What is penetration testing?
A penetration test is an authorised attempt to break into your own systems, carried out by security specialists so that you find the weaknesses before a criminal does. The testers use the same techniques a real attacker would use, but under a written agreement, within agreed limits and with care not to damage anything. At the end you know which weaknesses exist, which of them can really be used against you, and what to fix first.
People often confuse a penetration test with a vulnerability scan. A scan is an automated tool that checks systems against a list of known problems and prints a long report. It is quick, cheap and useful, and it is the first step of any test. But a scanner cannot think. It will not notice that one customer can read another customer's invoices by changing a number in the address bar, or that three small flaws can be chained into a complete takeover. Finding those takes a person, and that manual work is what makes a test a penetration test.
- Automated tool
- Finds known, listed weaknesses
- Many false alarms to sort
- Hours to run
- Good for monthly checks
- Manual work by a specialist
- Finds logic flaws and chained attacks
- Each finding proven
- Days to weeks
- Good before launch and yearly
Tests differ by how much the tester is told in advance. In a black-box test the tester starts with nothing but a name or an address, like an outsider. In a grey-box test they receive ordinary user accounts and some documentation, which is the most efficient option for most applications. In a white-box test they also see source code and architecture, which finds the most in the least time. The right choice depends on what you want to learn.
Tests also differ by target. A web application test looks at sign-in, sessions, access control, input handling and business logic. An API test examines the interfaces that apps and partners call. A mobile test covers the app on the phone as well as the server behind it. A network test looks at servers, firewalls and remote access. A cloud review checks how accounts, storage and permissions are configured. Most clients start with whatever faces the internet and holds customer data.
Finally, a test is a snapshot. It tells you about the system as it was during the test window. Software changes every week, so the value comes from testing regularly, fixing what is found and building the lessons into how you develop. That is why we treat the report as the middle of the job, not the end.
A good test therefore answers three questions for the people who run the business: how could someone get in, what could they reach if they did, and what should we fix first.
How we deliver a penetration test, step by step
Every test Organizein runs follows seven stages, based on the OWASP testing guides and the Penetration Testing Execution Standard. The method is the same for a three-day test of one web app and a month-long test of a whole estate. What changes is the depth of each stage. Nothing is touched until you have signed an authorisation, and you can stop the test at any moment.
Scoping and authorisation
We agree exactly what will be tested: addresses, applications, user roles and environments. We also agree what is off limits, the test window, and who to call on each side if something unexpected happens. You sign a letter authorising the test. Where systems run with a cloud or hosting provider, we check its testing policy and notify it when that is required.
Reconnaissance
The tester maps the target as an attacker would: domains and subdomains, exposed services, technologies in use, sign-in pages, forgotten test sites and information about the company that is publicly available. This often finds systems the client did not know were online, which are added to the scope only with your written approval.
Discovery
Automated scanners run first to cover known weaknesses quickly. The tester then works through the application by hand, role by role and function by function, looking for the flaws tools miss: broken access control, weak session handling, unsafe file uploads, injection points and errors in business rules such as discounts, limits and approvals.
Controlled exploitation
For each suspected weakness, the tester tries to prove that it can be used, going only as far as needed to show the impact. That might mean reading one record that should be private, not downloading a database. We do not run denial-of-service attacks, delete data or change live records unless you have asked for it in a test environment.
Analysis and risk rating
Each confirmed finding is scored with the Common Vulnerability Scoring System and then adjusted for your business: a medium-rated flaw on the payments page may matter more than a high-rated one on an internal wiki. Findings that can be chained together are described as one attack path, so you see the real route to harm.
Report and walkthrough
The report has two parts. A summary for management explains the overall position in plain language. A technical section gives each finding with its rating, where it is, how to reproduce it, evidence such as screenshots, and a specific fix. We then walk your developers through it on a call and answer their questions.
Retest
When your team has fixed the findings, we test each one again and update the report to show what is closed. The retest is included in the price if it takes place within the agreed period. On request we issue a short letter confirming that a test was done and that listed issues were resolved, which you can share with customers.
Who carries out the test
A test is led by a senior tester who plans the work, does the manual testing and writes the report, supported by a second tester on larger scopes. A different senior person reviews every report before it reaches you. You are given the names and background of the testers before the test begins, and you can ask about their experience and qualifications.
How we communicate during the test
You get a named contact and a private channel for the duration. We send a short note at the start and end of each test day. If we find something critical, such as exposed customer data or signs that someone else has already broken in, we do not wait for the report. We call your contact immediately so that you can act the same day.
How we protect your systems and data
Testing is done from fixed addresses that we give you in advance, so your team can tell our traffic from a real attack. We prefer a staging environment that mirrors production. When production must be tested, we agree quiet hours and avoid destructive techniques. Any data seen during a test is kept to the minimum needed as proof, stored encrypted and deleted after the engagement on an agreed date.
Confidentiality
We sign a non-disclosure agreement before scoping. Reports are sent through an encrypted channel to named recipients only. We never name a client or describe a finding outside the engagement without written permission, which is also why you will not see client case studies of security work on this site.
