Pentest for web applications and APIs - White-box pentest with proof per confirmed finding, €1,000 to €2,000 excl. VAT, depending on the size of the application. Book an intro call · Pentest subscription
A vulnerability scan reports possible vulnerabilities. A penetration test actively looks for ways in and can show which of them can be exploited and how far an attacker gets with them. The name in an offer alone does not show the method. This article shows the differences using primary sources: BSI, NIST SP 800-115, the OWASP Web Security Testing Guide and the PCI SSC. It explains how to recognize a scan that is sold as a pentest, where automated pentests fit and when to use which method. How often to test is covered in How Often Should You Run a Penetration Test.
Table of contents
- Pentest vs vulnerability scan: the short answer
- Definitions: vulnerability scan and penetration test in the primary sources
- Pentest vs vulnerability scan: the difference in a table
- How to recognize a scan that is sold as a pentest
- Where automated pentests fit
- When to use which method
- Our approach at WZ-IT
- Common misconceptions
- Status and methodology
- Further guides
Pentest vs vulnerability scan: the short answer
A vulnerability scan compares systems automatically with known vulnerabilities and returns a list of possible findings. A pentest attacks the system like an attacker and can prove what can be exploited. The line is not clear-cut in the sources: an invasive vulnerability scan can, according to the BSI, use exploits and thereby prove that a system is vulnerable, while the exploit module in the BSI's pentest sequence is optional. In our view most applications need both: the scan frequently and broadly, the pentest after significant changes and at a fixed interval.
| Question | Short answer |
|---|---|
| What does a scan check? | Known vulnerabilities, outdated software and misconfigurations, by comparing versions with databases (NIST SP 800-115, section 4.3) |
| What does a pentest check? | Whether and how vulnerabilities can be exploited, including combinations of several weaknesses (NIST SP 800-115, section 5.2) |
| Where is the difference? | Mainly in the goal: the scan identifies, ranks and reports vulnerabilities, the pentest looks for ways to exploit them (PCI SSC, section 2.1). The proof alone does not separate the two: according to the BSI the exact proof only comes from exploitation, an optional module in the pentest guide's sequence (BSI guide, modules 3 and 4, in German), and an invasive vulnerability scan also uses exploits (BSI IS web check guide, section 3.2, in German) |
| Is automated the same as scan? | No. The BSI counts the largely automated IS web check among the pentests but requires manual verification. In our view, proof on the target and review of the severe findings by people show whether an automated method is a pentest; the source code determines the level of information (section Where automated pentests fit) |
| Is a scan enough? | For an overview of known weaknesses yes. As proof of exploitability a scan is usually not enough, while an invasive vulnerability scan can provide it with exploits. PCI DSS lists scans (11.3) and pentests (11.4) as separate requirements (PCI SSC, SAQ D) |
Definitions: vulnerability scan and penetration test in the primary sources
Vulnerability scan
In SP 800-115 (September 2008) NIST describes vulnerability scanning as a method that records hosts and their attributes and additionally tries to identify vulnerabilities. For this, operating systems and applications are identified and matched against information on known vulnerabilities in the scanner's database. According to NIST, network-based scans can have a high rate of false positives, and a person with expertise has to interpret the results.
In its practical guide to IS penetration tests (version 1.2, November 2016, in German) the BSI names the same mechanism: most vulnerability scanners decide from the versions of operating systems and applications whether a vulnerability exists. The result can be inaccurate because non-invasive scanners only query banners. According to the BSI a scan is worthwhile when a large number of systems has to be checked in a short time. An invasive scanner additionally tries whether vulnerabilities can be exploited with exploits, to gain certainty about their exploitability (module 3 of the guide). Tools exist as free and as commercial programs (section 2.2.4 of the guide).
Penetration test
The BSI defines in the same guide (our translation): "An IS penetration test is a proven and suitable approach to determine the attack potential on an IT network, an individual IT system or a (web) application." The success prospects of a deliberate attack are assessed. The BSI describes the exact proof like this: "The exact proof that a vulnerability exists only takes place when it is exploited." Exploitation is not part of every pentest, though: in the guide's sequence the "exploits" module is optional, and the BSI recommends a moderate attack strength: vulnerabilities are proven, but only actively exploited when this cannot be avoided and the exploits have been sufficiently tested. The guide knows several test depths, from a technical security audit through non-invasive vulnerability scans to the invasive vulnerability scan with exploits (sections 1.3 and 3.1.2). A pentest can therefore also report vulnerabilities that were not exploited.
