Penetration Testing Companies: How to Choose One, What It Costs, and What You Get
How to vet penetration testing companies: pen test vs vulnerability scan, test types, cost drivers, report quality, and when FTC, PCI or insurers require one.
By Scott McAuley · Nov 16, 2026 · 13 min read
Penetration testing companies are paid to break into your network on purpose, under a written contract, so you find out what an attacker could reach before one does. For a small business, the right firm is independent of your IT provider, scopes the work to a written asset list, has humans exploit what the scanners only flag, and hands you a report your insurer and auditor will accept.
Most owners meet the subject sideways. A cyber insurance renewal asks whether you have had a penetration test in the last 12 months, a large customer's security questionnaire asks for the executive summary, or a card processor mentions PCI. Then a search for a pen testing firm returns forty vendor listicles, each of which ranks its own company first. The useful question is not which company is best; it is which kind of test you need, how to tell a real test from a relabeled scan, and what the report has to prove. This guide expands the testing section of our small business cybersecurity guide into a buyer's walkthrough, and it ends with a readiness checklist you can hand to whoever runs your IT.
What Penetration Testing Companies Actually Do
A penetration test is an authorized, time-boxed attack against systems you own, carried out by people who think like attackers. The National Institute of Standards and Technology describes it in NIST SP 800-115, the Technical Guide to Information Security Testing and Assessment, as testing that mimics real-world attacks to find ways around a system's security features. The key word is mimics: the tester does not stop at a list of possible weaknesses. They try to use them, chain them together, and see how far they get.
A typical engagement has five stages:
- Scoping and rules of engagement — the written list of IP addresses, applications, and accounts in scope, the testing window, what is off-limits (production databases, patient-care systems), and who to call if something breaks
- Reconnaissance — what an outsider can learn about you from public records, DNS, exposed services, and leaked credentials
- Discovery and exploitation — automated scanning followed by manual attempts to exploit what was found
- Post-exploitation — once inside, how far the tester can move: from one laptop to a file server, from a standard user to a domain administrator
- Reporting and retest — findings ranked by real business risk, evidence for each, fixes, and a second pass to confirm the fixes worked
The firms in this market range from global consultancies to two-person boutiques to platforms that sell automated testing by subscription. Size matters less than method. A small firm with senior testers and a clean report often serves a 40-person business better than a large firm that sends its most junior staff.
What Is the Difference Between a Penetration Test and a Vulnerability Scan?
A vulnerability scan is automated: software compares your systems' versions and configurations against public vulnerability databases and lists what matches. A penetration test is human-led: a tester tries to exploit those weaknesses and shows what an attacker could actually reach. Scans measure breadth; pen tests prove impact. Most regulated businesses need both, on different schedules.
| | Vulnerability scan | Penetration test |
|---|---|---|
| Who does it | Software, reviewed by IT | Human testers, using tools |
| What it finds | Known weaknesses that match a signature | Exploitable paths, including chained low-risk issues |
| False positives | Common; needs triage | Rare; each finding is demonstrated |
| Typical cadence | Monthly to quarterly, and after changes | Annually, and after significant changes |
| What it proves | What might be wrong | What an attacker could do |
The confusion is not accidental. Some vendors sell an automated scan with a branded cover page as a penetration test. Regulators treat the two as different activities: the FTC Safeguards Rule lists penetration testing and vulnerability assessments as separate requirements, and PCI DSS says the same thing in separate requirements for scanning and testing. If a quote has no human hours in it, it is a scan.
Scanning still matters. It is the continuous part of the job, and it belongs inside your managed security program; vulnerability scanning is part of our cybersecurity services stack alongside endpoint detection and response. The pen test is the annual exam that checks whether that program actually holds up.
Types of Penetration Testing a Small Business Might Need
You rarely need every type. Pick by asking where an attacker would start and what they would want.
External network test. The attacker's view from the internet: your firewall, VPN, remote desktop gateways, email, and anything else with a public IP address. This is the most common starting point and the one insurers ask about most.
Internal network test. Assumes the attacker is already inside, through a phished laptop or a malicious insider, and measures how far they can move. For most small businesses this is where the uncomfortable findings live: flat networks, shared local admin passwords, and old service accounts with domain-wide rights.
Web application test. For customer portals, online booking, payment pages, and anything your business built or customized. It follows a methodology such as the OWASP testing guide and needs authenticated access to test properly.
Cloud and Microsoft 365 review. Tests the tenant configuration: conditional access gaps, legacy authentication, over-shared SharePoint sites, and admin accounts without phishing-resistant MFA. For an office that lives in Microsoft 365, this can matter more than the firewall.
Wireless and physical. Guest Wi-Fi that reaches the internal network, or a lobby port that hands out an address. Useful for clinics and offices with public waiting rooms.
Social engineering. Phishing campaigns and pretext phone calls against staff. Valuable, but often better handled as an ongoing phishing-simulation program than as a once-a-year test.
Red team vs penetration testing. A red team exercise is a longer, stealthier campaign with a goal (reach the payroll system) that also tests whether your monitoring notices. It is built for organizations with a security operations team to test. Most small businesses get more from a well-scoped penetration test.
