TemplateLast reviewed November 4, 2026

Job Safety Analysis Template with Risk Scoring (Free)

Free job safety analysis template: a macro-free .xlsx that scores each task step's hazards before and after controls on a 5x5 risk matrix.

A crew about to start a risky job usually talks it through first. Our job safety analysis template puts that conversation on paper and scores it. List the job’s steps, name the hazard in each one, and score it by likelihood and severity before and after controls. The workbook works out every risk score, rates it against bands you set, plots each step on a 5x5 risk matrix, and keeps the crew sign-off and a toolbox talk attendance log in the same file.

Download the job safety analysis template. It’s a plain .xlsx with no macros and no sign-up. The sample job is a rooftop compressor replacement, so you can see every score and count working before you clear it and write up your own.

Video walkthrough

This short captioned walkthrough covers the JSA sheet, the risk matrix and the toolbox talk log.

Job Safety Analysis Template walkthrough

What a job safety analysis is

A job safety analysis, often shortened to JSA and also called a job hazard analysis (JHA), takes one job and breaks it into the steps a technician actually performs. For each step you ask what could hurt someone, then decide what will prevent it or make it less serious. OSHA’s publication 3071, Job Hazard Analysis, explains the method in detail and is the best place to start if your company hasn’t written one before.

The JSA template follows that step, hazard and control layout and adds a score. The 1-to-5 likelihood and severity scales, the bands and the matrix are this template’s own way of ranking hazards. They’re a planning aid for your crew and aren’t drawn from any regulation.

Who this template is for

It’s built for field service and mechanical contractors who send small crews into non-routine work: a compressor swap on a roof, a boiler tear-down, a lift with a crane, a panel change in an occupied building. It also suits a service manager or safety lead who reviews the crew’s JSA before the job starts and wants the same format every time.

The template answers a few questions about one job:

  1. Which steps carry the worst hazards before anyone does anything about them?
  2. Do the controls you wrote down bring each score down, and by how much?
  3. Which hazards are still rated High or Extreme with the controls in place?
  4. Who on the crew has reviewed and signed the JSA?

What’s in the workbook

There are six sheets, in this order:

  1. Instructions covers setup, the color legend, the limitations, and a note that the formulas, layout and sample figures were written from scratch for this file.
  2. JSA holds the job header (job, location, date and who prepared it) and up to 50 steps. Each step has a task, a hazard, a likelihood and severity before controls, the controls, and a likelihood and severity after controls. The risk scores, ratings and Row check are calculated.
  3. Risk Matrix is where you set the rating bands. It shows two 5x5 grids that count how many steps sit at each likelihood and severity, one before controls and one after.
  4. Sign-off lists everyone who reviewed the JSA, with their company or role, the date and whether they signed.
  5. Toolbox Talks logs each safety briefing: the date, topic, presenter, number of attendees and whether a sign-in sheet was kept.
  6. Dashboard pulls it together: steps analyzed, high or extreme hazards before and after controls, average scores, the risk-score reduction, sign-off, talk attendance and the data checks.

Blue cells are for typing and yellow cells calculate. The Dashboard results are green.

How to fill in a JSA

  1. On JSA, type the job, the location, the date and your name.
  2. Clear the sample steps and list your job’s steps in the order the crew will do them. Keep each step short: “remove access panels,” not a paragraph. A step with two separate hazards takes two rows.
  3. For each hazard, score likelihood and severity before controls, each a whole number from 1 (lowest) to 5 (highest).
  4. Write the controls. Then score likelihood and severity again as if those controls are in place.
  5. On Risk Matrix, check the three band limits. The sample uses 4, 9 and 14.
  6. Fill in Sign-off before the work starts, and log each toolbox talk on Toolbox Talks as you hold it.
  7. Fix any Row check that doesn’t read OK, then read the Dashboard.

JSA sheet showing the rooftop compressor replacement job header and eight task steps, each with a hazard, likelihood and severity before controls, a risk score and rating, the controls, and the scores and rating after controls, with a Row check of OK on every step

Row check catches the mistakes that would throw off the totals. Check inputs means the step number, task or hazard is blank, or one of the four scores is missing or isn’t a whole number from 1 to 5. Add controls means a step has a hazard and scores but no controls written. Check scores means the score after controls is higher than the score before, which usually means the two were typed the wrong way round. Every one of those adds to Rows to check on the Dashboard, and the figures there read “Check data” until it’s back to zero.