NIST describes the pentest as security testing in which assessors mimic real-world attacks to find ways of circumventing security features. Most pentests look for combinations of vulnerabilities that together allow more access than a single weakness.
The OWASP Web Security Testing Guide (version 4.2) places the pentest as one technique among several: it should not be regarded as the primary or only testing technique. If source code is available, it should be given to the testers, because code can reveal vulnerabilities that a black-box test misses.
Grey area: the BSI IS web check
The BSI does not draw the line between "automated" and "manual". In its pentest guide it calls the IS web check a special case of the IS penetration test. According to the BSI page its tests are carried out "for the most part ... using automated methods via the Internet". The IS web check guide (version 1.6, January 2020, in German) adds the condition in its test sequence (our translation): "Every automated scan must therefore be verified and extended by manual methods." It also says that many false positives occur and that some vulnerabilities are missed because the scanners' intelligence is not sufficient to draw conclusions from results.
Pentest vs vulnerability scan: the difference in a table
The comparison follows the PCI SSC Penetration Testing Guidance (Information Supplement, version 1.1, September 2017, section 2.1) and adds BSI, NIST and OWASP.
| Feature | Vulnerability scan | Penetration test |
|---|---|---|
| Goal | Identify, rank and report vulnerabilities that could lead to a compromise if exploited (PCI SSC) | Find ways to circumvent security features; assess the success prospects of a deliberate attack (PCI SSC, BSI) |
| Approach | Usually a variety of automated tools combined with manual verification of identified issues (PCI SSC) | A manual process that may include scanners or other automated tools (PCI SSC); combination of several vulnerabilities (NIST) |
| Result | List of potential risks from known vulnerabilities. The ranking is not uniform: for PCI scans it follows CVSS for external scans and the organization's own risk-ranking process for internal scans (PCI SSC) | Description of each verified vulnerability and/or potential issue discovered, including how and to what extent it may be exploited (PCI SSC) |
| Proof | Comparison of versions and responses, false positives possible (NIST, BSI); an invasive vulnerability scan additionally uses exploits and thereby proves vulnerability (BSI IS web check guide) | The exact proof comes from exploitation of the vulnerability; in the BSI guide's sequence this module is optional (BSI) |
| Duration | Typically seconds to minutes per host (PCI SSC) | Days or weeks, depending on scope (PCI SSC) |
| Frequency | At least quarterly and after significant changes (PCI SSC) | At least annually and upon significant changes (PCI SSC); repeat tests every two to three years (BSI) |
| Limit | Comparison with databases of known vulnerabilities; according to NIST high false-positive and high false-negative rates for application scanners (NIST); according to OWASP application security assessment software is ineffective at in-depth assessment and test coverage (OWASP) | Snapshot without a guarantee that all vulnerabilities are found; according to the BSI this applies to the IS web check as it does to other security tests (BSI IS web check) |
The supplement dates from 2017 and refers to the numbering of PCI DSS at the time (requirement 11.3 for the pentest). In PCI DSS 4.0, scans are under 11.3 and pentests under 11.4. The frequencies in the supplement match the wording of requirements 11.3.1, 11.3.2, 11.4.2 and 11.4.3 in the PCI SSC's SAQ D (version 4.0; for the status of the version see "Status and methodology"): scans at least every three months, internal and external pentests at least every 12 months and after any significant infrastructure or application upgrade or change. According to the SAQ D, PCI DSS is intended for all entities that store, process or transmit cardholder data or could impact the security of the cardholder data environment. The deadlines therefore apply to entities within the scope of PCI DSS, not to every organization. According to the PCI SSC, a merchant clarifies whether and which attestation is required with the compliance-accepting entity, typically an acquirer or a payment brand (PCI SSC FAQ 1215).
