The recruiting dashboard most teams inherited has a “time to hire” number on it, expressed in days, watched like a stock price, and largely useless as a measure of how well the team is hiring.
This piece is the alternative. Five measures that predict hiring outcomes, ordered from leading indicators (predict the future) to lagging indicators (confirm the past). All five are concrete enough to instrument inside a normal ATS, including the one we make.
What “vanity metric” actually means
The label comes from outside hiring. Eric Ries coined the distinction in 2010 in Harvard Business Review: a vanity metric is one that makes you feel good but does not guide a decision. An actionable metric shows clear cause and effect, so when it moves you know what to do.
Time to hire is the textbook hiring vanity metric. It rises, it falls, and the team rarely knows why. Beneath the metric is a stack of roles that closed fast, closed slowly, got rescued in week six, or quietly stayed open while the dashboard moved on. The average makes for a clean number to put in a quarterly review. It does not tell you whether the next hire is going to work out.
The five measures below replace it. The first three are leading indicators: they move while a role is still in flight, so the team can act. The fourth is mixed: it summarises past hires by source, which informs where to invest the next sourcing dollar. The fifth is the lag: it confirms after the fact whether the previous four were telling the truth.
What time to hire actually measures
Mechanical issue first. Time to hire averages every closed role in a window. SHRM’s 2025 benchmarking report puts the global average around 44 days, climbing past 60 for technical roles. The headline number is real. Its usefulness is not. A team that fills one role in 8 days and another in 80 reports a 44-day average that describes neither role honestly. A team that filled three roles in 60 days each and let two open roles sit unfilled for six months reports a clean 60-day average and a hidden failure.
The metric flatters whichever shape the hiring process happens to take. It does not tell the team whether the process is good.
Why dashboards default to it anyway: it is easy to compute, the data is in the ATS already, the number is benchmarkable externally, and it reports comfortably up the chain because it sounds like progress whether or not it is. Once a team puts time to hire on the dashboard, the dashboard starts driving behaviour. Goodhart’s observation applies: when a measure becomes a target, it stops being a good measure.
The time-to-X family
Time to hire is the most visible vanity metric, but it has siblings. Time to fill (role posted to role filled) starts the clock earlier. Time to start (offer accepted to first day) starts it later. Time to productivity (first day to fully ramped) lives entirely past the hire. Each of these is occasionally useful as a sanity check after the fact. None of them should sit on the dashboard as a target. They all average across roles, they all reward rushing whoever happens to be available, and they all let a team that filled the easy roles look productive while the hard roles stay open. The fix is not to pick the right time-to-X. The fix is to measure something else.
1. Time to first qualified interview
Leading indicator.
The clock starts when the role goes live. It stops when a qualified candidate sits in front of the hiring manager for a real interview.
This measure isolates the part of the pipeline you most control (sourcing, screening, scheduling) from the part you do not, which is whether the candidate eventually accepts. It exposes the most common failure mode of SMB hiring, which is not that candidates take too long to decide. It is that the team takes too long to put a good one in front of the hiring manager.
In Join’s customer pipelines, this number is the single best predictor of whether a role closes well or drags. Roles that hit the first qualified interview inside 10 days tend to close on time. Roles that drift past 14 days tend to keep drifting, regardless of how strong the eventual candidates turn out to be. The 60-day bullets from the job description and the screening flow upstream are what make the 10-day mark hit-able; both upstream tools are covered separately.
If time to first qualified interview is over two weeks, the bottleneck is upstream of any candidate behaviour. It is almost certainly that the role sat in a pile, the screening pass is taking too long, or the hiring manager has not set aside calendar for first-rounds.
Target: under 10 working days for most SMB roles. Under 5 for high-volume roles.
2. Stage conversion rates
Leading indicator.
The percentage of candidates who advance from each stage to the next, tracked separately:
- Applicant → screened
- Screened → first interview
- First interview → second interview
- Final interview → offer
- Offer → accepted
Each of those rates tells a different story. A low applicant-to-screened rate usually means the job ad is wrong (too vague, too narrow, or attracting the wrong people) or the screening filter is throwing too much away; the knockout-questions-first approach tightens this conversion deliberately. A low screened-to-first-interview rate means the screening criteria are misaligned with what the hiring manager actually values. A low offer-to-accepted rate means the salary, the role description, or the experience the candidate had during the process did not match what they thought they were signing up for.
You do not need to optimise every rate. You need to know which one is the limiting step and work on that one.
Stage conversion also varies meaningfully by source. In Join’s multiposting customers, applicant-to-screened rates are systematically lower from Indeed (broadest pool, most marginal applicants) than from LinkedIn (narrower pool, higher baseline qualification). That variation is signal, not noise: treating boards as interchangeable is one of the more common SMB hiring mistakes.
Target: depends on the role, but the direction should be stable. A rate that swings widely between similar roles is the signal: something changed, or something was never working the way you thought it was.
3. Decision velocity
Leading indicator.
The time between a candidate completing their final interview and receiving a decision.
