In short
Too many simultaneous priorities produces slower delivery of everything and completion of almost nothing, because each item accrues switching cost and none accumulates enough attention to finish. The fix is a hard limit on work in progress, expressed as a number of active initiatives, plus an explicit decision about what stops. Ranking is not enough: a list of nine ranked priorities still has nine things running. Something has to be moved to not now, in writing, with the reason recorded.
Key takeaways
- Ranking without stopping changes nothing. The list must get shorter, not better ordered.
- Everything running at once means everything finishes late, and half of it never finishes at all.
- Say 'not now' in writing with a reason. Unstated deprioritisation is heard as a rejection or ignored entirely.
- Sunk cost keeps dead initiatives alive. What has been spent is not a reason to continue.
- One finished change beats three in progress, because only finished changes produce results.
The cost of running everything at once
A business with nine active initiatives is not moving nine times faster than one with a single initiative. It is usually moving slower on all nine, and the reason is arithmetic rather than motivation.
Three costs stack up. Each item consumes reorientation time whenever anyone returns to it, and that cost is paid every time, not once. Each item requires coordination, so overhead grows faster than the number of items. And unfinished work delivers nothing at all, which means that at any given moment, a business running nine initiatives at a ninth of its attention has produced exactly zero results and nine partial ones.
The uncomfortable version: three initiatives done sequentially finish at months two, four and six. The same three run in parallel finish at months six, six and seven, if they finish. The parallel version delivers its first result four months later and is far more likely to have one abandoned midway.
Parallel work does not divide the time. It multiplies the coordination and defers every completion to the end.
Ranking is not prioritising
Most prioritisation exercises produce an ordered list and change nothing, because ordering is not the operative decision. If all nine items remain active, the ranking describes preference while the capacity is still split nine ways.
Prioritising means something specific and uncomfortable: some things stop. Not deferred vaguely, not deprioritised in a document while continuing quietly, but stopped, with the decision recorded.
| Looks like prioritisation | Actually is |
|---|---|
| A ranked list where everything continues | A preference order, with unchanged capacity |
| “Focus on the top three” with six others running | Nine initiatives, three of which get slightly more attention |
| Deprioritising without telling anyone | An initiative that continues on goodwill and produces nothing |
| Three active, six explicitly not now | Prioritisation |
The third row is the most damaging in practice. An initiative that has been quietly deprioritised does not stop consuming resources: someone is still attending meetings about it, someone is still being asked for updates, and the person doing the work still believes it matters. All the cost, none of the output.
Set a number and hold it
The most reliable mechanism is a limit chosen in advance, before anyone is arguing about a specific initiative. Numbers chosen while looking at a list get bent to fit the list.
For most businesses under fifty people, two or three active operational initiatives is realistic, alongside delivering the actual work. Anything above four in a business of that size means either the initiatives are trivial or several are notional.
Active initiatives is a realistic ceiling for a small business that also has clients to serve. The number people usually want is six, and the number of those that reach a measurable outcome is rarely more than two.
Then hold it mechanically: nothing new starts until something finishes or something is stopped. This is the same logic that makes work in progress limits effective at a constrained step, described in workflow bottlenecks, applied to the initiative level rather than the task level. It works for the same reason: finishing is what produces value, and capping concurrency is what makes finishing happen.
Choosing what makes the cut
With a limit set, the selection question becomes tractable. Four tests, applied in order, resolve most lists.
- Is it upstream of the others? An initiative that removes the cause of two other items on the list wins regardless of its own size. Documentation frequently sits here: it is dull, it is upstream of onboarding, automation and resilience simultaneously.
- Would you start it today? Ignore what has been spent. If the answer is no, it belongs in the not-now column even if it is half built.
- Can it finish inside a quarter? Initiatives longer than that need splitting into something that produces a result, or they will be overtaken by events.
- Is there a named owner with actual capacity? An initiative owned by someone with no hours is not active, it is aspirational, and it occupies a slot that something real could use.
The first test is the one that most often changes the answer, and it is the one that instinct gets wrong. Instinct ranks by pain, which selects the loudest symptom. Causal position selects the thing generating the symptoms, which is usually less urgent-feeling and worth considerably more. That distinction is the whole basis of how an operations audit orders its findings.
Stopping things that are already running
Stopping is harder than not starting, for reasons that are psychological rather than analytical, and it needs to be handled explicitly or it does not happen.
Sunk cost. The months already spent feel like an argument for continuing. They are not. They are gone either way, and the only question is whether the remaining effort is the best available use of that capacity. The person who has invested most in an initiative is the person least able to assess this, which is not a criticism, it is a reason to have someone else ask the question.
The visibility problem. Stopping something looks like failure in a way that never finishing it does not. A business where quiet abandonment is safe and explicit stopping is embarrassing will accumulate a long tail of undead initiatives, each consuming a little attention forever.
The practical remedy is to make stopping a normal, recorded event with a stated reason:
- Say it out loud, to everyone who was involved. Silence is heard as continuation.
- Record why, in a sentence. The reason matters when someone proposes it again in eight months.
- Capture what was learned or built, so the effort is not entirely lost.
- Free the capacity explicitly and say where it went. Otherwise it evaporates.
- Stopping rule
- A condition, agreed before an initiative starts, under which it will be halted: a date, a result not achieved, a resource no longer available. Agreed in advance, stopping is the execution of a plan. Agreed afterwards, it is a judgement on the people involved, which is why it rarely happens.
Holding the line when new work arrives
The limit will be tested within weeks, usually by someone senior with a good idea. Refusing is the wrong response and so is agreeing.
The third option is to make the trade visible. Show what is currently active, state that the new item can start when one finishes or in place of one, and ask which. This is not obstruction: sequencing is precisely the decision that person is entitled to make, and most of the time, seeing the three current items causes them to choose to wait.
What this prevents is the silent version, where the new item is added, nothing is removed, and the limit becomes decorative within a month. A limit that bends under the first bit of pressure was never a limit; it was an intention, and intentions do not survive a busy quarter. That is the same failure pattern described in why process changes don't stick, and it has the same remedy: decide in advance what happens the first time someone asks for an exception.
The Mayim Ops assessment is built to produce one first move rather than a list, precisely because a business that receives twenty recommendations will attempt all of them at a twentieth of the attention each, and finish none.
Frequently asked questions
How do you prioritise when everything is a priority?
Set a hard limit on how many initiatives can be active at once, usually two or three for a small business, and force the list to fit the limit. Ranking alone does not help, because a ranked list of nine still has nine things running and consuming attention. The decision that matters is what stops.
What is a work in progress limit?
A cap on how many items may be active at the same time. Counterintuitively, lowering it usually increases how much gets completed, because each item finishes sooner and less capacity is lost to switching between half-finished work.
How do you decide which projects to stop?
Ignore what has already been spent and ask what each initiative would be worth if you were starting it today, then whether it is upstream of the others. Stop the ones that would not be started now, and say so explicitly rather than letting them fade, because unstopped initiatives keep consuming attention.
Why does doing fewer things at once produce more output?
Because switching between tasks carries a real cost in reorientation, and because unfinished work delivers no value at all. Three initiatives at a third of attention each finish later than the same three done sequentially, and the sequential version starts producing results from the first completion.
How do you say no to a new priority from a senior stakeholder?
Do not say no; make the trade visible. Show what is currently active, state that the new item can start when one of them finishes or in place of one of them, and ask which. This converts a refusal into a sequencing decision, which is the decision that person is entitled to make.