a framework. OKRs, EOS, Hoshin Kanri, the operating system du jour. The framework is often the right move. The mistake is implementing the framework without the operating system underneath it — which is what produces the strange outcome of a company with detailed OKRs and worse execution than it had six months earlier.
The operating system underneath a fifty-person company is five explicit decisions, each of which sounds simple and each of which has to be made deliberately or it defaults to the wrong thing.
The five decisions
One: what is your goal syntax, and who owns it? Choose one framework — OKRs, KPIs, OGSM, EOS Rocks, SMART. Not two. The framework matters far less than the consistency. Then name the human being who owns the goal system. Not "leadership" — a person, by name. Usually a COO, Chief of Staff, or, if you have neither, the CEO. Their job is to make sure every goal has a single named owner, an update cadence, and a structural connection to the work layer. Without an owner, the goal system rots within six months.
Two: how do you express the strategy in three priorities? The annual strategy resolves into exactly three strategic priorities. Not five, not eight, not the four-with-a-stretch-fifth that everyone keeps trying to negotiate. Three. For each priority, you need: the named outcome (what "done" looks like in twelve months), the single accountable owner, the bandwidth assigned in specific people and percentages, and the lead indicator you watch weekly. If you cannot write a priority's paragraph without invoking specific people and their available capacity, the priority doesn't really exist — and that discovery is the most valuable output of the exercise.
Three: how does the work connect to the goals? Every initiative — every multi-week directed effort that consumes meaningful senior capacity — links to one of the three priorities. If an initiative doesn't link, it's reclassified: either it becomes a fourth priority (which means one of the existing three gives up capacity) or it moves to a separately-capped operational layer. The connection is made at the initiative level, not the task level. Don't tag tickets. Tag the three-to-eight-week efforts where strategy and work actually meet.
Four: what's the operating cadence? Three rhythms, none optional at fifty people.
The weekly is a one-hour leadership meeting. Not status updates — the system already provides those. Agenda: what decisions are pending, what's stalled and needs intervention, what changed this week that changes the plan. If the meeting doesn't produce decisions, it's broken.
The monthly is a two-hour strategy review. Agenda: each priority's lead indicator, initiatives that have stalled or drifted, and one or two explicit prioritisation decisions. Output is always a decision, never just an update.
The quarterly is a half-day QBR. Did we close or move our three priorities? What did we learn? What are the next three? Output is the next quarter's three priorities, with named owners and specific bandwidth.
Five: what's the single system of record? One system holds the goals, the initiatives, the owners, the status and the risks. Not three systems with an integration. One. The test is simple: on any random morning, can the COO answer three questions without scheduling a meeting? What are our three strategic priorities? What work is moving each of them? Which are on track and which are at risk? If the answer requires three tools and a Slack thread, the system of record isn't doing its job, and the company will pay for it in drift.
What it looks like in week six
Here's the operating model in motion, six weeks in.
The COO opens the system Monday morning. Priority 1 (net revenue retention to 110%) has three active initiatives. Two are on track. The third is stalled because the engineer it depends on is sitting at 130% allocation across three priorities. The system has flagged the over-allocation. The initiative owner has been notified.
The Tuesday leadership meeting opens with the over-allocation. "We have a 130% on Priority 1 — do we pull capacity from Priority 3, delay the initiative, or descope it?" The team decides in fifteen minutes. The decision is logged. The system is updated. Three people leave the meeting knowing exactly what changed and why.
There is no Friday status deck. There is no Sunday-evening reconstruction by the CEO of what happened during the week. There is one decision, surfaced by the system on Monday, made on Tuesday, recorded by Tuesday afternoon.
That's not a goal framework doing its work. That's an operating model.
Why companies skip this
Most fifty-person companies don't build this because each individual decision feels too heavy. Picking one framework feels like closing options. Naming three priorities feels like rejecting people's pet projects. Asking for bandwidth specifics feels invasive. Consolidating onto one system feels like a months-long migration. The decisions get postponed individually, and the operating model never assembles.
What gets built instead is a half-implemented framework grafted onto the old informal model, plus a few new meetings to compensate. Six months in, the company concludes "OKRs don't work for us" — when what actually didn't work is that they were never deployed onto a real operating system.
The Vindaris view
A fifty-person company that commits to three real priorities, backed by named owners, real bandwidth, and traceable work in a single system, outperforms a fifty-person company with twelve goals across a dashboard every time. The framework is the easy part. The operating model underneath — the cadence, the system of record, the ownership discipline — is what makes strategy executable.
The framework is the language. The operating model is the grammar. You need both, and most companies are trying to write strategy with vocabulary alone.