Prices: anonymized ranges of published offers
Prices depend on scope and method. The table lists ranges from our own survey of publicly visible price information of several providers (as of 19 November 2026). The providers are deliberately not named and the survey data is not published, so the ranges cannot be verified from outside. They are not a market overview, and the services are not equivalent: the offers differ in method, application size and depth of reporting.
| Method | Published price range |
|---|---|
| Vulnerability scan | One-off scan of a server or a domain from a two-digit euro amount (gross consumer price); platforms with an annual fee from €0 in an entry plan up to four to five digits in euros |
| Manual web app pentest | Billed by effort; one provider states a day rate of around €1,600 per testing day; the total price follows from the day rate and the size of the application |
| Automated pentest per test | Around €1,250 to €3,500 or around US$3,000 to US$5,100 per single test; euros and dollars not converted |
How to recognize a scan that is sold as a pentest
The name "penetration test" in an offer alone does not show the method. Seven questions to the provider help to tell the two methods apart. They are a WZ-IT heuristic and derive from the BSI's requirements for preparation, execution and report of a pentest (BSI guide) and from PCI DSS 11.4.1 (SAQ D). No source defines these seven questions as a clear-cut rule. The column "Sign of a scan" is, unless marked otherwise, our classification.
| Question to the provider | Sign of a pentest | Sign of a scan |
|---|---|---|
| How are test object, scope and authorization agreed? | Written order with test period, test object and test depth (BSI) | Start by entering a domain, with no agreement on scope and depth |
| Which access is used? | Credentials and, where needed, tokens or certificates for protected areas (BSI IS web check guide) | Only a logged-out view from outside (our assessment) |
| How is each finding proven? | Description of each verified vulnerability and/or potential issue discovered, stating how and how far it can be exploited (PCI SSC) | Entry with CVE number and base score, without proof on the target |
| Who verified the findings? | Verification by manual methods (BSI IS web check guide) | Raw output of the tool |
| Which methodology is used? | Documented methodology with industry-accepted approaches, e.g. OWASP WSTG (PCI DSS 11.4.1) | No statement on methodology |
| What does the report contain? | Test object and test conditions, management summary, technical description per vulnerability with recommendation (BSI) | List of possible risks by severity |
| What was not tested? | Coverage and limitations are stated (our assessment) | No statement on coverage |
A single feature proves nothing. If proof and verification are missing, the method is, in our classification, a scan, regardless of the name in the offer. The sources describe both methods but do not state this rule expressly.
Where automated pentests fit
The sources do not draw the line uniformly by manual work. The PCI SSC describes the pentest as a manual process that may include tools. The BSI, by contrast, counts the largely automated IS web check among the penetration tests but requires manual verification in its sequence. From this we derive three questions by which an automated method can be described. Criteria 1 and 3, proof and review by people, are in our view the most telling features for whether a method is a pentest or a scan. Proof alone does not separate them, because an invasive BSI vulnerability scan can also prove vulnerability. Criterion 2, the source code, determines the level of information and thereby what a test can detect, not the classification: the BSI counts black-box tests and the PCI SSC black-box and gray-box tests among the pentests, but the BSI recommends white-box tests. The classification is our reading of BSI, OWASP, NIST and PCI SSC, not a definition by these bodies.
| Criterion | Question | Source |
|---|---|---|
| 1. Proof on the target | Is the finding proven on the running target, where possible by exploitation, and can the proof be reproduced? | BSI: the exact proof only takes place through exploitation, an optional module in the guide's sequence (guide); an invasive vulnerability scan also proves vulnerability with exploits (IS web check guide) |
| 2. Source code (level of information) | Is the source code included in the test? | BSI generally recommends white-box tests, i.e. extensive information about the target (guide); OWASP: source code should be given to the testers (WSTG) |
| 3. Review by people | Does a person review the severe findings before they appear in the report? | BSI: verification by manual methods (IS web check guide); OWASP: false positives are to be weeded out before the report (WSTG) |
If proof or review by people is missing, we classify an automated method as a scan. If source code is missing, vulnerabilities can go undetected: OWASP names vulnerabilities that can be found in source code and that a black-box test misses. Whether the test is then a pure black-box test depends on the remaining level of information; OWASP also describes the gray-box test with partial knowledge of the application. If proof and review by people are given, we classify the method as a pentest with automated means; whether source code is also included determines the level of information of the test. Whether a run covers permission and logic flaws can be checked in the report: a coverage table per role and function states what was tested and what was not (our assessment).