When Is a Penetration Test Required?
A penetration test is required when a regulation, contract, or insurer says so, and for small businesses that happens more often than owners expect. The main drivers:
The FTC Safeguards Rule. Non-bank financial businesses (tax preparers, mortgage brokers, auto dealers, financial advisors) are covered. Under 16 CFR 314.4(d)(2), a covered firm without effective continuous monitoring must conduct annual penetration testing and vulnerability assessments at least every six months. Section 314.6 exempts firms that keep information on fewer than 5,000 consumers from that testing clause, so count before you assume either way. Our financial services practice works through this rule with clients.
PCI DSS. If you accept cards, PCI DSS v4.0.1 requirement 11.4 calls for internal and external penetration testing at least once every 12 months and after significant changes, plus testing of any segmentation you rely on to shrink your scope; the standard itself is in the PCI Security Standards Council document library. Whether it applies to you depends on your self-assessment questionnaire type: a business that sends all card data to a hosted payment page has a much lighter load than one that stores card data. Our sister company Lone Star Payments explains how to tell which questionnaire you file in its PCI compliance guide for small business.
HIPAA. The current Security Rule requires a risk analysis and periodic technical evaluation but does not name penetration testing. The proposed update HHS published in January 2025 would require penetration testing at least every 12 months and vulnerability scanning at least every six months; it had not been finalized as of this writing. Many practices test now anyway, because their risk analysis points to it. Our step-by-step security risk assessment walkthrough shows where testing fits.
Cyber insurance and customer contracts. Neither is law, but both are binding. Insurance applications increasingly ask about testing, and an inaccurate answer can undermine a claim later; we cover the current carrier expectations in cyber insurance requirements for Houston businesses. Large customers in healthcare, energy, and defense supply chains now ask vendors for pen test summaries in security questionnaires.
Frameworks. The CIS Critical Security Controls list penetration testing as Control 18, starting at Implementation Group 2, the tier meant for organizations that handle sensitive data. That matters in Texas: Senate Bill 2610, which took effect September 1, 2025 and created Chapter 542 of the Business and Commerce Code, shields businesses with fewer than 250 employees from exemplary damages after a data breach if they maintained a written cybersecurity program that meets the law's standard. The Center for Internet Security notes that the Texas law cites the CIS Controls. Actual damages are still recoverable, and the details belong with counsel, but a documented testing program is the kind of evidence that law rewards.
How Much Does a Penetration Test Cost?
The cost of a penetration test is set almost entirely by scope, so any price quoted before scoping is a guess. Published price ranges online come mostly from vendors selling their own tests, and they vary too widely to be useful. What actually moves the price:
- Asset count — the number of external IP addresses, internal hosts or subnets, and user accounts in scope
- Test types — external only, external plus internal, and whether web applications or a cloud tenant are added
- Application complexity — the number of user roles and pages in a web app; authenticated testing takes longer than anonymous testing
- Method — mostly automated with a human review, or mostly manual by senior testers
- Retest — whether confirming your fixes is included or billed again
- Compliance format — PCI or other framework-specific reporting adds work
How to buy well: ask for a fixed price tied to a written asset list, with the retest included and the tester's hours stated. Compare quotes on hours of human testing, not on the headline price. A quote that is far cheaper than the others with no hours stated is usually a scan.
There is also a free tier worth knowing about. CISA's Cyber Hygiene services offer no-cost vulnerability scanning to U.S. governments and to public and private critical infrastructure organizations. It is not a penetration test, but if you qualify it is a useful continuous check on your internet-facing systems.
How to Vet Penetration Testing Companies
The firms that rank highest in search are not necessarily the right fit for a 30-person office. Ask every candidate the same questions and compare the answers in writing.
Who will actually test? Ask for the named testers and their certifications. Common, meaningful ones include OffSec's OSCP, GIAC's GPEN, and CREST certifications. Certifications are not everything, but a firm that will not tell you who is testing is a firm that may subcontract it.
What is your methodology? A good answer references a published standard, such as NIST SP 800-115, the OWASP testing guide for applications, or the Penetration Testing Execution Standard, and describes how much of the work is manual.
Can we see a sample report? This is the single best filter. A redacted sample shows whether findings are demonstrated with evidence and written so a non-specialist can act on them.
How do you protect our data? Testers will touch sensitive systems. Ask about professional liability and cyber insurance, how evidence is stored and destroyed, and, for healthcare clients, whether they will sign a business associate agreement when they may encounter patient data.
What happens if something breaks? Ask for the rules of engagement, testing windows, and an emergency contact on both sides before work starts.
Is the retest included? A test without a retest leaves you with a list of problems and no proof they were fixed.
Red flags. A guarantee that you will "pass", a price with no scope, no named testers, a report delivered in hours, or a firm that also wants to sell you the fixes for everything it finds.
