Should you do the hardest thing first?
The short answer
Sometimes. The advice is stated as a universal rule and it is not one, because two well-understood ordering disciplines answer it differently and both are correct for their own objective.
Shortest job first minimises average waiting time — clear the small things and the queue drains fastest. Priority scheduling respects importance over size — do the consequential thing regardless of how long it takes. Neither dominates, and no ordering rule is best in general.
So the real question is not which is right. It is what today is short of.
When hardest-first is right
When the hard thing is the day's actual output. If one item is what the day is for, doing it first is not a discipline exercise — it is protecting the thing that matters from a day that will otherwise fill up. Everything else will still be possible at four o'clock. That will not.
When it has no deadline and the small things do. Dated work keeps arriving, and undated work never reaches the front on its own. That is starvation, and putting the important thing first is the manual correction for it. The mechanism is here.
When you keep not doing it. A thing avoided for a week does not get more likely at 4pm. If it has survived several days of being scheduled later, later is demonstrably not working.
When hardest-first is wrong
When the hard thing is blocked and you have not checked. This is the most expensive version of the mistake. You commit the best hours to something that needs an answer from someone else, discover it at eleven, and have now lost both the morning and the small things you could have cleared. Thirty seconds of checking prevents it.
When the queue behind it is jammed. Strict order has a named failure — head-of-line blocking — where one long item holds up everything queued behind. If six people are waiting on six short replies, making them wait on your four-hour task is a choice with a cost, and it is usually not the one you would make deliberately.
When you genuinely cannot start it. Hardest-first is advice about ordering, and if the obstacle is that no first action has been named, ordering does not touch it. That is a different problem with a different fix.
When "hardest" means longest rather than most important. These get conflated constantly. A four-hour task is not important because it is four hours. If the hard thing is merely large, shortest-job-first has the better argument: clear the small ones, shrink the queue, then go.
The version worth actually using
Ask two questions before the day starts, in this order:
- What is this day for? Usually there is one answer. That item goes first, whether or not it is the hardest.
- Is it blocked? If yes, the real first task is the unblocking one, and it is probably short.
If those produce nothing — no single thing the day is for, nothing blocked — then the day's problem is volume rather than order, and the useful move is a limit rather than a sequence: three things and no fourth.
Why the universal version persists
Because it is memorable, and because it is right often enough to survive.
Most days do contain one thing that matters more than the rest, and most days do fill with smaller work if nothing protects the important item. As a default it beats no rule at all — and arrival order, which is what you get with no rule, is the one discipline known to be bad for essentially every objective.
But a default is not a law, and the cost of applying it on the wrong day is a wasted morning and a blocked queue. Checking takes half a minute.
Where to go next
- Is there a best prioritisation method? — the ordering rules and what each is optimal for
- Why your important work never gets done — starvation and the standard fix
- The Top 3 method — when the problem is volume rather than order
Using this in Decide One
Decide One asks what today looks like before offering a method, because the answer to this question changes with the day. That is the reason it ships three methods rather than one rule.
A line you did not finish is unfinished. It is not marked against you, and a day you did not open is not counted.
Open Decide One — free to use, with no account. Nothing you write leaves your device.
Decide One is a priority instrument. Bring the day into view, choose a method, give the first thing real time.