The Minimum Job-Search Record, and What Makes a Tracker Trustworthy

A job search falls apart quietly. Not because you stop applying, but because you lose the thread: which follow-up is due, which company went silent weeks ago, whether your interview rate is telling you to fix targeting or just apply more. This guide is for anyone running a search across many companies at once who wants a tracker that answers "what do I do next?" on demand. It sets a definition of done for that tracker - the fields it must hold, the status vocabulary that keeps it sortable, and the follow-up trigger that keeps it honest - so you could hand it to a friend and both of you would grade the same row the same way.

What is the minimum you must record per application?

The mandatory core is six fields: company name, role, date applied, application status, contact person, and follow-up date. Every serious source converges on this set, and none argues for fewer. Add four more when you can capture them cheaply - posting link, location, salary range, and source - because they cost seconds at log time and answer real questions later.

Think of the six-field core as five questions the tracker must always answer: where you applied, when you applied, who you contacted, what interviews are on your timeline, and what came of each one. If a row cannot answer those, it is not tracked, it is just remembered, and memory is exactly what fails at 100-plus applications.

The temptation is to build a beautiful spreadsheet with twenty columns. Resist it. Every column you add is a column you must fill, and unfilled columns are what make a tracker feel abandoned. The differentiator between a working tracker and a dead one is not schema - every guide lists nearly identical fields - it is whether you actually update it.

How big does the tracker get, and how long must it persist?

Plan for 100 to 200 rows over three to six months. That is the common total volume before landing a role, at a strategic cadence of two to three applications per day or 10 to 15 per week. At that scale the tracker is not a nice-to-have; it is the only thing standing between you and losing track of which follow-ups are due.

The search itself runs long, and the numbers are ranges, not a single figure. The median time to first offer was recently measured at 68.5 days, a 22% jump from the prior quarter. Employer-side time-to-hire averaged 42 days across industries, and that stretches by seniority.

The practical consequence: your records must persist for months, and rows will sit dormant for weeks before a company responds. That dormancy is the trap the follow-up trigger exists to catch, because a row you applied to 40 days ago looks identical to one from yesterday unless the tracker flags the age for you.

The tracker's real output is a ratio, not a list

The single most useful thing a tracker produces is your application-to-interview rate, not your applied-count. A raw count of applications says nothing about whether your search is working. What tells you is the fraction of applications that turn into interviews, measured after a real sample.

The base rates are sobering. Only about 2.4% of applicants reach the interview stage for any given role, which works out to roughly 42 applications per interview on average. Cold online applications convert at 0.1% to 2% into an offer. Against those numbers, "I applied to 60 places" is meaningless until you know how many produced a screen.

A raw applied-count says nothing. The interview rate after twenty rows is the diagnostic that tells you what to fix.

There is also a ceiling. Zippia's data shows candidates who apply for 21 to 80 jobs enjoy about a 30.89% probability of an offer, while those who exceed 81 applications drop to roughly 20%. More is not better past a point, because volume above the ceiling usually means worse targeting. So the review step is not busywork: after 20 to 30 applications, compute your rate. If it is far below average, fix targeting or fit before adding rows. If it is at or above average, more volume is a reasonable lever.

When should a follow-up flag fire?

Fire the flag at day 5 to 7 for startups and 1 to 2 weeks for larger companies. A single global threshold misfires, because the same silence means different things at different company sizes. Encode the trigger so overdue rows flag themselves - the strongest templates simply turn old application dates red - and never rely on remembering.

The survey evidence clusters in the one-to-two-week band, but there is real disagreement, and it maps to context.

Practitioners now favour the earlier edge for lean employers: wait 5 to 7 business days before the first follow-up on a startup, send a thank-you 24 to 48 hours after an interview, and treat two silent follow-ups as the signal to move on. At a large company, two weeks of silence is common and does not mean rejection - shortlists take longer to build, and closing that row early throws away a live opportunity.

The stop rule matters as much as the start. Send one follow-up at the flag, one more a week later, then stop. If there is still nothing by day ten after the first, move on. A second silent follow-up does not help, and a third reads as pushy to the exact people you are trying to impress.

The contact field is where the leverage hides

Log the person, not just the posting. The contact field is the highest-yield column in the tracker and the one most seekers under-fill. Referrals make up only about 7% of the applicant pool but account for 30 to 50% of successful hires, so a row that names a recruiter or a referrer captures your best odds, while a row with only company and role gives you nothing to act on.

This is also where geography changes the workload. In Refolk's index of professional profiles, there are 92,124 contactable recruiters and talent-acquisition professionals in the United States against 6,643 in the United Kingdom - a ratio of about 13.9 to 1.

The practical read: a UK seeker faces a far thinner outbound-recruiter pool, so each captured contact is worth more and the contact field should be weighted harder, with referrals prioritised. A US seeker has more recruiters to reach but the same base-rate problem, so the contact still needs to be logged the moment you have it. Finding the right named person is exactly the friction that eats an afternoon, which is where Refolk helps: describe the person you need and it returns real contacts to drop straight into the tracker.