How the risk score works

Each hazard gets two numbers. Likelihood is how likely the hazard is to hurt someone on this job. Severity is how bad the harm would be. Both run from 1 to 5, and the risk score is likelihood times severity, so it runs from 1 to 25.

The score turns into a rating through the three limits on the Risk Matrix sheet. With the sample bands, a score up to 4 rates Low, up to 9 rates Medium, up to 14 rates High, and anything above that rates Extreme. Change a limit and every rating on the JSA, every matrix color and the Dashboard counts follow it. The limits have to be whole numbers from 1 to 24, each higher than the one before. If they aren’t, Bands check on Risk Matrix says so and the Dashboard holds its figures at “Check data.”

Take the first sample step, setting up a ladder at the rooftop hatch. A fall from a ladder onto a roof deck could be serious, so severity stays high whether or not controls are in place. Before controls, the score is 15, which rates Extreme. Tying the ladder off and having a second tech foot it lowers the likelihood but not the severity, so the score after controls is 10, still High. It stays in the warm colors on the matrix, so it’s the step a supervisor should look at first.

Lockout and tagout on the unit’s disconnect is different. It starts at 15, but locking, tagging and testing for zero voltage cut the likelihood to the bottom of the scale, so the score after controls is 5. Severity stays at the top of the scale because a shock would be just as serious; the controls only make one less likely.

Reading the risk matrix

The Risk Matrix sheet has two grids. Likelihood runs down the side from 1 at the top to 5 at the bottom, and severity runs across from 1 on the left to 5 on the right. Each cell counts the steps scored at that likelihood and severity, and its color is the band that cell’s score falls in.

Risk Matrix sheet with the Low, Medium and High band limits and a Bands check of OK, above two 5x5 grids counting steps by likelihood and severity before controls and after controls, colored from green to red by rating band

Compare the two grids and you can see what your controls did. Before controls, 2 steps sit at likelihood 3, severity 5, and 2 more at likelihood 2, severity 5. After controls, those steps have moved up the grid: 2 sit at likelihood 1, severity 5. Steps whose severity also fell move left as well, and 2 steps end up at likelihood 2, severity 2. A step that stays in the same cell after controls is a step where the controls on paper didn’t change the score.

Worked example: a rooftop compressor replacement

The sample JSA covers 8 steps for replacing the compressor on a rooftop unit: ladder setup, lockout, panel removal, refrigerant recovery, brazing, the crane lift, start-up and clean-up. Every figure is a sample input to replace. The Dashboard works out:

MeasureResult
Steps analyzed8
High or extreme before controls5
High or extreme after controls1
Average risk before10.88
Average risk after5.25
Risk-score reduction51.7%
Highest residual score10
Sign-off complete75.0%
Toolbox talks logged3
Talk attendees18
Rows to check0

Dashboard sheet listing steps analyzed, high or extreme hazards before and after controls, average risk before and after, the risk-score reduction, highest residual score, sign-off complete, toolbox talks logged, talk attendees, rows to check and a Data check of OK

5 of the 8 steps start out High or Extreme. With the controls in place, the count drops to 1: the ladder setup, which keeps the highest residual score at 10. Brazing the replacement valve shows a middle case. It starts at 12 (High) and the fire watch and cleared combustibles bring it to 6 (Medium).

The risk-score reduction of 51.7% compares the total of the scores entered before controls with the total after. It tells you how much the crew’s own scoring moved, which is useful for comparing one JSA with the next. It isn’t a measured drop in injuries or real-world risk, and it’s only as good as the scores typed in.

Sign-off complete reads 75.0% because one of the four people listed, the apprentice, hasn’t signed yet.

Sign-off and toolbox talks

The Sign-off sheet needs a name, a date and Yes or No for signed on every row. Company or role is optional, which is handy for a subcontractor like the crane operator in the sample. Print the JSA and Sign-off sheets together if your crew signs on paper.

Toolbox Talks is a running log of the short safety briefings you hold. Each row needs a date, a topic, a presenter, the number of attendees (a whole number from 0 to 500) and Yes or No for whether a sign-in sheet was kept.

Toolbox Talks sheet logging three safety briefings with date, topic, presenter, attendee count, whether a sign-in sheet was kept and a Row check of OK on each

The sample logs 3 talks with 18 attendees in total. That total counts attendance, not people, so a technician who sat in on all three talks counts three times.

How it fits with our other templates

A JSA covers one job. A few other documents sit around it:

You’ll find the rest of the set on our free field service templates page.

