Penetration testing cost: what a pen test costs and what sets the price
Penetration testing cost is set by the scope: how many applications, interfaces, user roles, servers and networks are tested, whether the testers start with login details or from the outside, how deep they go, and whether the fixes are retested. Most pen tests are quoted in tester days, so the same company can receive quotes that differ several times over simply because each supplier is testing something different. This guide explains the types of test, the cost drivers, the rules that require testing, the costs people forget and how to compare quotes on the same basis.
It does not quote our own prices. It explains how a pen test budget is built, so any quote you receive, from us or from anyone else, makes sense. For where testing fits in a wider review, see our security audit service, which explains the difference between an audit, a vulnerability scan and a penetration test.
The short version
- Scoping: agreeing the targets, the rules, the test window and who to contact if something breaks.
- Testing: tester days spent on the targets, manual work supported by tools.
- Reporting: findings, evidence, severity and advice, with a summary for management.
- Fixing and retesting: your team's work to fix findings, and a retest to confirm.
Types of penetration test
| Test type | What it covers | What sets the effort |
|---|---|---|
| Web application | A website, portal or web app, logged in and logged out | Number of roles, functions, forms and integrations |
| API | The interfaces behind apps and integrations | Number of endpoints, authentication methods and data flows |
| Mobile app | The app on the device and the services behind it | Platforms, local data storage, the back end |
| External network | What is reachable from the internet | Number of addresses, services and exposed systems |
| Internal network | What an attacker inside the network could reach | Network size, directory services, segmentation |
| Cloud configuration | Accounts, identities, storage and network settings in the cloud | Number of accounts, subscriptions and services |
| Social engineering | Phishing and pretexting against staff, with consent from management | Number of people and scenarios |
Tester days are most of the bill
A pen test is skilled people's time. What an hour costs follows where the team works and how specialised it is. Eurostat's release of 31 March 2026 puts average hourly labour costs across the whole economy in 2025 at €34.9 in the EU and €38.2 in the euro area, from €12.0 in Bulgaria to €56.8 in Luxembourg (read on 12 November 2026). Experienced security testers are paid well above those averages, but the figures explain why quotes from different countries differ. Compare the number of tester days and what they cover, not the day rate alone.
What drives penetration testing cost
| Driver | Why it adds work | How to keep it in check |
|---|---|---|
| Size of the target | More functions, endpoints, addresses or accounts take more days to cover | Test the systems that hold the most valuable data first |
| User roles | Each role must be tested for what it can reach that it should not | Give testers one account per role |
| Starting point | Testing from outside without information takes longer than with documentation and credentials | Share documentation; it buys depth instead of discovery time |
| Depth | Chaining findings into a realistic attack takes more time than listing them | Agree which risks matter most to the business |
| Environment | Testing in production needs care and test windows; a test copy needs preparing | A production-like test environment where possible |
| Retesting | Confirming fixes is a second, smaller round of work | Ask for it in the first quote |
| Reporting needs | Reports for a regulator, auditor or large customer need more detail and evidence | Tell the supplier who will read the report |
Rules that require or expect testing
For companies in scope of NIS2, cybersecurity risk-management measures must include security in development and maintenance, including vulnerability handling and disclosure, and policies and procedures to assess the effectiveness of those measures (Directive (EU) 2022/2555, Article 21(2)(e) and (f)). Penetration testing is a common way to produce that evidence. For identified financial entities, DORA goes further: they must carry out advanced testing by means of threat-led penetration testing at least every three years, covering several or all critical or important functions on live production systems (Regulation (EU) 2022/2554, Article 26). The OWASP Web Security Testing Guide, version 4.2 at the time of reading, is a widely used reference for what a thorough web application test covers (all read on 12 November 2026). A test meant as evidence for a regulator usually needs more scoping and reporting than one done for your own assurance.
The hidden costs of a pen test
- Your own team's time. Preparing accounts, environments and documentation, and being available during testing.
- Fixing findings. Often the largest cost, and the reason for testing in the first place.