One warning on the evidence. The famous "85% of jobs come from networking" figure is weak - it traces to a non-random LinkedIn poll of roughly 3,000 self-selected respondents, yet it is still the answer most search engines return. Referrals are genuinely powerful on better data, so log your contacts. Do not rebuild your entire search around a single unrandomised statistic.

Build the tracker: the procedure

The whole system takes about an hour to stand up and 20 minutes a week to run. The order below puts instrumentation before volume, because a pipeline you cannot read is a pipeline you cannot fix.

There is a real disagreement about step order worth naming. Volume-first sources push filling the pipeline to 10 to 15 applications a week before building instrumentation. Quality-first sources argue for tracking fewer, better-targeted applications, pointing at the 81-application ceiling. My position: build the instrumentation first. It costs an hour, and without it you cannot tell whether your problem is volume or targeting - which is the only question the argument is really about.

How this goes wrong: failure modes and false positives

Most broken trackers fail the same handful of ways, and every failure has a tell you can check for. The point of a standard is that you can grade a tracker against these without debate.

Stale tracker. Weekly batch updates instead of point-of-apply logging. The false positive is dangerous: rows look handled but statuses lag by days, so the follow-up trigger fires on wrong dates. Check: any row whose last-updated date precedes its last known event. Update latency, not schema, is the dominant failure across every source.

Free-text status. "Waiting", "maybe", "sent email" defeat sorting. The pipeline counts look fine but you cannot filter to "everything at Screen" or "everything overdue". Check: does every status cell match the fixed dropdown list, with no exceptions?

Silent ghosting misread as rejection. Two weeks of silence at a large company is common and often does not mean no. Closing a live enterprise row too early is a false positive that throws away a real chance. Check: company size against the follow-up window before marking a row dead.

Follow-up spam. Chasing at day 2 or 3 feels proactive and reads as pushy. Check: no follow-up logged before the day-5-to-7 floor.

Volume theatre. Panic-applying inflates the row count without moving outcomes; past 81 applications the offer probability falls to about 20%. The tell is a high applied-count next to a near-zero interview rate. Check: compute the application-to-interview ratio after 20 to 30 rows before adding more volume.

No contact captured. Only company and role logged, no person, when a name was available. A promising row with no one to follow up is wasted leverage. Check: any Applied-status row with an empty contact field.

"85% via networking" over-weighting. Building the whole tracker around referrals on the strength of a weak poll. Check: is the source a randomised study? Usually not.

Watch the mental cost too. In one survey, 72% of US job seekers said the search hurt their mental health and about 66% reported burnout. A tracker that reduces uncertainty - that always answers "what next?" - is partly a defence against that. Volume theatre makes it worse; a readable pipeline makes it better.

A status vocabulary you can copy

Fix the status list before you log a single row, because retrofitting one across 80 rows of free text is miserable. The list below is designed to sort cleanly and to make the "what next" obvious from the status alone.

Saved     - found it, not applied yet; next action is apply or drop
Applied   - submitted; follow-up trigger is now armed
Screen    - recruiter or phone screen scheduled or done
Interview - in a hiring-manager or panel loop
Offer     - offer received; next action is decide or negotiate
Rejected  - explicit no; closed
Ghosted   - two silent follow-ups passed the window; closed

Treat Ghosted and Rejected as different closes on purpose. Ghosted rows are the ones to review for source quality later - if one channel produces mostly ghosts, that is a signal about where to spend effort, and you only see it if you did not lump silence in with explicit rejections.

Verify before you call the tracker done

Run this checklist against your tracker before you trust it to tell you what to do next. If any item fails, the tracker is decorative, not operational.

How to keep the tracker current

A tracker is only as good as its freshest row, so the maintenance job is small and constant, not large and occasional. Do three things and it stays trustworthy for the length of the search.

First, log at the moment of applying, every time, no exceptions. This is the one habit that prevents the dominant failure. It takes under a minute and it is the difference between a follow-up trigger that fires on real dates and one that fires on guesses.

Second, run the weekly review as a fixed appointment. Read the pipeline for which sources are producing interviews and which companies have gone quiet, recompute your interview rate, and name the next action for every live row. Twenty minutes a week buys you a whole-search view that no amount of remembering can match.

Third, adjust after the sample is real. After 20 to 30 applications, your interview rate is meaningful. Below average means fix targeting - better-matched roles, warmer contacts, tailored materials - rather than adding volume. Tailoring each application to its posting is exactly the per-row work that drags at scale, and it is where Refolk removes friction: it writes the resume from your history, retunes it per posting, drafts the cover letter, and scores your fit, so the quality lever is actually pullable instead of aspirational. Keep the numbers, follow the ratio, and the tracker will always answer the only question that matters mid-search: what do I do next.