Decision Latency: Why Your Company Got Slower as It Grew
Every founder past about fifty people notices the same thing: decisions that once took a day now take three weeks, and nobody can explain where the time goes. The instinctive diagnosis is cultural — the company has lost its edge, people are too comfortable, we need to reinstate urgency. That diagnosis is almost always wrong, and acting on it makes things worse. What you have acquired is decision latency, which is structural, measurable, and fixable in specific ways depending on where it comes from.
What’s inside
Latency is structural, not cultural
The five sources
How to measure yours
The fix for each source
Why demanding more urgency backfires
The reversibility rule
Where Blomma fits
Latency is structural, not cultural
At ten people a decision takes a day because the three people who need to agree are in the room and each holds the full context. Nothing about their urgency is special; the conditions make speed free.
At a hundred people, the same class of decision needs input from people who are not in the room, who hold different fragments of context, whose priorities conflict, and none of whom has clear authority to settle it. The decision now has to travel. Every hop costs a scheduling delay, a context transfer, and a chance of the whole thing being deferred pending more information.
That is latency, and it is a property of the structure rather than the people. The same team that made fast decisions at ten will make slow ones at a hundred if the structure requires more hops. Which is why exhortation does not work — you are asking people to overcome a routing problem with effort.
The useful reframe: treat speed as a design output. If decisions are slow, something in the structure is adding hops, and the hops are findable.
The five sources
Five causes, each with a distinct signature.
Unowned decisions.Nobody has clear authority, so the decision becomes a negotiation about who decides as well as what to decide. Signature: long threads about process and being looped in, rather than about the substance. The most common source by a wide margin.
Too many consultees.The decision is owned, but the owner feels obliged to consult eight people, each of whom takes a week to respond. Signature: the owner can name the decision-maker but the calendar is the bottleneck. Usually a symptom of an owner who does not feel safe deciding.
Missing context.The decision keeps stalling because someone needs information that lives in one person’s head or in nobody’s. Signature: repeated “let me check with” and decisions that get made and then unmade when a constraint surfaces late. Underrated and cheap to fix.
Escalation to a single point.Everything contested ends up on one calendar — usually the founder’s. Signature: decisions queue in a visible way, and the queue length correlates with one person’s travel schedule. This is the source founders most often cause and least often see.
Fear of being wrong.The decision is owned, informed, and unmade because the owner has learned that a bad call has consequences and a slow call does not. Signature: excessive analysis, requests for more data, and decisions that arrive with elaborate justification. This one is genuinely cultural, and it is downstream of how you have responded to past mistakes.
How to measure yours
Unmeasured latency gets debated. Measured latency gets fixed. Three approaches, in increasing effort.
Pick ten recent decisions and time them.From the moment the question was first raised to the moment it was settled and communicated. Note where each one waited. This takes an afternoon and is usually sufficient to identify your dominant source — most companies have one that accounts for the majority.
Ask your leads what they are waiting on.In writing, unsoftened. The list is your latency, itemised. Ask specifically what they are waiting on from you, because that is the item people minimise in conversation.
Instrument a decision log.For significant decisions, record when raised, who owns it, who was consulted, when settled. Three months of this gives you a real distribution rather than anecdotes, and it makes the trend visible as you grow.
One thing to look for in all three: the difference between time spent deciding and time spent waiting. Almost always the waiting dominates, which tells you the problem is routing rather than deliberation.
The fix for each source
The fixes are specific, and applying the wrong one is why “we need to move faster” initiatives fail.
For unowned decisions: write down decision rights.A short list of the decisions that matter and, for each, who decides, who is consulted, and what threshold escalates. It reads bureaucratic and it removes the largest single source of latency, because it converts a question of authority into a question of substance.
For too many consultees: cap the consultation and name it.Two or three consultees, with a deadline — input by Thursday or the decision proceeds without it. And make it explicit that the owner is expected to decide without unanimity, because most over-consultation is an owner protecting themselves.
For missing context: write things down.The reasoning behind past decisions, the real constraints, what the board actually cares about. Unglamorous and the highest return per hour available. A few hours of writing removes weeks of downstream stalling.
For single-point escalation: redesign the escalation paths.For each recurring conflict, ask where it resolves. Any conflict that resolves only at your desk needs either an assigned owner or an explicit rule. Then hold to it when you disagree with an outcome, which is the actual test.
For fear of being wrong: change what happens after a bad call.This is behavioural and it is yours. If a reversible decision that went badly resulted in the decision being taken back, you have taught the organisation that speed is punished. The fix is visible, specific tolerance for reversible mistakes — and it takes months to be believed, because people are going by evidence rather than announcements.
Why demanding more urgency backfires
The standard founder response to slowness is to raise the temperature: more urgency, tighter deadlines, visible impatience. It reliably makes latency worse, for three reasons worth understanding.
It converts latency into risk avoidance. Under pressure with unclear authority, the rational move is to escalate rather than decide — because escalating is defensible and deciding is not. So the queue on your desk lengthens.
It removes the pause where the real problem would surface. A team that is slow because context is missing needs someone to notice the missing context. Urgency makes them proceed on assumptions instead, which produces rework that costs more than the original delay.
And it degrades the signal. If everything is urgent, nothing is, and your organisation loses the ability to tell a genuine priority from your general impatience. That cost compounds and it is difficult to reverse.
The alternative is unglamorous: find the source, fix the routing, and let speed be a consequence. Companies that are genuinely fast at scale are not more urgent. They have fewer hops.
The reversibility rule
One principle worth adopting explicitly, because it resolves a large fraction of latency on its own.
Sort decisions by how hard they are to undo. For reversible decisions, the standard should be speed: the owner decides, quickly, with limited consultation, and you review the outcome rather than the decision. For genuinely irreversible ones — a market entry, a senior hire, a platform commitment, anything with a long tail — slow is correct and the deliberation is not latency.
Most companies apply uniform care to both categories, which means they are simultaneously too slow on the reversible majority and insufficiently careful on the irreversible minority.
Say the rule out loud and apply it visibly. The first time someone makes a fast reversible call that goes badly and you respond by asking what they learned rather than taking the decision back, you have changed the operating speed of the company more than any process would.
Where Blomma fits
The uncomfortable part of this diagnosis is that two of the five sources are usually caused by the founder — single-point escalation, and the fear of being wrong that follows from how past mistakes were handled. Those are the hardest to see from the inside, and the people best placed to tell you report to you.
Blomma is an always-on AI career coach with no stake in your company. Use it to run the ten-decision audit honestly and identify your dominant source rather than the one you would prefer. Use it to work out whether your escalation paths quietly end at your desk, and what it would actually take to move two of them. Use it to prepare the conversation where you hand a decision to someone and commit to not taking it back — and the harder moment three weeks later when it goes sideways and every instinct says intervene. Where the situation warrants a human who has designed at your stage, bring one in..
Your company did not lose its urgency. It acquired hops, and hops are a design choice you can revisit. Find where the waiting happens, fix the routing, and speed comes back on its own.