- Retesting. Without it, nobody knows whether the fixes worked.
- Test environment. A safe copy of production with realistic data takes work to prepare.
- Third-party permission. Cloud and hosting providers may have rules on testing; some systems need their owner's consent.
- A report nobody acts on. Findings without owners and dates cost the test and fix nothing.
Pricing models and who carries the risk
| Model | How it works | Who carries the risk |
|---|---|---|
| Fixed price per scope | An agreed price for named targets and a number of tester days | The supplier carries the estimate; the buyer carries the risk of a scope that leaves out important systems |
| Day rate | Tester days billed as used | The buyer carries overrun risk; flexible for unclear scopes |
| Subscription or retainer | Recurring tests, scans and retests over a year | Shared: suits systems that change often |
| Automated scan service | A tool run on a schedule with a generated report | The buyer carries the risk of what a tool cannot find |
How to compare pen test quotes
- Which targets are in scope, by name, and which are not?
- How many tester days, and is the work manual or mostly automated?
- Do testers start with credentials and documentation, or from outside?
- Which method or standard is followed?
- Is a retest of fixes included, and within what time?
- What does the report contain, and is there a management summary?
- Who are the testers, and what experience do they have with this kind of system?
- How are test data and findings stored and deleted afterwards?
Ways to spend less without cutting quality
- Close the obvious gaps first. A security audit or scan before the pen test lets testers spend days on real depth.
- Share documentation. Testers then find more in the same days.
- Test before launch. Fixing before go-live costs less than fixing in production.
- Scope by risk. Test the systems with the most valuable data most deeply.
- Plan the fixes. Budget the fix and retest time when you order the test.
Related reading: NIS2 compliance, incident response plan, client portal cost and web app development cost. All our services are on the services page, and every guide is in the guides library.
Questions people ask
How much does a penetration test cost?
It depends on the scope: how many applications, interfaces, user roles, servers and networks are tested, whether the testers start with credentials or from outside, how deep the test goes and whether a retest is included. Pen tests are usually quoted in tester days, so the scope sets the price. Any figure given before the scope is written down is a guess.
Why do pen test quotes differ so much?
Mostly because they test different things. One quote may be an automated scan with a short report, another several days of manual testing by experienced testers with a retest. Compare the scope, the method, the number of tester days and the report, not the total alone.
How often should we do a penetration test?
At least before a new application or major change goes live, and then on a regular cycle that matches the risk, often once a year. Some rules set their own cycle: under DORA, identified financial entities must carry out threat-led penetration testing at least every three years.
Is a vulnerability scan the same as a penetration test?
No. A scan is an automated tool that lists known weaknesses; it is quick and repeatable. A penetration test is a person trying to break in within agreed rules, chaining weaknesses together and showing how far an attacker gets. Both have a place, and they are priced very differently.
What should a penetration test report include?
The scope and rules of the test, the method, each finding with its severity, evidence and the steps to reproduce it, the business impact, and clear advice on fixing it. A summary for management and a retest of the fixes make the report far more useful.
Sources
- Eurostat: EU hourly labour costs ranged from €12 to €57 in 2025 (31 March 2026, read 12 November 2026)
- Directive (EU) 2022/2555 (NIS2), Article 21 (read 12 November 2026)
- Regulation (EU) 2022/2554 (DORA), Article 26 (read 12 November 2026)
- OWASP Web Security Testing Guide (version 4.2, read 12 November 2026)
Your penetration test scope checklist
Write one line for each before you request quotes. The answers make quotes comparable.
- Why the test: assurance, a launch, a customer, an auditor or a regulator.
- Targets: applications, interfaces, addresses, cloud accounts, by name.
- User roles, with one test account each.
- Starting point: with credentials and documentation, or from outside.
- Environment: production or a test copy, and test windows.
- Systems owned by third parties, and their consent.
- Who reads the report, and what they need from it.
- Retest of fixes, and when.
- A contact on your side during testing.
- How test data and findings are stored and deleted.