This one matters more than any other measure of process quality, and almost no SMB tracks it. The reason it matters is that strong candidates are weighing other offers, and the team that decides last loses them, even if the salary and the role were better. SHRM’s 2024 Talent Trends report puts candidate ghosting among the top three recruiting hurdles named by U.S. employers (~46% of organisations); ghosting concentrates in exactly this window, post-final-round and pre-offer.
A 48-hour decision keeps strong candidates engaged. A 10-day decision puts them in someone else’s pipeline. The fix is structural: write the interview scorecard before the final round, score before debrief, and front-load the reference call so it does not sit between the decision and the offer. In Join’s customer base, the bottleneck at this stage is almost always hiring-manager calendar discipline at the leadership level, not the candidate.
Target: under 48 hours for senior roles. Under one week for everything else. If you cannot hit those numbers, the problem is calendar discipline, and no recruiting tool will fix it for you.
4. Source yield by channel
Mixed indicator: built from past hires, informs the next sourcing decision.
Hires per channel per quarter, divided by applicants from that channel. Sometimes called “source of hire” or “yield by source”.
The version that matters is not “where did the applicants come from” but “where did the hires come from”. They are not the same. LinkedIn Talent Solutions’ research on source of hire, drawing on their internal data and a Harvard Business School field study, puts employee referrals and LinkedIn at the top of the yield ranking and traditional job boards near the bottom. The job-board pool is larger; the share of that pool that becomes a hire is smaller. SMB teams chasing volume on the broadest board are typically optimising for the wrong half of the funnel.
The Join multiposting vantage on this is direct: when customers stop chasing volume and start measuring hires-per-channel, the channel mix usually rebalances away from broadest-reach boards toward narrower, more targeted ones. The high-volume boards stay useful for specific role types, just not as a universal default. The country-by-country board comparisons get more useful once a team knows which boards produced hires last quarter.
Target: less about hitting a number, more about updating the sourcing mix every quarter on the basis of actual hires, not applicant volume.
5. First-year retention as a quality-of-hire proxy
Lagging indicator: confirms what the leading metrics predicted weeks earlier.
The percentage of hires from a given cohort who are still in role twelve months later, with the hiring manager’s after-the-fact rating of “would re-hire” as a secondary check.
Quality of Hire is the recruiting metric everyone wants and almost nobody measures. SHRM has flagged it directly as the most-requested, least-measured talent metric: only a fraction of organisations track it in a data-driven way. The pure version is a composite of performance ratings, ramp time, and retention measured at multiple intervals. Most SMBs do not have the data infrastructure for that. The practical proxy is much simpler: of the people you hired twelve months ago, how many are still here, and would the manager hire them again knowing what they know now?
The number to compare against: roughly 40% of all employee turnover happens in the first year, per Work Institute’s retention research. That share is the cohort where bad hiring decisions show up first. A first-year retention rate that lags the industry baseline is the lagging confirmation that the leading indicators were lying, somewhere upstream.
In Join’s customer base, the strongest cross-correlation we see is between decision velocity and first-year retention. Hires that came from rushed final-round windows (over 10 days from final interview to offer) are disproportionately the ones that show up in 90-day attrition. The leading metric warned about the lag months in advance; the lag confirms it.
Target: above industry benchmark for the role and locale. Check the cohort at 90 days for early-attrition signal, then again at 12 months for the full picture.
The five at a glance
The five measures, with their indicator type and target, are summarised below.
| Measure | Type | What it tracks | When it fails | Target |
|---|---|---|---|---|
| Time to first qualified interview | Lead | Role-live → qualified candidate in front of the hiring manager | Over 2 weeks signals an upstream bottleneck (pile, slow screening, no calendar) | Under 10 working days; under 5 for high-volume |
| Stage conversion rates | Lead | Advance rate at each step (applicant → screened, screened → interview, etc.) | Wide swings between similar roles | Direction stable. The change is the signal |
| Decision velocity | Lead | Final interview → decision | 10-day decision puts strong candidates in someone else’s pipeline | Under 48 hours for senior; under one week otherwise |
| Source yield by channel | Mixed | Hires per channel divided by applicants per channel | Volume from a channel does not match hires from that channel | Updated quarterly; sourcing mix follows yield, not volume |
| First-year retention | Lag | % of hires still in role at 12 months; manager “would re-hire” | Below industry baseline for the role and locale | At or above industry benchmark; 90-day check first |
What to put on the dashboard
If you keep one number, keep time to first qualified interview. It is the single best predictor of whether the rest of the process has space to work properly. Slow upstream means rushed downstream, and rushed downstream is how teams hire people they later regret hiring.
If you keep five, keep that one plus stage conversion rates (per-role), decision velocity (per-hire), source yield by channel (per-quarter), and first-year retention (per-cohort, 90 days first and 12 months later). Five numbers, with the right cadence for each, not five numbers updated daily.
If you keep time to hire on the dashboard at all, keep it as a sanity check, not a target. A team optimising for a low time-to-hire number will close fast on whoever is available, and “whoever is available” is rarely the right hire. The five measures above are what make the time-to-hire number trustworthy when someone asks, even though none of them is the time-to-hire number itself.
The point of measuring is not to drive a number down. It is to know where in the process the next bad hire is going to come from, and to fix it before they sign the contract. Five numbers that show you that, instead of one number that hides it.
