The Hiring Season Curve for One Role and Metro, Month by Month

You are trying to decide which weeks to concentrate your applications for one specific role in one specific city. This guide is for a job seeker who is tired of "best month to apply" listicles that assume everyone is a Q1 office hire, and it delivers a repeatable method: build a month-by-month posting curve from public data, separate a real seasonal peak from a one-off spike, and turn the peak into a dated ramp-and-coast calendar. I carry one worked example the whole way through, including the wrong turns.

Why generic "best month" advice misleads you

The generic answer is wrong because the peak belongs to your occupation, not the calendar, and the gap between them can reach seven months. Education job ads stay below their annual average until May and then pick up in August, while office roles cluster in Q1. A teacher who follows generic advice applies roughly seven months out of phase.

The popular numbers are also not primary data. The widely repeated "January +15% postings" and "Q1 +22% activity" figures trace to secondary blogs, not to any government release. Meanwhile the platform with the largest posting dataset now contradicts that story: LinkedIn's labor-market seasonality research finds that job postings tend to peak from April through July across many markets, and May is often the highest posting month depending on country and market. So the single most-cited peak is disputed by the largest single source of postings.

The mechanism is simple. Hiring tracks the customer's operating cycle, not the fiscal-budget January reset. Schools hire for an academic year. Retail hires for a holiday season. Hospitals hire continuously because patients arrive continuously. Any curve that ignores that is fitting one industry's shape to everyone.

The peak belongs to your occupation and metro, not to a calendar month everyone else copied from a blog.

The one example I carry through

I will use one worked case for the whole guide: a registered nurse in the Dallas-Fort Worth metro, and I will hold a high school teacher in Phoenix alongside it as the contrast. These two are chosen deliberately because they produce opposite conclusions from the same method - one is strongly seasonal, one is not - which is exactly the point the listicles miss.

First, some stock context a posting index alone cannot supply. In Refolk's index of professional profiles, the labor pool behind each role differs a lot, and pool size reframes how much competition a peak actually brings.

The US carries about 34 percent more teachers than registered nurses. If you are a nurse eyeing a move abroad, geography changes the picture again.

The US registered-nurse pool is roughly 24 times the UK's. None of this tells you when to apply - the posting curve does that - but it tells you how crowded the peak will be when you get there.

The two public sources that have monthly, metro-level data

Two primary sources give you monthly data at the granularity you need, and both are free. The Indeed Hiring Lab Job Postings Index is daily, refreshed weekly, and each series - the national trend, occupational sectors, and sub-national geographies - is seasonally adjusted separately. For the US, indices are calculated for states and the largest metro areas by population, broken out at the Metropolitan Statistical Area level, and mirrored on FRED. BLS JOLTS is the other: every month BLS publishes hires, openings, and separations at national and state level from a survey of 21,000 establishments.

Each source has a shape you must respect.

Read the Indeed catch carefully, because it causes a common error. The index is the percentage change in seasonally adjusted postings since February 1, 2020, where a reading of 101 is 1 percent above the baseline. It is not a count of jobs. You use it for change within one series over time, never to compare the size of two metros as if the numbers were volumes.

The other Indeed limit is suppression. For a niche occupation in a single metro, they apply sample-size thresholds, and if the audience is too small they aggregate further or do not publish. That will bite a specialized role in a small city, and I will show what to do about it.

The procedure: from a role and a city to a curve

Here is the whole method. It runs about four hours the first time and under an hour on repeat, because you keep the spreadsheet. Notice the order: I derive the industry and metro curve first and treat the national January story as one pattern among many, rather than starting from a universal peak and bolting on exceptions.

Step 1 to 3, worked

For the Dallas-Fort Worth nurse, I map the role to a healthcare occupational sector, the metro to the Dallas-Fort Worth-Arlington MSA, and pull the Indeed healthcare series for that MSA from FRED plus JOLTS healthcare hires for Texas. I take everything available, which gives me well past the 36-point minimum. For the Phoenix teacher, I map to an education sector and the Phoenix MSA and do the same.

Step 4, the simple method

You do not need statistical software for a first pass. The deviation method is enough: for each year, compute the year's average posting level, then for each calendar month compute how far that month sat above or below that year's average, then average those deviations across years. That gives a 12-point curve. If you want the rigorous version, the US Census Bureau's X-13ARIMA-SEATS software is free, and the seasonal R package wraps it.

The three wrong turns, worked in the example

Every real attempt hits at least one of these, so I walk each one as it happened in the Dallas nurse and Phoenix teacher cases.

Wrong turn one: mistaking a spike for a season

For the Phoenix teacher, one year showed a huge January jump. Read alone, that looks like a January peak, and a listicle would have confirmed the bias. The test is recurrence. The operational floor is three or more years of data, because X-13ARIMA needs a minimum of three years of monthly data to capture seasonality and errors out on shorter series. Indeed uses the same floor - its projected seasonal factors are based on estimates from the preceding three years, built on historical patterns in 2017, 2018, and 2019. Checked across three years, the teacher's January jump did not repeat; it was a one-off, likely a single large district opening. The recurring signal was the August rise, consistent with education staying below average until May then picking up in August.