What a finding looks like in our report
Each finding fits on one or two pages and follows the same order: a title in plain language, the severity and its score, the affected address or function, a short explanation of the risk to your business, numbered steps to reproduce it, evidence, and a recommended fix with a reference. Developers should be able to reproduce and repair an issue from the report alone, without a call.
How we help after the report
Testers stay available to your developers during the fix period. They answer questions about a finding, review a proposed fix before it is released and suggest a simpler remedy when the ideal one is not practical. Where the same kind of flaw appears in several places, we explain the pattern so that it can be removed at the source, for example in a shared library.
How regular testing is organised
Clients who test more than once a year usually agree a calendar: a full test of the main application annually, shorter tests after major releases, and monthly vulnerability scans between them. Using the same team each time means testers already know the system and spend the days looking for new problems, not relearning old ground.
Weeks. An illustration of a typical order of work, not a quote.
What is included in our penetration testing service
The scope names each target and the type of test. These are the parts most engagements are built from.
Web application testing
Manual and automated testing of sign-in, sessions, access control, inputs and business logic.
API testing
Authentication, authorisation, rate limits and data exposure in REST and GraphQL interfaces.
Mobile app testing
iOS and Android apps: local storage, network traffic, reverse engineering and the server side.
Network testing
External and internal tests of servers, firewalls, VPN and remote access.
Cloud configuration review
Identities, permissions, storage and network rules in AWS, Azure and Google Cloud.
Vulnerability assessment
Scheduled scans with results checked by a person and false alarms removed.
Report and walkthrough
Management summary, technical findings with proof and fixes, and a call with your developers.
Retest and letter
A second test of fixed findings and a confirmation letter on request.
Each engagement is led by a person, not a tool. Scanners are used to save time on known issues, and the days in your quote are spent mainly on manual testing and on writing findings your developers can act on.
Who uses this service, and why
Most tests are ordered for one of four reasons. Knowing yours helps us set the scope so that the report answers the question you are really being asked.
Companies with a customer questionnaire
A large customer asks for evidence of a recent independent test before signing.
Teams preparing for certification
ISO 27001, SOC 2, PCI DSS and similar standards expect regular testing and evidence of fixes.
Products about to launch
You want to find serious flaws before real users and real data arrive.
Firms that had a scare
An incident or a near miss has shown that nobody is sure how exposed the business is.
If the driver is a certification, our compliance support team can align the test scope with the standard so that one piece of work serves both needs.
Black box, grey box or white box: how we choose
The more a tester knows at the start, the more of the paid time goes into finding problems instead of mapping the system. The table shows what each approach is suited to.
| Approach | What the tester receives | Best for | Trade-off |
|---|---|---|---|
| Black box | A name or address only | Seeing what an outside attacker sees | Less coverage for the same time |
| Grey box | User accounts and basic documentation | Most web, API and mobile tests | Needs test accounts for each role |
| White box | Accounts, source code and architecture | Critical systems and pre-launch reviews | Requires sharing code under NDA |
For most applications we recommend grey box. It tests what matters most, whether one user can reach another user's data or an ordinary user can act as an administrator, without spending days guessing at things you could simply tell us. Weaknesses found in code are easiest to prevent at the source, which our DevOps engineers do by adding security scans to the release pipeline.
What decides the cost and timeline of a test
A penetration test is priced by the tester days needed to cover the scope properly. We give a fixed price once we have seen the target or a description of it. These factors decide the number of days. The pricing page describes how fixed quotes work.
- The size of the application: number of screens, functions and API endpoints.
- The number of user roles, since access between roles must be checked in every direction.
- The approach: black-box tests need more time for the same coverage.
- The number of external addresses or internal hosts in a network test.
- Whether testing must happen outside business hours.
- Reporting needs, such as mapping findings to a particular standard.
Be wary of a quote that is far cheaper than the rest. It usually means a scanner report with a cover page. Ask how many days of manual testing are included and who will do them.
What we measure and report
A single test gives a picture. Repeated tests show a trend. For clients who test with us regularly we track these figures from one engagement to the next.
Findings by severity
The number of critical, high, medium and low findings in each test.
Time to fix
How long confirmed findings stay open, by severity.
Retest pass rate
The share of findings that are properly closed at the first retest.
Repeat findings
Weaknesses of the same type that return, which point to a gap in development practice.
Coverage
The share of applications, APIs and external addresses tested in the past year.
Exposed surface
Internet-facing systems found in reconnaissance that were not on your own list.
Mistakes we see around security testing
A test is only as useful as what happens before and after it. These are the ways we most often see the value of a test lost.
Testing once and filing the report
A report that leads to no fixes changes nothing. Assign each finding an owner and a date.
Scoping too narrowly
Testing the main site while the old admin panel sits on a forgotten subdomain misses the real risk.
Treating a scan as a test
An automated report cannot find broken access control or flawed business rules.
Testing only before an audit
Flaws found a week before the auditor arrives cannot be fixed in time.