For the network layer the PCI guidance states (section 4.2.2) that simply running an automated tool does not satisfy the penetration testing requirement; the tester must interpret the results and determine whether additional testing is needed. Whether a specific automated run satisfies PCI DSS 11.4 is not answered in this article.
Two statements in the sources limit what a test achieves. The OWASP guide quotes Gary McGraw: in practice, a penetration test can only identify a small representative sample of all possible security risks in a system. The BSI states that an IS web check is only a snapshot and gives no guarantee that all vulnerabilities are found.
When to use which method
| Situation | Method | Reason |
|---|---|---|
| Many servers and services, ongoing overview of known weaknesses | Vulnerability scan | According to the BSI, scans are worthwhile when many systems must be checked in a short time (guide) |
| Web application after commissioning or after a significant change | Pentest | According to the BSI a pentest is usually sought after first commissioning or significant changes (guide) |
| Check roles, tenants and permissions | Pentest with test accounts | Our assessment: permissions and tenant separation can be checked with test accounts per role. According to OWASP, application security assessment software is ineffective at in-depth assessment and test coverage (OWASP) |
| Proof whether a reported finding is exploitable | Pentest | The exact proof only takes place through exploitation, an optional module in the BSI guide's sequence. According to the guide, a non-invasive scan does not exploit vulnerabilities, an invasive scan uses exploits (BSI) |
| Check configuration and hardening of servers, cloud and network | Security audit | Simplified WZ-IT classification: review of configuration and hardening rather than application logic. The BSI describes technical security audits based on versions, configuration and patch level (guide) |
| Static website without accounts and inputs | A scan is often enough | Our assessment: little application logic that a pentest could additionally check |
| Evidence for customers, auditors or supervisors | Method as specified by the recipient | Our assessment: method and interval usually follow from the requirement of the body that asks for the evidence. For PCI DSS, a merchant clarifies the attestation and its scope with the compliance-accepting entity (PCI SSC FAQ 1215); overview in the article How Often Should You Run a Penetration Test |
For web presences the BSI recommends a sequence: first an IS quick revision of the basic protection, then the web application without an upstream security gateway with an IS web check, finally the gateway and further interfaces with an IS penetration test (guide, section 2.1.2). A pentest also does not replace quality assurance: the BSI recommends reproducible tests that run again after every change.
Our approach at WZ-IT
We offer both methods separately. The vulnerability scan checks systems automatically for known vulnerabilities. For web applications and APIs there is the automated white-box pentest. This is how it runs:
- Handover: you hand over the source code, a running test instance and test accounts per role and tenant.
- Ownership check: we verify ownership of the target.
- Inventory: we record the routes of the application in an inventory.
- Attack: the attacks run against the test instance in an isolated test environment, never against the production system.
- Proof: confirmed findings come with proof. Every finding carries one of four states: confirmed, refuted, suspicion without confirmation, or not tested.
- Manual review: the team reviews severe confirmed findings by hand.
- Report: the report contains a coverage table that shows what was tested and what was not.
This sequence describes the WZ-IT offer and is not a statement by the cited sources.
The one-off pentest costs €1,000 (small), €1,500 (medium) or €2,000 (large) excl. VAT, as a fixed price by size of the application. Anyone who wants to test regularly uses the pentest subscription, quarterly or monthly. The subscription costs, excl. VAT per year, €2,500 (small), €3,750 (medium) or €5,000 (large) with quarterly tests and €4,000 (small), €6,000 (medium) or €8,000 (large) with monthly tests. On request we fix the findings and take over hosting and operations.