Limitations to know before relying on it

  • The scores are your crew’s judgment. The template checks that they’re whole numbers from 1 to 5 and that the score after controls isn’t higher than before. It can’t tell you whether a control is adequate.
  • The bands, scales and matrix are this template’s own method. The workbook is a planning record. It isn’t a form any regulator, customer or general contractor issues, and it doesn’t tell you which rules apply to your work.
  • One JSA sheet holds one job with up to 50 steps. Copy the sheet for another job, but the Risk Matrix and Dashboard read only the original JSA sheet.
  • Sign-off and Toolbox Talks hold 200 rows each. There are no electronic signatures, and nothing is sent to the office or synced between crews.

Grab the job safety analysis template, clear the sample job, and write up your next one before the crew heads out.

Frequently asked questions

  1. What is a job safety analysis?

    A job safety analysis, also called a job hazard analysis, looks at one task step by step, names the hazards in each step, and sets the controls that remove or reduce them. OSHA publication 3071, Job Hazard Analysis, describes the method.

  2. How does the risk score work?

    Each hazard gets a likelihood and a severity from 1 to 5, before controls and again after. The score is likelihood times severity, from 1 to 25, and the Low, Medium, High and Extreme bands come from thresholds you set on the Risk Matrix sheet. The risk-score reduction compares the scores you entered. It is not a measured change in injury rates.

  3. What is the difference between a JSA and a toolbox talk?

    A JSA is written for one job and its steps, usually before the work starts, and the crew signs it. A toolbox talk is a short safety briefing on one topic for the whole crew. The template keeps both: the JSA and sign-off for the job, and an attendance log for the talks.

  4. Can I change the risk bands?

    Yes. Type your own upper limits for Low, Medium and High on the Risk Matrix sheet, as whole numbers from 1 to 24, each higher than the one before. Ratings on the JSA, the matrix colors and the dashboard counts all follow the new bands.

  5. Is this an OSHA form?

    No. It is a planning record for your own crews. It does not replace any form a regulator, customer or general contractor asks for, so check which rules apply to your work.

Trust signal

Fact Checked & Editorial Guidelines

Every post on this site is fact-checked against the policy below before the "Last reviewed" date is updated. If a single item below fails verification, the post does not go live.

  • Every claim traces to a source.

    Pricing, feature lists, integrations, and headquarters are taken from vendor product pages, documentation, or signed contracts — never repeated from secondary blogs. Where a claim is sourced from a single vendor's marketing, it is qualified as such.

  • Vendor relationships are disclosed in-line.

    If a review covers a platform whose vendor has provided trial access, sandbox access, or paid placement on a sister property, that relationship is stated in the review's methodology footer — not buried in a sitewide disclosure page.

  • Pricing is rechecked at every review cycle.

    Vendor pricing changes constantly. The 'Last reviewed' date on each post is the date the price line was last re-verified against the vendor's public pricing page. If you spot a stale price, the contact page accepts corrections.

  • Corrections are logged, not silently rewritten.

    Material factual corrections after publication get a correction note dated and appended to the post. We don't pretend the prior version never said what it said.

Spotted an error? Send a correction via thecontact page — corrections are logged with a dated note on the post.

Trust signal

Editorial Review & Methodology

Reviews and comparisons on this site follow a single documented methodology — the same rubric, applied identically to every platform, on every review cycle.

  • Five-criteria scoring rubric, applied identically to every platform.

    Usability, pricing transparency, feature depth, support quality, and integrations. Each criterion scored 0–10 with documented weighting. The rubric is published on the methodology page and does not change between platforms in the same review.

  • Hands-on testing where vendor trial access permits.

    If a vendor offers trial or sandbox access, the reviewer spins up an account and works through the documented evaluation script before scoring. Where access is enterprise-gated, the access type is disclosed and scoring draws on product documentation, verified buyer reviews, and analyst sources.

  • Editorial independence from commercial relationships.

    No vendor pays for placement, previews scores, or controls the content of a review. Affiliate links, where present, do not change ranking — picks are ordered by score, not by commercial yield. If a conflict of interest exists for a specific review, it is disclosed within that review.

  • Reviews get re-checked, not just re-dated.

    Each 'Last reviewed' update means the rubric was re-applied — pricing, feature inventory, integration list, and any material vendor changes since the prior review. A bare date bump without re-evaluation is not a re-review.

The full rubric, weighting, and review-cycle process is on themethodology page.