What a Good Penetration Test Report Contains
The report is the product. Insurers, auditors, and customers will read the executive summary; your IT team will live in the findings. A good one has:
- An executive summary in plain English: overall risk, the two or three findings that matter most, and what an attacker could have reached
- Scope and method — exactly what was tested, when, from where, and what was excluded
- Findings ranked by business risk, not just by a CVSS score, each with evidence (screenshots, commands, the data reached)
- Attack narratives showing how low-risk issues chained into a serious one, which is where pen tests earn their fee
- Specific remediation for each finding, written for whoever manages the system
- Retest results, ideally as a separate letter you can hand to a customer or carrier
If the report is mostly raw scanner output with a cover page, you bought a scan.
Should Your IT Provider Run Your Penetration Test?
Usually not, and the reason is independence, not skill. The team that designed and manages your network is the wrong team to grade it; they test against their own assumptions, and any finding is a finding against their own work. Auditors and insurers increasingly expect testing by a party independent of the people operating the environment.
The productive split is this: an independent testing firm attacks, and your IT provider prepares, coordinates, and fixes. Before the test, your provider builds the asset inventory, confirms backups and monitoring are in place, and arranges access and testing windows. During the test, they watch for alerts and answer questions. After it, they work the findings and schedule the retest. That is the arrangement we recommend to our managed clients; how this fits a broader security program is covered in our guide to managed IT security services.
One more scope item that is easy to miss: AI tools. AI phone agents, chat assistants, and workflow automations are new places your data lives and new credentials to protect. If you run them, include their accounts and integrations in scope. Talos Automation's guide to AI voice agents for business is a useful map of the data those systems touch.
The Pen Test Readiness Checklist
Hand this to whoever runs your IT and mark each line done, scheduled, or not applicable.
Before you request quotes
- Written list of external IP addresses, domains, and public applications
- Internal network map, including subnets, VLANs, and any segmentation you rely on for PCI scope
- Identify the driver: FTC Safeguards, PCI DSS, HIPAA risk analysis, insurer, or customer contract
- Decide which test types you need: external, internal, web app, Microsoft 365 or cloud
Before testing starts
- Signed rules of engagement with scope, windows, exclusions, and emergency contacts
- Business associate agreement or NDA signed where testers may see sensitive data
- Recent verified backup of critical systems
- Monitoring and EDR left on, with the IT team told to watch, not block, the named tester IPs
After the report
- Every critical and high finding assigned an owner and a date
- Retest scheduled and completed
- Executive summary and retest letter filed with insurance and compliance records
- Next test date set: 12 months out, or sooner after a significant change
Where This Lands
Choosing among penetration testing companies is mostly about choosing the right test and recognizing a real one: human-led, scoped in writing, independent of the people who run your network, and documented in a report that proves impact and confirms the fixes. The rest of the security baseline it tests against is in our small business cybersecurity guide.
This is standing work for our compliance support practice: we help Houston businesses work out which test they actually need, prepare the environment, coordinate with an independent testing firm, and fix the findings, then hand you the evidence in writing. If you want to know where you stand first, start with a no-obligation IT assessment.
Test before someone else does.
Scott McAuley is the founder and CEO of Texas Management Group, and founder of Talos Automation and Talk Is Cheap — 25+ years running IT, communications, and automation for Texas businesses.
Frequently Asked Questions
How much does a penetration test cost for a small business?
It depends almost entirely on scope, so treat any price quoted before scoping as a guess. The drivers are the number of external IP addresses and internal hosts, whether web applications or a cloud tenant are included, how much of the testing is manual, and whether a retest is included. Ask for a fixed price tied to a written asset list, with the retest priced in.
What is the difference between a penetration test and a vulnerability scan?
A vulnerability scan is automated: software matches your systems against public vulnerability databases and lists possible weaknesses. A penetration test is human-led: a tester tries to exploit those weaknesses and chain them together to show what an attacker could actually reach. Scans measure breadth; pen tests prove impact. Most regulated businesses need both, on different schedules.
How often should a small business get a penetration test?
Once a year and after any significant change is the common benchmark. PCI DSS requires internal and external testing at least every 12 months and after significant changes, and the FTC Safeguards Rule requires annual testing for covered firms without continuous monitoring. With no regulatory driver, annual testing plus monthly or quarterly vulnerability scanning in between is a sensible baseline.
Is penetration testing required for HIPAA compliance?
Not explicitly today. The current HIPAA Security Rule requires a risk analysis and periodic technical evaluation but does not name penetration testing. The proposed update HHS published in January 2025 would require it at least every 12 months, with vulnerability scanning every six months, but it had not been finalized as of this writing. Many practices test anyway because their risk analysis points to it.
Should our IT provider do our penetration test?
Usually not. The team that built and manages your network is the wrong team to grade it, and that is an independence problem, not a skills problem. Hire an independent testing firm, and have your IT provider scope the assets, coordinate access and testing windows, watch for alerts during the test, and fix the findings afterward.
About the Author
Scott McAuley is a Marine Corps veteran and 25-year IT executive. He is President & CEO of Texas Management Group, founder of Talos Automation, and creator of Talk Is Cheap, and was named IT Services CEO of the Year 2023. Full bio at scottmcauley.com →