Every application you send is an experiment: a hypothesis (“I'm a fit for this role”) tested against a real market. Most job seekers run dozens of these experiments and keep none of the results. Then, when the search stalls, they have nothing to diagnose with, so they fall back on the two default strategies: apply to more jobs, or rewrite the resume top to bottom. Both are guesses. A tracker with five fields replaces the guessing.
The five fields worth logging (and nothing else)
Elaborate 20-column spreadsheets die within a week. The discipline that survives is minimal:
| Field | Why it earns its place |
|---|---|
| Company + role | Identification, and so you can find the exact posting again before an interview. |
| Date applied | Everything downstream is a rate over time. Also prevents the embarrassing double-apply. |
| Match quality (e.g. your Yes Score) | The single most diagnostic field. It lets you separate 'my targeting is off' from 'my resume is off'. |
| Status | Applied → screen → interview → offer/rejected/ghosted. Four or five states, no more. |
| Resume version sent | Which tailored version went out. Without this you can't connect outcomes to choices. |
The three rates that diagnose a search
1. Response rate: surfaced or invisible?
Any human response (screen invite, rejection email written by a person, recruiter question) divided by applications. A healthy cold-application response rate runs roughly 10-25%. Far below that, your problem is upstream of your qualifications: findability (vocabulary, format) or targeting (wrong jobs).
2. Interview rate: the headline number
Interviews divided by applications, trailing 30 applications. Cold-apply norms run about 2-7%. Under 2%, fix the funnel before adding volume. Above 7%, your materials are working; spend your energy on interview prep and on applying to more of whatever you've been targeting.
3. Interview-to-offer: a different muscle
Offers divided by final-round interviews. If you reach interviews regularly but offers never land, your resume is done improving; the work is interview performance, and no amount of further tailoring will move it.
Two searchers, same effort, different information
Hover or tap an item for detail.
The monthly review (15 minutes)
- Compute the three rates over the trailing 30 applications.
- Split by match quality:what's the interview rate on applications where you covered 70%+ of requirements vs. the long shots? This one cut usually settles the “should I apply to fewer, closer jobs” question with your own data.
- Look at ghost patterns: if one job board or one role title produces only silence, stop feeding it.
- Pick exactly one changefor the next month: tighter targeting, a summary rewrite, more referrals. One variable per month, or you can't attribute the result.
Spreadsheet or tool?
A spreadsheet with the five columns above works, and working is what matters. The practical failure mode is entry friction: after the fifteenth application of a tiring week, the spreadsheet doesn't get opened, and gaps make every rate unreliable. Tools that log automatically remove that failure mode. (YesPile's tracker, for instance, records the company, role, date, match score, and exact resume version the moment you optimize for a job, so the log builds itself and you only ever update statuses.)