Your attack surface changes every week. Code ships, cloud resources spin up, and third-party integrations come and go. A traditional penetration test captures a single moment in that change, so the report starts aging the day it is delivered.
That gap is why penetration testing deserves a closer look. This guide covers the main types, how the process works, what it costs and how often to run it, how to choose a provider, and how continuous testing compares with the traditional point-in-time model.
What Is Penetration Testing?
Penetration testing is an authorized, simulated cyberattack against an organization’s systems, applications, or people. Its purpose is to find and prove exploitable weaknesses before a real attacker does.
You will often hear it called ethical hacking. Testers use the same techniques an attacker would, but they work under an agreed scope and rules of engagement, and the goal is validated, prioritized risk, not damage.
Penetration Testing vs. Vulnerability Scanning
A vulnerability scan is automated and lists possible issues. A penetration test goes further. It validates which of those issues are actually exploitable and shows the real business impact.
The distinction matters when you buy. A scan tells you what might be wrong, a pentest proves what an attacker could do, and some vendors sell a glorified scan as a pentest.
Aspect | Vulnerability Scan | Penetration Test |
|---|---|---|
Method | Automated scanning for known issues | Manual and automated testing that exploits findings |
Depth | Broad coverage of known vulnerabilities | Deeper testing of how weaknesses combine |
Output | A list of potential issues | Validated findings with proof of exploitation |
Who performs it | Scanning software, often run by internal teams | Skilled penetration testers |
For a deeper comparison, read Penetration Testing vs Vulnerability Testing: 6 Key Differences.
Why Penetration Testing Matters
There are two main drivers: preventing breaches and meeting compliance requirements.
1. Breach Prevention
A pentest turns abstract risk into a prioritized, fixable list before an incident does. You learn which weaknesses an attacker could actually use, and you fix those first.
2. Compliance
Frameworks differ in how directly they ask for testing:
- PCI DSS: Explicit. Requirement 11.4 requires internal and external testing at least every 12 months and after any significant change.
- SOC 2: Doesn’t name penetration testing, but auditors commonly expect it as evidence that controls work.
- ISO 27001: Doesn’t prescribe a testing method, but testing is a common way to show technical vulnerabilities are managed.
- HIPAA: Doesn’t name it today. HHS proposed requiring annual testing in January 2025, but the rule is not final.
Since requirements change, confirm the current expectations with your auditor or assessor, including for frameworks such as HITRUST. Sprocket’s page on compliance frameworks such as SOC 2, PCI DSS, HITRUST, and ISO 27001 shows how testing evidence maps to each.
What Are the Types of Penetration Testing?
Penetration tests are categorized by how much the tester knows and by what they target, and most programs combine several.
By Knowledge Level
Black Box
In a black box test, the tester starts with no prior knowledge or access and does their own reconnaissance, the way an outside attacker would. It is the most realistic simulation of an external threat.
Gray Box
In a gray box test, the tester has partial information or limited credentials. This simulates an insider, or an attacker who already has a foothold through compromised access.
White Box
In a white box test, the tester has full access to architecture, documentation, and sometimes source code. It trades realism for thoroughness, which makes it well suited to a deep review.
By Target
External Penetration Testing
External penetration testing targets internet-facing assets such as domains, applications, APIs, and cloud services. It finds real-world exploitable risks and forgotten attack paths.
Internal Penetration Testing
Internal penetration testing simulates an insider threat or an attacker who has already breached the perimeter. Testers move laterally and escalate privileges across internal systems to show how far an attacker could get.
Web Application Penetration Testing
Web application testing tests modern web applications against real-world attack techniques. It finds vulnerabilities, including flaws in authentication, authorization, and business logic, before they become breaches.
Social Engineering
Social engineering testing tests the human layer through controlled phishing and vishing campaigns. It shows how employees respond and whether security awareness training holds up.
Scope can also extend to wireless, mobile, and cloud as part of a broader program. A full-scope, objective-based engagement is often called adversary simulation, or red teaming.
What Does the Penetration Testing Process Look Like?
Credible testing maps to recognized standards such as PTES, OWASP, MITRE ATT&CK, and NIST. Providers describe the steps differently, but most engagements follow the same six.
1. Planning & Scoping
The engagement starts by defining targets, rules of engagement, and objectives. Clear scope prevents wasted effort and blind spots, because testers know what is in bounds and nothing important gets left out.
2. Reconnaissance & Discovery
Testers map the attack surface to understand what is actually exposed. That includes domains, IP ranges, services, and assets that nobody formally tracks.
3. Scanning & Enumeration
Automated tooling surfaces candidate weaknesses. These are inputs to human testing, not conclusions on their own.
4. Exploitation
Testers attempt to exploit findings to confirm real impact. This is what separates a validated finding from a theoretical one.
5. Post-Exploitation & Lateral Movement
Once testers have a foothold, they show how far an attacker could go by escalating privileges and pivoting between systems. This reveals the true attack path, which is often a chain of several smaller weaknesses.
6. Reporting, Remediation & Retesting
Findings should arrive with proof of exploitation, severity tied to business impact, and clear remediation guidance. After your team applies fixes, a retest confirms that each fix actually worked.
How long does a penetration test take? A small application or environment can take days, while a large or complex one can take weeks, depending on scope and attack surface size.
To get the most out of a test, read how to prepare for a penetration test.
How Much Does Penetration Testing Cost?
There is no universal price, and a generic flat quote is a red flag. Cost is driven by scope and by what you get back, so two quotes at the same price can buy very different amounts of testing.
Most penetration tests fall between roughly $5,000 and $30,000, though individual quotes vary widely with scope. Typical ranges by test type look like this.