Keeping developers out of the room
The people who must fix the findings should hear them explained and ask questions.
The 90 days after a test
The weeks after the report are where security actually improves. In the first fortnight, critical and high findings should each have an owner and a date, and we are available to review fixes as they are made. Around week four to eight, depending on your release cycle, we carry out the retest and update the report.
By the third month the useful question is no longer whether these findings are closed but why they existed. We offer a short session with your developers on the patterns behind the findings, and suggest checks that would catch them earlier, such as code scanning in the pipeline or a review step for access rules. The date and scope of the next test are agreed before the engagement closes, so that testing becomes routine instead of a reaction.
What to prepare, and what to ask any testing provider
Scoping is quicker and the quote more accurate when a few facts are to hand. The first list is what we ask for. The second is what we would ask any testing company before letting it near our systems.
What to prepare before scoping
- A list of the applications, APIs and addresses you want tested, and which hold customer data.
- The reason for the test: customer request, certification, launch or general assurance.
- Test accounts for each user role, in a staging environment if one exists.
- Any technical documentation, such as an API specification.
- The names of a technical contact and an emergency contact.
- Your hosting or cloud provider and any limits it places on testing.
Questions to ask any provider
- How many days of manual testing are in the quote, and who will do them?
- Which methodology do you follow, and can we see a sample report?
- How will you avoid disrupting our live systems?
- What happens if you find something critical during the test?
- Is a retest included, and within what period?
- How is our data stored during the test and when is it deleted?
Terms used on this page
Security reports can be hard to read. These are the terms that appear in ours, explained plainly.
- Vulnerability
- A weakness in software or configuration that could be used to cause harm.
- Exploit
- The technique or code that takes advantage of a vulnerability.
- CVSS
- The Common Vulnerability Scoring System, a standard way to rate severity from 0 to 10.
- Grey-box test
- A test in which the tester is given user accounts and limited documentation.
- Rules of engagement
- The written limits of a test: what may be tested, when and how.
- Retest
- A second check, after fixes, to confirm that each finding has been closed.
Questions about penetration testing
What is the difference between a penetration test and a vulnerability assessment?
An assessment lists known weaknesses using scanners. A penetration test adds manual work to find flaws tools miss and to prove which weaknesses can really be exploited.
How long does a penetration test take?
A single web application usually takes one to two weeks including the report. Larger scopes with several apps or networks take three to four weeks.
How much does a penetration test cost?
It is priced by the tester days needed for your scope. We give a fixed quote after a short scoping call or questionnaire.
Will the test disrupt our systems?
It should not. We avoid destructive techniques, prefer staging environments and agree quiet hours for production. You can pause the test at any time.
Do you need access to our source code?
Only for a white-box test. Most tests are grey box and need user accounts and basic documentation, not code.
Is penetration testing legal?
Yes, when the owner of the systems has authorised it in writing. We never test anything without a signed authorisation that names the targets.
How often should we test?
At least once a year and after any major change, such as a new application, a redesign of sign-in or a move to a new cloud setup.
What do we receive at the end?
A management summary, a technical report with each finding, its rating, proof and a recommended fix, a walkthrough call and a retest.
Is a retest included?
Yes, once, within the period agreed in the quote. The report is updated to show which findings are closed.
Can you test our mobile app?
Yes. We test iOS and Android apps, including data stored on the device, network traffic and the APIs behind the app.
Do you test cloud environments?
Yes. We review identities, permissions, storage and network rules in AWS, Azure and Google Cloud, following each provider's testing policy.
Will the test help with ISO 27001, SOC 2 or PCI DSS?
Yes. These standards expect independent testing. We can map findings to the relevant controls and provide evidence of the retest.
What standards do you follow?
The OWASP Web Security Testing Guide, the OWASP Mobile and API guidance, and the Penetration Testing Execution Standard, with CVSS for scoring.
Who will carry out the test?
A named senior tester leads each engagement. You receive the testers' names and background beforehand and can ask about their qualifications.
What if you find a critical issue?
We call your emergency contact immediately instead of waiting for the report, and give enough detail for you to act the same day.
Do you fix the problems you find?
The report explains how to fix each one. If you want help, our development and DevOps teams can do the repairs as separate work.
Can you run phishing simulations?
Yes, as agreed in writing with your management, with results reported in aggregate so that the aim is training, not blame.
How is our data handled?
We keep only the minimum needed as evidence, store it encrypted and delete it on the date agreed in the contract.
Do you sign an NDA?
Yes, before scoping. Security findings are never shared or referred to outside the engagement.
How do we get a quote?
Send the form on this page and say what you would like tested. We reply with a few scoping questions, then a fixed price and dates.
A ten-page guide to penetration testing
Further reading
Related services and reading
Guides
How Much Should a Website Cost?
Are you thinking of redesigning your already existing website or simply wish to launch a new one?
What Is CRM
Best Interaction services with Customers.
What Is Digital Marketing
Build Brand Recognition as an Industry Leader and Enhance Profitability
Further reading
Independent sources on the standards and platforms behind this work.