Wrong turn two: borrowing the national curve for a local market

The flagship metro is not always the hottest. As of one October 31 reading, the Job Postings Index in Atlanta was 110, while nearby Valdosta, Warner Robins, Macon, and Hinesville all exceeded 130. Demand was more resilient in the mid-size markets next door. For the Dallas nurse this matters twice: the national healthcare curve is flat, but the local one might not be, and a smaller metro within commuting range might run a higher index. The fix is to pull your MSA series, not the national one, and to check the neighbors.

Wrong turn three: ignoring the posting-to-hire lag

This is the single biggest error the listicles omit. Applying at the posting peak means your interviews land after the peak, when the pipeline is already full. Multiple benchmarks converge on a lag of roughly three to nine weeks. SHRM puts the average US time to fill at 44 days. Average time to hire is about 24 days but varies by industry, with retail and hospitality fastest at 14 to 20 days and healthcare and engineering slowest at around 49 days. So to be interviewed during a posting peak, your application must land three to seven weeks earlier.

Definitions matter here, and confusing them shortens your lead time. Time to hire is days from a candidate applying to accepting an offer. Time to fill is the total time a position stays unfilled, requisition to accept. Use time to fill for calendar planning; the candidate-side number understates how early you must move.

For the Phoenix teacher whose real peak is August, a 5 to 6 week offset means the ramp begins in late June and early July. Apply in August itself and you interview into September, past the hiring window for the academic year.

When the answer is "there is no season"

For some roles the correct output of this method is not a month but a rule: apply on posting. Healthcare demand is consistent and growing, and its monthly deviation band is narrow. BLS projects roughly 1.9 million healthcare openings per year from 2024 to 2034 from growth and replacement needs. That is the Dallas nurse's result. The deviation curve comes back nearly flat, the recurrence test finds no month that reliably beats the others, and waiting for a peak simply wastes months of eligibility.

This is not a failure of the method. Non-seasonal is a strategy. From the same public data, the Phoenix teacher gets a dated ramp window and the Dallas nurse gets a standing instruction to apply the moment a fit appears. The matrix below is how to read your own curve once you have it.

When you are trying to see how peers in a non-seasonal role actually move - who changed employers, and when - a posting index will not show you individuals. That is where searching the profile graph directly removes the friction.

Building the ramp-and-coast calendar

The output is a dated calendar with two kinds of weeks: ramp weeks, when you push volume, and coast weeks, when you skill-build and prepare. You place the ramp by taking your confirmed posting peak and subtracting your industry's offset.

For the Phoenix teacher, the confirmed peak is August, the offset is five to six weeks, so the ramp runs roughly the last week of June through early August, six to eight weeks of concentrated applications. The coast period is the winter trough, spent on portfolio, references, and credential renewal rather than sending applications into a thin market.

Here is a reusable skeleton. Fill the bracketed values from your own curve.

Occupation + metro: [role] in [MSA]
Confirmed posting peak month(s): [month], recurs across [3+] years
Industry time-to-fill: [days] -> offset [weeks]
RAMP window (apply hard): [peak minus offset] through [peak]
  - Target volume: [N] applications/week
  - Tailor every application; the peak is crowded
COAST window (skill-build): [trough months]
  - No cold volume; refresh resume, references, portfolio
Recheck date: [month] each year, re-pull latest 12 months

During the ramp, volume alone is not enough because the peak is the most crowded moment of the year. Tailoring each application to the posting is what separates you from the flood, and that is slow by hand. Refolk writes your resume from your own history, tailors it to each posting, drafts the cover letter, and scores how well you actually fit, which is what makes a six-week ramp of well-fitted applications feasible instead of a choice between volume and quality. You can start from Refolk when your ramp window opens.

Keeping the curve current

A curve built once decays, so treat it as a standing series you refresh, not a one-time verdict. The Indeed index updates weekly and JOLTS updates monthly, so re-pull the latest 12 months once a year, ideally a month or two before your ramp window opens, and re-run the deviation calculation. Official statistics agencies go further than the three-year floor - standard seasonal adjustment uses a symmetric filter drawing on 6 to 10 years of original data - so the longer your series, the more stable your factors become. Each year you add makes the peak more trustworthy.

Two things to watch on refresh. First, if your segment starts getting suppressed because the metro sample thinned, aggregate up to the state or a broader occupation family and note that the curve is now coarser. Second, if a new large employer entered your metro, a fresh spike will appear that has not yet passed the three-year recurrence test; hold it as a candidate, not a confirmed peak, until it repeats.

The discipline that separates this from a listicle is the three-year rule and the lag offset. Anyone can name a month. Deriving your own curve, proving the peak recurs, and shifting it earlier by your industry's time to fill is what turns "best time of year to apply for jobs" from a headline into a calendar you can actually work from.