Common misconceptions
| Misconception | Correct | Source |
|---|---|---|
| A vulnerability scan is a small pentest. | The scan identifies, ranks and reports vulnerabilities. The pentest looks for ways to exploit them and can prove exploitability. An invasive scan can additionally use exploits but does not replace the pentest. | PCI SSC, BSI, BSI IS web check guide |
| A pentest finds all vulnerabilities. | It shows what was found at the time of testing with reasonable effort and the agreed methods, with no guarantee of completeness. | BSI IS web check |
| Automated always means scan. | The BSI counts the largely automated IS web check among the pentests but requires manual verification. In our view, proof and review by people count. | BSI, BSI IS web check guide |
| A clean scan report means: secure. | Scanners produce false positives and miss vulnerabilities: NIST names high false-positive and high false-negative rates for application scanners, OWASP calls application security assessment software ineffective at in-depth assessment and test coverage. | NIST, OWASP |
| Anyone who scans does not need a pentest. | PCI DSS lists both as separate requirements with different intervals. | PCI SSC |
| A pentest replaces quality assurance and code review. | The BSI says no for quality assurance: an IS penetration test does not replace quality assurance, which must be integrated into the lifecycle. The BSI calls code review a subfield of quality assurance. OWASP: many flaws are hard to find with other forms of testing such as penetration testing; source code analysis is the technique of choice for them. | BSI, OWASP |
| White box is less realistic than black box. | The BSI generally recommends white-box tests, because black-box tests can miss vulnerabilities. | BSI |
Status and methodology
Research status: 19 November 2026. The cited documents were retrieved on that date and the quoted passages checked against their wording. The sources differ in age: BSI guide to the pentest 2016 (version 1.2), BSI guide to the IS web check 2020 (version 1.6), short description of the IS web check 2022, NIST SP 800-115 from 2008, PCI SSC Information Supplement from 2017, OWASP Web Security Testing Guide version 4.2 (version 5.0 is in development). The sources describe the separation of scan and pentest in a similar but not identical way: OWASP also calls the pentest black-box testing and puts a combination of several techniques at the centre, the PCI SSC describes it as a manual process, the BSI distinguishes black-box and white-box tests and classes the largely automated IS web check as a special case. In our assessment individual tools and attack types have changed since the older sources; we did not compare specific versions. The BSI guides cited here are German-language documents (the BSI's English overview page links to them); quotes from them are our translations.
The wording of PCI DSS 11.3 and 11.4 comes from the PCI SSC's SAQ D for PCI DSS 4.0 of April 2022. The PCI SSC has announced that PCI DSS 4.0 will be retired on 31 December 2024, after which version 4.0.1 of June 2024 is the only supported version. It contains no new or deleted requirements but clarifies some of them (PCI SSC blog). The SAQ D for 4.0.1 and the main document of the standard were not reachable for automated retrieval (HTTP 403). We therefore did not check whether the quoted deadlines are word-for-word unchanged in 4.0.1.
The price ranges come from our own survey of publicly visible price information of several providers as of 19 November 2026; the providers are deliberately not named. The survey data is not published, so the ranges cannot be verified from outside. Prices can change, and not every price page gives a date. The seven provider questions, the three questions for automated pentests, the assignments in "When to use which method" and the statements marked "our assessment" are our classification. We deliberately make no statement on which test satisfies a standard. That depends on the respective framework and the body that assesses it.
Further guides
- How Often Should You Run a Penetration Test: intervals by standard, triggers and release rhythm
- Automated white-box pentest for web applications and APIs
- Pentest subscription: quarterly or monthly
- Vulnerability scanning in operations
- Security audit for servers and cloud
- AI code audit, code review or penetration test: which one fits?
Pentest or scan for your application? We assess your application and tell you which method makes sense. Book an intro call · Request a pentest
Sources
- BSI: Practical guide to IS penetration tests (Praxis-Leitfaden für IS-Penetrationstests), version 1.2, November 2016, in German
- BSI: Practical guide to the IS web check (Praxis-Leitfaden für den IS-Webcheck), version 1.6, January 2020, in German
- BSI: IS web check, short description, as of 29 June 2022, in German
- BSI: IS penetration tests and IS web checks
- NIST SP 800-115: Technical Guide to Information Security Testing and Assessment, September 2008
- OWASP Web Security Testing Guide, version 4.2
- OWASP Web Security Testing Guide: Introduction and Objectives
- PCI SSC: Information Supplement Penetration Testing Guidance, version 1.1, September 2017
- PCI SSC: SAQ D for service providers, PCI DSS 4.0, April 2022
- PCI SSC: FAQ 1215, What is a PCI DSS Self-Assessment Questionnaire?, April 2024
- PCI SSC blog: Just Published, PCI DSS v4.0.1