Test Type | Typical Range |
|---|---|
External network | $5,000 to $20,000 |
Internal network | $7,000 to $35,000 |
Web application | $5,000 to $30,000+ |
Full-scope adversary simulation (red team) | $50,000 to $150,000+ |
What Drives the Cost
Three factors set most quotes:
- Scope & attack surface size: Every additional application, IP range, API, user role, or environment adds tester time.
- Test type & complexity: An authenticated web application with many user roles takes more work than a small external perimeter.
- Tester expertise: Senior testers cost more per day, and they find the chained, complex issues that junior testers and tools miss.
Attack-surface-based scoping usually gives better value than counting individual IPs or targets, because it follows what is actually exposed instead of an inventory that may already be out of date.
The Hidden Costs of Traditional Pentests
The quote is rarely the whole cost. Buyers often miss three things:
- Retest fees: Some providers include one round of retesting after fixes, and others charge for it separately.
- Re-scoping & new SOWs: Anything left out of the original scope tends to come back later as a change order or a new statement of work.
- The exposure gap: Once the testers leave, new assets and changes go untested until the next engagement.
Pricing models differ too. Per-engagement pricing pays for one test at a point in time. Subscription pricing pays for testing over a period, and only some subscriptions, such as continuous penetration testing, keep testing as your environment changes.
Sprocket’s Approach
Sprocket handles retesting differently. Unlimited retests are included at no charge, with no re-scoping, no new statement of work, and no per-test or per-scope-change invoice. Sprocket’s pricing depends on the scope of your environment and the testing you need.
See Sprocket’s full pricing details or learn more about continuous penetration testing.
How Often Should You Perform Penetration Testing?
Many organizations test at least annually, often because of compliance. PCI DSS, for example, sets annual testing as a floor, and SOC 2 and ISO 27001 auditors commonly expect regular testing evidence.
More frequent testing is warranted in some cases. High-risk web applications, financial systems, and fast-changing cloud environments benefit from quarterly or continuous testing, because their exposure changes faster than a yearly cycle can track.
Certain events should also trigger a test, including major infrastructure upgrades, cloud migrations, significant application releases, mergers and acquisitions, and security incidents.
The core limitation of an annual test is plain. It leaves your environment unvalidated for the rest of the year, and attack surfaces change far faster than that.
Point-in-Time vs. Continuous Penetration Testing
Point-in-time testing gives you a snapshot. Continuous testing validates risk as your environment changes.
Aspect | Point-in-Time | Continuous |
|---|---|---|
Coverage between tests | None until the next engagement | Ongoing as the environment changes |
Speed of remediation | Findings often arrive in a final report | Findings arrive as they are validated |
Retesting | Limited or billed separately | Built in, so fixes are verified as they happen |
Compliance currency | Evidence ages quickly after the test | Evidence stays current through the year |
Cost model | Priced per engagement | Priced over a period, with scope that tracks the attack surface |
Sprocket’s term for this approach is continuous offensive security testing, or COST, and it positions COST as an alternative to PTaaS. The difference is time. PTaaS modernizes how penetration testing is delivered, with live dashboards, collaboration, and faster retesting. But each engagement is still scoped in advance and bound to a fixed window, so when the window closes, active testing stops while the environment keeps changing. That is why PTaaS is not the same as continuous testing.
Read more on why PTaaS is not continuous
How to Choose a Penetration Testing Provider
Choose a provider based on its ability to identify and validate real security risks, not just the features it offers. These seven factors separate providers:
- Credentials & certifications: Look for relevant qualifications, such as CREST accreditation, where applicable.
- Testing approach: Understand how the provider combines manual testing, automated tools, and human expertise.
- Evidence & reporting: Check whether findings include proof of exploitation, clear risk explanations, and practical remediation guidance.
- Retesting policy: Ask whether fixes can be retested and whether additional fees apply.
- Independence: Consider whether the provider offers an independent assessment of your security.
- Compliance support: Confirm whether the testing scope and reporting meet your relevant compliance requirements.
- Testing frequency: Compare one-off engagements with continuous testing to determine which fits your security needs.
When you compare quotes, ask for a sample report and the number of tester-days behind each price. That is the most reliable way to compare like with like.
How Sprocket Security Approaches Continuous Penetration Testing
Sprocket combines AI agents with expert human penetration testers.
- Agents plus experts. The agents, Apex and Link, run reconnaissance, discovery, and exploitation across web applications and APIs. Expert testers then confirm exploitability by hand and chain findings the way an attacker would.
- Every finding is verified. No finding reaches your report until a penetration tester has confirmed it.
- Testing that tracks change. Attack surface management monitors your assets for change, so a new subdomain or service gets tested as it appears.
- Findings you can act on. Each finding includes proof of exploitation, a step-by-step attack narrative showing how an attacker would use it, severity, and clear remediation guidance.
- Unlimited retesting. After you apply a fix, you can request a retest to confirm it worked, as many times as you need, with no additional cost, no new statement of work, and no re-scoping.
See how Sprocket tests to learn where agents lead and where humans lead
Make Penetration Testing Part of Your Security Program
A strong penetration testing program matches the types of testing to your risks, follows a clear process from scoping to retest, sets a realistic budget, and tests often enough to keep pace with change. The value of any pentest is proof of what an attacker could exploit, and that proof is only current if testing keeps up with your environment.
Knowing today’s exploitable risk, not last year’s, is what keeps a security program ahead of attackers.
Get Started With Sprocket
If you want to see validated, exploitable findings for your own environment, Sprocket can scope testing around your attack surface.