The Hiring Momentum Score, to Prioritize, Watch, or Skip One Company

Deciding where to spend limited application effort is a judgement call you make one company at a time. This guide gives you a repeatable way to score whether a single company on your target list is expanding, holding, or contracting hiring in your specific function, using only public data, and to end on a clean verdict: prioritize, watch, or skip. It is for job seekers who have a shortlist and want to rank it by real momentum rather than employer brand.

Generic advice tells you to research a company and stops at Glassdoor reviews and the mission statement. That tells you nothing about whether the team you would join is growing right now. This is the upside read. It sits alongside the library's downside references - layoff risk and return-to-office reversal - but answers a different question: is this company adding headcount in my function, and does that make it worth my time?

Why raw posting counts lie, and what to score instead

Score the live-req count and its 30-day trend, not the advertised count. Two independent facts make a raw posting number nearly useless on its own. First, the ghost-job floor: a 2025 Greenhouse study found between 18% and 22% of all online job postings are ghost jobs, positions that do not exist or that employers are not planning to fill immediately. Second, the hires-per-posting ratio has halved since 2019, from 8 per 10 to 4 per 10, so a company can carry a high static count while almost no one is actually being hired.

Both facts point the same way. The advertised count is inflated by ghosts and decoupled from real hiring. What survives both problems is velocity: the change in verified, live reqs in your function over a fixed window. A rising trend is harder to fake than a snapshot, because it requires the company to keep opening and refreshing reqs on its own applicant tracking system (ATS) - the software like Greenhouse, Lever, or Workday that runs its careers page.

A static high posting count can hide a hiring freeze; a rising 30-day trend cannot.

So the momentum score has five inputs, in order of weight: 30-day velocity in your function, the share of open roles that are expansion rather than backfill, the live-req count after discounting ghosts, a stability flag from tenure, and context events like funding or layoffs held as notes rather than multipliers. The rest of this guide builds each input and then combines them.

The four signals that separate expansion from backfill

Read four signals to tell net-new growth from replacement. Practitioners distinguish net-new headcount from a replacement, leadership hires from bulk hiring, whether one function grows while others freeze, and whether roles cluster or stand alone.

• Net-new vs. replacement. A backfill replaces someone who left and inherits existing relationships; an expansion hire is net-new headcount and carries a stronger signal.
• Leadership vs. bulk. A single VP or director hire in your relevant function is a strong signal that a team is being built out rather than topped up.
• Function vs. company. A company can hire in one function while freezing or cutting others. Verify your specific team is growing before you assume company-wide budget expansion.
• Clustering and first-time titles. Multiple related reqs opening together, and titles the company has never posted before, read as expansion. A lone recurring title reads as backfill or evergreen.

The word "evergreen" matters here. An evergreen listing is a role kept perpetually open to build a pipeline, not to fill an immediate seat. Analyst roles are disproportionately used this way because backfills for them are predictable. When you see one analyst or support req that reappears every few months, do not count it as demand.

Each signal tells you something and each can lie. A cluster proves coordinated intent, but a cluster of the same evergreen title proves nothing. A first-time title proves a new capability is being built, unless it is a rebrand of a role that was closed last month. A leadership hire proves a team is being resourced, unless that leader is a like-for-like replacement for someone who just left, which turns it back into a backfill.

How funding and context events should move the score

Weight funding cautiously: a round is permission to hire, not proof of hiring. The evidence is blunt. Top-quartile fundraisers grew headcount about 6% over twelve months versus about 12% for all other companies, and the difference is not statistically significant. The mechanism is that money now buys fewer people: Series A startups raised about $160k per employee in 2020 and more than $320k in 2025, and the median Series A headcount fell from 57 employees in 2020 to 44 in 2024.

That does not mean funding is noise. It means a round moves a company from skip to watch, and only confirmed reqs move it to prioritize. The counter-case shows what conversion looks like: one company raised $400 million and announced plans to grow headcount 50% to 450. That is a round paired with a posted, quantified plan - watch it become reqs, and if they appear in your function, prioritize.

The macro backdrop sets your "normal." On a three-month moving average basis, hiring held flat at 3.3% for six straight months, and JOLTS showed 6.5 million openings with an openings rate of 3.9%, the lowest in more than eight years outside March and April 2020. Against a soft market, a company sustaining velocity near the practitioner ceiling of 30% quarterly headcount growth is genuinely surging; flat postings are merely steady, not weak.

Reading tenure as a stability signal without the age trap

Tenure reads as churn only after you net out age and sector. Median tenure with a current employer was 4.1 years in January 2026, up from a low of 3.9 years in January 2024. But the headline hides the trap: workers aged 55 to 64 had 9.6 years of tenure, more than three times the 2.7 years for workers aged 25 to 34. A young, high-growth team therefore looks unstable on raw tenure without anyone actually leaving.

Sector matters just as much. Use this baseline before you flag anything.

