Manual Tester Job Demand in 2026: What the Numbers Really Say

Yes. The most reliable numbers come from the Bureau of Labor Statistics (BLS), which tracks software QA analysts and testers as their own group. Here is what it reports:

• 187,600 people had these jobs in 2025.

• 198,200 jobs are expected by 2035. That is 6% growth, or about 10,600 new jobs.

• Growth for all jobs combined is expected to be 3%.

• The median pay was $104,300 in May 2025. The lowest-paid 10% earned under $61,440, and the top 10% earned over $167,010.

Six percent over ten years works out to roughly 1,000 new jobs a year. That is steady, not fast. BLS expects software developers to grow 10%, so testing is growing slower than the people who write the code.

Why grow at all? BLS points to more software built for AI, connected devices, and robots, plus more money going into security software that needs testing. Every new product with software inside needs someone to check it. That is the main engine behind manual tester job growth, and it doesn’t care whether the checking is done by hand or by script.

A catch you should know about. BLS doesn’t split manual testers from automation testers. It describes the job as using manual or automated tests, plus exploratory testing. So manual testers are in the count, but you can’t read 187,600 as “manual-only jobs.” The mix inside that number is changing, and that is the next section.

The total is steady, but three things are moving underneath it.

1. The checking that repeats is going to scripts.
When a team ships every week, nobody can re-check every old feature by hand before each release. That re-checking is called regression testing, and it’s the first thing teams hand to scripts. Katalon’s survey of 1,400+ QA professionals found that about 45% automate regression testing, and it is the most automated kind. The same survey found that 82% still use manual testing in their daily work. So manual work isn’t gone. The repeating part of it is.

2. The job title is changing.
One 2026 hiring review says ads titled “manual QA tester” are down in many markets, while ads for quality engineers, SDETs, and test automation roles hold up. (An SDET is a tester who also writes code.) The desk is the same, but the name and the asks are different. If you only search “manual tester,” you’ll miss jobs. Search QA analyst, QA engineer, and test engineer too.

3. The ads ask for more.
A job tracker called Skillenai read 66 “Manual QA Engineer” ads from a 90-day stretch. The skills named most were Jira, SQL, manual testing, API testing, and Confluence. That’s a small sample. But look at the pattern. SQL and API testing sit right next to manual testing. Employers want a manual tester who can also look inside a database and check how apps talk to each other.

The job itself is also wider than running cases. BLS lists the work as writing test plans, spotting project risks, running manual, automated, and exploratory tests, reporting defects, and giving feedback on how usable the software is. Running the same list of cases by hand is one small piece of that.

BLS shows where testers work most:

• Computer systems design firms: 30% of tester jobs, with a median pay of $99,960.

• Finance and insurance: 13% of jobs, with a median pay of $104,260.

• Manufacturing: only 9% of jobs, but the highest median pay at $129,930.

(Source: BLS, May 2025 wages.)

Banks and insurers test rules, not just screens. Say a claim is filed on the last day of a policy. Does the system pay it? A script can only check what someone told it to check. A tester who knows insurance asks the question nobody wrote down. testRigor makes the same point: complex business rules in fields like healthcare and finance still need people to check them.

BLS doesn’t say why manufacturing pays more. It does name connected devices and electric vehicles as growing sources of software work, and testing something that touches real hardware often needs a person holding the device.

Some kinds of testing lean on people no matter the industry:

• Exploratory testing - You use the product with no script and look for what nobody planned for. A user pays, the session times out, they log back in, and the cart is empty. You find that by wandering, not by following steps.

• Usability checks - A screen can work perfectly and still confuse a first-time user. A script has no way to feel lost.

• Accessibility checks - Can someone using a screen reader finish checkout? You have to try it.

• Features that keep changing - Early in a project, the screens change every few days. Writing scripts for them is a waste, so people test by hand until things settle.

AI is already in QA teams. It just isn’t running them. The World Quality Report found that 89% of companies are piloting or using generative AI in their quality work, but only 37% have it in real production. Companies reported an average productivity gain of 19%, while about a third saw very little.

AI also creates new checking work. CodeRabbit studied 470 open-source code changes. The ones with AI help had about 1.7 times more issues, 10.83 per change against 6.45, and logic mistakes were 75% more common. The study guessed which changes used AI from signs in the code, so treat it as a strong clue, not proof. Logic mistakes are the kind a tester with product knowledge catches. More code written faster means more to check.

