Save the Posting, Not the Link
Six weeks after you applied, a recruiter emails to arrange a call. You open the row in your tracker, click the link to remind yourself what the job actually was, and get a page saying the position is no longer available.
Now you’re preparing for a conversation about a role you can only half remember, and the one document that would have told you exactly what they said they wanted is gone. There is no archive you can rely on and no way to ask for it without looking like you didn’t read it.
This is entirely preventable and it costs about thirty seconds per application, at a moment when you already have the posting open in front of you.
Why the link breaks
Adverts are taken down when a role is filled, when a hiring freeze lands, when the advertising contract lapses, or when the board simply rotates old records out. Copies on other boards expire on their own schedules, so the aggregator version can outlive the employer’s or vanish first — see one vacancy, many listings for why the copies behave independently.
None of these events is announced, and all of them tend to happen during the exact window when you’re waiting to hear back. The link in your tracker is a pointer to something you do not control, which makes it a poor place to keep information you’ll need later.
What you actually need it for
Four things, in rough order of how often they come up:
Interview preparation. You want the requirements in the employer’s own words, so you can prepare examples against them rather than against your memory of them.
Checking a repost. The same role reappears two months later. Comparing the new text against the old one tells you whether the requirements moved — a seniority bumped, a technology added, a “nice to have” promoted — which is genuinely informative about why the first process didn’t conclude.
The compensation conversation. If a range was published, that range is a fact you may want to refer to. If it wasn’t, knowing it wasn’t is also useful.
Proving to yourself that two listings are one job. The dedupe check in never apply to the same job twice is much easier when you can compare descriptions rather than titles.
Capture fields, not just a screenshot
A screenshot is better than nothing and worse than text, because you can’t search it and you can’t paste from it.
What you want is the description text plus a handful of fields lifted off the top of it. Usefully, there’s a fairly standard set — job-posting data is exchanged between systems as structured records, and Serply’s Google Jobs endpoint describes postings returned with the employer, listed perks and remote flags among their fields. Those are the same attributes worth pulling out by hand, because they’re the ones that get lost, contradicted, or quietly reworded between one copy of a listing and the next.
A capture block that fits in a notes field or a linked doc:
Employer Northwind Ltd
Role (as
written) Senior Analyst (m/f/d)
Ref REQ-4471
Location Manchester
Remote "hybrid, 2 days on site" — their wording
Comp £52-58k published
Perks listed pension, 28 days, learning budget
Closing date none stated
Source URL <the listing you used>
Captured 2026-08-11
--- full description text below ---
The two that people skip and later want are remote and ref. Copy the employer’s exact wording on remote work rather than paraphrasing it, because “remote”, “hybrid” and “flexible” are each used to mean at least three different arrangements, and the version you’ll be held to is theirs. The reference number is what resolves a duplicate argument in one second instead of five minutes.
Where to put it
Not in the tracker itself. A spreadsheet cell holding two thousand words of job description makes the sheet unusable, which is how trackers get abandoned.
One document per application, in whatever notes tool you already use, named company-role and linked from the tracker row. The tracker keeps the row; the doc keeps the evidence. That’s the same split as the notes-per-conversation habit in the notes you take after every conversation — and if you’re already keeping one doc per process, the posting text is simply the first thing in it.
When to do it
At the moment you decide to apply, before you start writing anything. Not after you submit.
Two reasons. The obvious one: after submitting, you’re finished with the tab and you’ll close it. The less obvious one: reading the posting closely enough to pull five fields out of it is a better preparation for the application than skimming it, and you were going to have to read it anyway.
So it slots into the top of the applying block:
- Open the posting.
- Search the tracker for the company name.
- Create the row.
- Create the doc, paste the description, fill the five fields.
- Now start the application.
Steps 3 and 4 together take under a minute and they’re the only parts of this that require any discipline.
What not to save
Everything you binned. You triage far more postings than you apply to. Capturing all of them turns a useful archive into a landfill and makes the whole habit feel like admin. Capture at the point of applying, not at the point of finding.
Long, un-skimmed pastes with no fields. If you paste four screens of boilerplate and never pull the specifics out, you’ve deferred the reading rather than done it. The five fields are the part that earns the time.
Anything you’d be uncomfortable holding. Some postings include a named contact’s details. Keeping a person’s email in a doc for a live process is normal; keeping a directory of them long after is not, and it’s worth clearing out at the end of a search along with everything else.
The one-line version
The listing is temporary. Your record of it doesn’t have to be, and the difference between the two is thirty seconds spent while it’s still on screen.
It is also, quietly, one of the few parts of a job search where a small amount of work now removes a specific, predictable problem later — rather than the usual arrangement, where you do the work and find out nothing for three weeks.