JobStraight · Research

What 2,411 job postings actually ask for

Job advice is mostly opinion. This is not. We fetched complete job descriptions from 77 company boards, read the requirements section of each, and counted which skills employers actually name — then published the method, the sample sizes, and everything we discarded.

Last updated 2026-08-17 · Source: Greenhouse public board API · Free to cite with a link.

What we did

Job aggregators return a truncated preview of each posting — usually the opening paragraph, which ends before the requirements. Counting skills in that text gives numbers that are simply wrong; our own first attempt produced SQL at 5% for analyst roles. So we used the Greenhouse board API, which returns the complete posting body, and read only the requirements section. Postings naming more than twelve distinct technologies were discarded as aggregator listings. Roles with fewer than 100 qualifying postings were dropped entirely rather than published with a caveat.

Marketing

Based on 177 postings. Percentages are the share whose requirements section named the skill.

Skills named in Marketing job postings, with the share of postings naming each
SkillShare of postings%
Salesforce
17%
A/B testing
13%
Degree required
12%
Stakeholders
11%
Looker
7%
SEO
7%
Excel
5%
AWS
5%
Tableau
5%

Product manager

Based on 423 postings. Percentages are the share whose requirements section named the skill.

Skills named in Product manager job postings, with the share of postings naming each
SkillShare of postings%
Degree required
22%
SQL
16%
A/B testing
13%
Stakeholders
13%
LLM
8%
AWS
8%
Machine learning
6%
Agile
6%
Azure
5%
GCP
5%
Kubernetes
5%

Software engineer

Based on 1,460 postings. Percentages are the share whose requirements section named the skill.

Skills named in Software engineer job postings, with the share of postings naming each
SkillShare of postings%
Python
33%
AWS
29%
Java
24%
Kubernetes
23%
Degree required
20%
Mentoring
19%
React
17%
TypeScript
16%
CI/CD
13%
C++
13%
Azure
12%
GCP
12%
Git
11%
SQL
10%

DevOps / SRE

Based on 128 postings. Percentages are the share whose requirements section named the skill.

Skills named in DevOps / SRE job postings, with the share of postings naming each
SkillShare of postings%
AWS
70%
Kubernetes
67%
Python
58%
Terraform
51%
CI/CD
41%
GCP
34%
Linux
30%
Degree required
28%
Mentoring
25%
Git
23%
Azure
20%
Docker
17%
Go
16%
SQL
13%

Designer

Based on 115 postings. Percentages are the share whose requirements section named the skill.

Skills named in Designer job postings, with the share of postings naming each
SkillShare of postings%
Figma
34%
Mentoring
14%
A/B testing
10%
Stakeholders
8%
Degree required
6%

Data scientist

Based on 108 postings. Percentages are the share whose requirements section named the skill.

Skills named in Data scientist job postings, with the share of postings naming each
SkillShare of postings%
SQL
86%
Python
84%
A/B testing
49%
Machine learning
37%
Degree required
28%
Spark
23%
Mentoring
16%
Stakeholders
9%
Snowflake
9%
Looker
9%
Airflow
9%
LLM
8%
Tableau
7%
dbt
6%

Method, and what this is not

Where does this data come from?

Complete job descriptions from the Greenhouse public board API, which returns the full posting body rather than the truncated preview aggregators serve. 2,411 postings across 6 roles met the inclusion criteria, drawn from 77 company boards.

Why are some roles missing?

A role is published only when at least 100 postings qualify. Smaller samples gave figures we could not stand behind: the data-analyst bucket read 68% for SQL at n=37, almost certainly a small-sample artefact, so it is excluded rather than published with a caveat.

How are skills counted?

Only the requirements section is read; the company blurb and benefits mention technologies incidentally and would inflate every count. Matching is word-boundary only, and any posting naming more than twelve distinct technologies is discarded as an aggregator listing.

Does this reflect the whole job market?

No, and it should not be read that way. These are venture-backed and mid-size technology companies using Greenhouse. It does not represent IT services, government or small businesses.

Citing this

Free to quote, republish and build on, including commercially. A link back is all we ask — no permission request and no attribution wording you have to match.

JobStraight, "What 2,411 job postings actually ask for", 2026-08-17. https://jobstraight.com/skills-in-demand/

Pay data for the same roles, measured the same way, is in the advertised-salary report.

Use this

If a skill appears in most postings for your target role, it belongs on your resume with an example behind it. If it appears in under a tenth, it is not what is keeping you from interviews.

Check your resume against a job description, rehearse the questions a posting implies, or learn the skills these postings name.

Keep reading

The AI skills employers actually ask for in 2026
Which AI-related skills appear in real job postings, which are hype, and how to describe your own AI use on a…
How to read a job description (and when to apply anyway)
Which requirements are real and which are wishlist, what three to five years actually means, and how to…
Every role we track
Browse 45 job roles with 5,567 live openings, each linking to real postings, measured pay and how to get into…
Read any job post like an ATS
Free job description keyword extractor: paste a posting to get the tools, skills and seniority an ATS screens…