Then there’s the AI inside products. When a chatbot answers a customer or an app summarizes a document, no script knows if the answer is right. testRigor says people are still needed to check AI output for errors and bias. That is new work, and it didn’t exist a few years ago.

There’s a harder side, too. AI is good at the routine pieces: drafting test cases, making test data, scanning long logs. That means fewer people are needed for routine work. The testers who use AI for those parts, then spend the saved time on judgment, come out ahead. That’s why one step in the path below is about using AI tools, not avoiding them.

Manual tester demand hasn’t left. It sits inside ads that read a little differently. Here’s a typical one. We wrote it to show the pattern, so it’s not a real posting. The skills in it match what trackers like Skillenai keep finding.

Tool names matter, so here’s what the counts say. TestDino found over 8,800 Indeed results for “automation testing Selenium” in February 2026. It also says Playwright ads roughly tripled between 2024 and 2026. Those counts overlap, because one ad can list both tools. So don’t treat it like a race. Check the ads near you, pick one tool, and the other gets easier to learn later.

One more thing. A line marked “a plus” is nice to have, so don’t skip an ad because you’re missing one. But every line you can already do puts you ahead of the next person.

A hybrid tester tests by hand and also uses tools. Most ads now describe this person. You don’t have to learn it all at once. Here’s an order that builds on itself.

Step 1: Make your test basics solid.
Everything else sits on top of this. Learn to write test cases, test plans, and bug reports that a developer can follow without calling you. Learn how a feature moves from idea to release. Proof to build: a sample test plan for an app you use, and three bug reports. JanBask’s Manual Testing course covers this in 4 weeks, including the software life cycle, test case design, test plans, and levels of testing. You only need basic computer skills.

Step 2: Add SQL and API testing.
These two show up in almost every ad. Both are easier than they sound, and neither needs heavy coding. Proof to build: a Postman collection for a free public API, and ten SQL queries you wrote and can explain.

Step 3: Learn one automation tool.
Pick Selenium or Playwright, based on the ads near you. You’re not trying to become a developer. You’re learning enough to automate the boring checks you’d otherwise repeat. JanBask’s QA Software Testing program covers steps 2 and 3 in 8 weeks, with manual testing, Java, Selenium, TestNG, Playwright, API testing, GitHub, and Jenkins, plus a capstone project. Proof to build: a GitHub page with a few automated tests and a short note on what each one checks.

Step 4: Use AI as a helper.
Let it draft test cases and read logs, then check what it gives you. Mistakes in AI output are where your testing instincts matter most. The AI in Automation Testing course runs 6 weeks and covers JavaScript, Playwright, API testing, ChatGPT, Playwright MCP, and GitHub Copilot. It uses JavaScript, while the program above uses Java, so pick the one that matches your target ads. Proof to build: a set of AI-drafted tests with your fixes marked.

You might also want a certificate. The ISTQB Foundation Level is the usual first one, and it’s the starting point for ISTQB’s other certificates, including AI Testing. It can help a résumé get read, but it won’t replace the proof above.

A course gives you order and someone to ask. The proof is what you’ll talk about in the interview.

The best next step depends on where you are today.

You haven’t started yet.
BLS says these jobs typically ask for a bachelor’s degree, but it lists no related work experience or on-the-job training as a requirement. Ads vary a lot, so read ten for jobs you’d want before you decide how much to learn. Start with steps 1 and 2. The median pay above covers every experience level, so a first job often pays less. Your first role is the hardest to land. A small set of bug reports and a Postman collection helps more than you’d expect.

You’ve tested manually for one to three years.
This is the best time to move. You already know how software breaks, and that’s the hard part. Do steps 2 and 3 now. SQL, API testing, and one tool are what change how ads treat you.

You’ve tested manually for five years or more.
Your product knowledge is of real value, and the tools will make it easier to use. But look at the PractiTest 2026 State of Testing report, a worldwide survey. It found that senior testers (10+ years) who build leadership and planning skills earn 10.6% more, while those who rely only on technical execution, like writing scripts, earn 13.8% less. So don’t just do the same work faster, by hand or by script. Add the tools, then move toward planning, risk, and leading a team.

If you’re not sure which group you’re in, JanBask offers free demo classes, and an advisor can look at your background and suggest a first step.

Manual tester job demand is real, and it’s moving toward testers who think, check, and use tools. That mix is something you can build, one step at a time.