The read: a fintech team at 2.5 years against a financial-activities baseline of 4.7 is worth a churn flag; a startup engineering team at 2.5 years staffed largely by people in their late twenties is not, because it matches the age-adjusted expectation. What tenure proves is retention over time. What it lies about is a young team, which looks like churn and is not, and a team that just doubled, which mathematically drops median tenure toward zero regardless of anyone quitting.

To settle the question, look at who left rather than at the median alone. A short search of departures from your target function over the past year, paired against the tenure figure, separates a young team from a leaking one.

Discounting false positives to a live-req count

Cut ghosts before you count. The documented discount takes two minutes: open the company's own careers page and search for the role title. If a job appears on LinkedIn or Indeed but is absent from the company's Greenhouse, Lever, or Workday portal, it may be stale, already filled, or cached by an aggregator. That single check removes most of the ghost floor.

Prevalence varies by study, which is why you verify rather than trust any single number.

After the careers-page cross-check, apply two secondary flags. A posting more than 30 days old with no visible activity is suspect, given the average US posting fills in about 41 days, and 25% of postings sit up more than 90 days. And the same role reappearing every few months usually points to an evergreen pipeline listing. Neither flag is a verdict on its own. There is no universal number of days; an older posting can still be active, while a recent repost can still be uncertain. Use age to lower confidence, not to auto-reject.

Public tools help you see trend without doing all the counting by hand. TrueUp aggregates hundreds of thousands of tech job postings and visualises company hiring trends, team sizes, and role types by headcount, funding stage, and growth rate. LinkedIn employee count is the most direct headcount-direction proxy. For the downside cross-check, layoff trackers cover confirmed cuts, though they miss companies below the WARN Act's 100-employee threshold and quiet attrition where roles simply are not backfilled. Whatever the aggregator says, the company careers page is the ground truth, because aggregators cache stale roles.

The scoring procedure, start to verdict

Run these eight steps in order on one company. Budget about 90 minutes the first time and 20 minutes for each weekly re-check once your search string and snapshot exist.

Turning the inputs into a verdict is a two-variable call: how fast the function is hiring, and how much of that hiring is real expansion. The matrix below is the decision.

Record the verdict in a fixed format so re-checks are comparable week to week. Copy this and fill one per company.

Company:
Function / title family:
Snapshot date:
Live reqs in function (careers page): ___  (prior snapshot: ___)
30-day velocity: ___%   Direction: rising / flat / falling
Expansion share: ___ of ___ live reqs tagged expansion
Leadership hire in function? yes / no
Tenure read: ___ yrs vs. sector baseline ___ yrs  Flag: stable / young / churning
Context notes (funding, earnings, layoffs):
Ghosts removed: ___ postings dropped after ATS cross-check
VERDICT: prioritize / watch / skip
Reason (one line, with the two numbers that drove it):

How this score goes wrong

The score fails in predictable ways, and every failure has a check. This section is the one to keep open while you work, because a false positive here costs you weeks of misdirected applications.

Two of these deserve extra weight. The funding halo is the most common misread on a startup shortlist, because a raised round is loud and public while the conversion into reqs is quiet. Anchor on the insignificant 6%-versus-12% headcount gap: a round without posted reqs in your function is a watch, full stop.

The tracker blind spot is the most dangerous, because it fails silent. Hiring freezes and quiet non-backfilling produce zero tracker entries, so a clean layoffs record can sit over a contracting team. The truer contraction signal is a clean tracker paired with a declining live-req count over successive snapshots. This is exactly why your own dated counts matter more than any single lookup: only a trend you recorded yourself catches the quiet freeze.

One more read to get right: recruiter silence. A recruiter not replying is not evidence a role is inactive; it simply adds no new information. Treat non-reply as neutral and let the req counts decide.

That leverage point is worth pricing into your own aim, too. Where the pool is thin, your application competes against fewer people, and a company's visible commitment there is more credible.

When you have narrowed to a live, clustered target, tailoring your application to each posting and scoring your own fit is where Refolk removes the manual work the score just justified, turning a prioritized company into submitted, tailored applications rather than a spreadsheet row.

Keeping the score current

A momentum score is a snapshot, so schedule the re-check. Velocity is only meaningful across dated counts, which means the first run creates the baseline and every weekly run creates the signal. Set a recurring 20-minute slot per prioritized company: re-pull the careers-page count, note the new 30-day velocity, and update the verdict if the direction flipped.

Re-run the whole eight-step procedure only when something structural changes: a funding announcement, an earnings call, a layoff-tracker entry, or a leadership hire in your function. Between those events, the weekly count update is enough. A prioritize can decay to watch in a month if velocity stalls, and a watch can graduate to prioritize the week a funded plan turns into clustered reqs. The point of the score is not a one-time ranking; it is a live read you keep accurate so your limited application effort always lands on the company that is actually building the team you want to join.