How Much Should Be In Flight at Once
There’s a particular kind of bad week that happens to organised job searchers. Two processes were going well, both ended in the same fortnight, and now there is nothing. Not one live conversation. And the reason is not that anything went wrong — it’s that while those two processes were absorbing all the attention, nobody was filling the top of the pipeline.
The metric that would have caught it isn’t applications sent per week. It’s coverage: how much is currently live at each stage, and whether that’s enough to produce an outcome in the timeframe you need one.
The pipeline has stages and each has a lag
Your tracker already has the stages — Found, Applied, Screening, Interviewing, Offer. The thing that makes coverage non-obvious is that each transition takes weeks, and the delays compound.
Suppose, in your field, applications take about two weeks to produce a screen, a screen takes another two to become an interview loop, and a loop takes three more to reach a decision. Those are illustrative numbers to make the arithmetic legible, not claims about your market — you can pull your own from the date columns in your tracker once you have a dozen completed rows.
With that shape, something you apply to today resolves in roughly seven weeks. Which means the emptiness you feel in the second week of August was created by what you didn’t do in the last week of June. The pipeline reports on your past self, and it does so with a delay long enough that you can’t feel the mistake while you’re making it.
That’s the whole case for measuring coverage instead of activity. Activity tells you what you did this week. Coverage tells you what’s going to happen.
A worked example
Say you want one offer, and you’re using these hypothetical conversion rates — again, placeholders you should replace with your own:
- 1 in 6 applications produces a screen
- 1 in 3 screens produces an interview loop
- 1 in 3 loops produces an offer
Working backwards from one offer: you need about 3 loops, which needs about 9 screens, which needs about 54 applications. Spread over the seven-week resolution window, that’s roughly 8 applications a week.
Now do the same arithmetic with numbers from a different situation — a strong internal referral channel where 1 in 2 applications produces a screen — and the requirement drops to about 18 applications for the same result. Same target, a third of the volume.
That contrast is the point of the exercise. The number of applications you need isn’t a property of job searching, it’s a property of your conversion rates, and the highest-leverage move is usually to improve a conversion step rather than to raise the input. Working out which of your steps is weakest is the subject of how many applications is the right number.
The coverage check, five minutes a week
Run this in your weekly review block. Look at your live-processes view and your Applied rows and write down four numbers:
COVERAGE — week of 22 June
Live processes (Screening/Interviewing/Offer): 3
Applied, still inside my ghost cutoff: 11
Sourced but not applied (Found): 6
Applications sent in the last 14 days: 7Then ask one question of each:
Live processes. How many will still be live in three weeks? A process at final-round stage is nearly resolved, which means it’s about to leave the pipeline whether it succeeds or not. Two processes both at final round is not two units of coverage — it’s a fortnight of coverage.
Applied and inside the cutoff. This is your real forward supply. If it’s near zero, you have a quiet month coming and nothing you do in that month will change it.
Found but not applied. This one is a warning light for a specific failure: sourcing is happening and applying isn’t. A Found pile that keeps growing means the applying block is being eaten by something else.
Recent sent volume. Compare against the rate your own arithmetic says you need. Not against anyone else’s number.
The trap: one process that’s going well
Every experienced searcher has done this. A process starts going well. There’s rapport, the work sounds good, they’ve mentioned next steps. And you quietly stop applying, because it feels a bit disloyal to keep looking and mostly because the anxiety has switched off.
Then it doesn’t happen — sometimes for reasons that have nothing to do with you, like a headcount freeze two days before the offer — and you restart from zero with six weeks of dead time ahead of you.
The rule that prevents it: maintain the input rate until you have signed something. Not until it feels likely. Not until they say you’re the leading candidate. Until it’s signed.
This isn’t cynicism about the specific process. It’s that the cost of the two extra applications a week is small, and the cost of an empty pipeline is a month of your life. Keep the sourcing and applying blocks running at a reduced rate if you must — but running.
When coverage is too high
The opposite state is real, and it looks like this: eleven live processes, four of which you can’t fully recall, and the preparation time each deserves divided by eleven, which is not enough for any of them.
Coverage has a ceiling set by attention, not by ambition. A live process needs real hours — scheduling, preparation, thinking about it in advance — and once you’re past what you can give, adding another one lowers your conversion on all of them. Beyond that point the correct move is to stop applying for a fortnight and work the processes you have. That is what running several live processes at once is about.
Most people are far more often in the empty state than the overloaded one. But if you notice that every process is getting a rushed, thin version of you, the pipeline is over-filled, and more volume is actively making things worse.
The one number, if you only keep one
Not applications sent. Not response rate. Conversations currently live, plus applications still inside your cutoff.
That single figure tells you whether the next six weeks contain anything. Write it on the top row of your tracker each week and watch the trend rather than the level, because the trend is visible weeks before the consequences are — and the whole reason to run a search as an operation is to see the quiet month while there’s still time to prevent it.