How To Prepare for a Penetration Test
Preparation decides how useful a test is. This checklist gets your business ready.

Key Takeaways
- Agree scope and objectives in writing.
- Tell the people who need to know and avoid surprising suppliers.
- Back up key systems before testing.
- Plan time to fix what is found.
A penetration test is more valuable when the groundwork is done. Poor preparation wastes tester time, causes confusion and sometimes disruption.
This guide answers the questions clients ask before a test. It reflects our experience delivering penetration testing and links to NCSC guidance. For the basics, read penetration testing explained.
Why does preparation matter?
A test is only as good as its scope and information. Testers who do not know what matters spend time on low-value targets. Teams who are not warned may mistake the test for an attack.
Preparation also makes sure the result is something you can act on.
↑ Back to ContentsHow do you define scope and goals?

Decide which systems are in scope, what you want to learn and any limits. Write it down with your tester, including systems that must not be touched.
Think about business priorities: customer data, payment systems and anything that would halt trading.
↑ Back to ContentsWhat information should you provide?
Provide network diagrams, application details, user roles and contact points. For tests with credentials, prepare accounts with the right permissions.
Good information helps testers find real weaknesses instead of guessing.
↑ Back to ContentsWho needs to be told?
Tell your IT team, hosting providers, managed service providers and anyone who monitors your systems. Some providers require prior approval for testing.
Decide who inside the business knows the test date and who does not, if you want to test detection.
↑ Back to ContentsWhat are rules of engagement?
These set testing windows, permitted techniques, escalation contacts and what to do if something breaks. They protect both you and the tester.
Include a named emergency contact available during the test.
↑ Back to ContentsWhat should you back up first?
Take backups of key systems and confirm you can restore them. Testing is designed to avoid damage, but systems can behave unexpectedly.
Check that monitoring and logging are on so you can see what happens.
↑ Back to ContentsHow do you handle third parties?
Cloud and hosting providers may have testing policies. Check them before you start. Suppliers who run your systems must agree to testing of their components.
Get written permission where needed.
↑ Back to ContentsHow should you handle findings during the test?
Agree how critical issues will be reported immediately, rather than waiting for the report. Decide who can authorise urgent fixes.
Keep a log of changes made during testing so results stay accurate.
↑ Back to ContentsWhat should you do afterwards?
Review the report with the tester, prioritise fixes, assign owners and set dates. Schedule a retest to confirm closure.
Share a summary with the board and, where relevant, customers.
↑ Back to ContentsWhat do we see go wrong?
In our experience, common problems include vague scope, missing contacts, unaware suppliers and no capacity to fix findings. All are avoidable with a short preparation meeting.
The best results come from teams that treat the test as a joint exercise.
↑ Back to ContentsWhat should a pre-test checklist include?
A practical checklist covers eight points: a signed scope document, named contacts on both sides, testing dates and windows, a list of systems excluded from testing, confirmation that backups work, notification of hosting and service providers, test accounts if credentials are needed and agreement on how critical findings will be reported. Tick each item off in writing so there is a record.
Share the checklist with your IT support provider early. They know which systems are fragile, which are shared with other customers and who must give permission, and they can often spot an omission in minutes that would otherwise surface halfway through the test.
↑ Back to ContentsHow do you decide what is in and out of scope?
Begin with the assets that would hurt most if compromised: customer databases, payment pages, remote access gateways and email. Then list the routes an attacker would use to reach them, such as your website, a VPN, a cloud tenant or a supplier connection. Those routes belong in scope even if they feel less important.
Be deliberate about exclusions. Production systems that cannot tolerate any risk, fragile legacy equipment and third-party platforms you do not control are common exclusions, but each one should be recorded with the reason. Anything excluded is untested, and the report will say so.
↑ Back to ContentsHow should you prepare your own staff and monitoring?
Decide whether your security monitoring should react to the test. If you want to test detection, tell only a small group and let the alerts fire. If you only want to find vulnerabilities, tell the monitoring team and give testers a list of source addresses so that alerts can be tagged rather than escalated.
Either way, brief the helpdesk. A phone call saying that the website appears to be under attack is much easier to handle when the helpdesk knows a test is under way. A short note with dates and a contact number is enough.
↑ Back to ContentsWhat do you need to check with cloud and hosting providers?
Major cloud platforms allow certain testing without prior approval, but conditions apply and they change. Hosting companies and managed service providers often have their own rules, and shared hosting can be off limits entirely. Check the current policy for every provider whose systems are in scope.
Keep written confirmation. If a provider’s monitoring blocks the testers, or if a shared system is affected, you will want proof that the activity was authorised and agreed in advance.
↑ Back to ContentsHow do you handle personal data during a test?
Testers may see personal data if a vulnerability gives them access to it. Under UK GDPR you should limit that exposure by agreeing in advance what testers will do when they find data, such as recording that access was possible without copying records. Include confidentiality terms in the contract.
Ask how the tester stores evidence, how long they keep it and how it is destroyed. Reputable providers will already have clear answers and will be happy to put them in writing.
↑ Back to ContentsWhat should happen on the day of the test?
Keep a named contact available throughout, with a mobile number, and make sure someone with authority can approve emergency changes. Agree how often testers will update you, particularly if they find something critical such as exposed credentials or an open database.
Avoid making unrelated changes to systems during the test window. A new firewall rule or a software update can confuse results and make findings harder to reproduce later.
↑ Back to ContentsWhat are the most common mistakes before a test?
The most common mistakes are a vague scope, no named contacts, forgetting to tell a hosting provider and starting without a tested backup. Another is choosing a test date that coincides with a system change or a busy trading period, which makes any disruption more costly.
A second group of mistakes comes after the test: reading the summary, filing the report and never fixing the findings. Treat the report as a project plan, not a certificate.
↑ Back to ContentsHow long before the test should you begin preparing?
Allow at least two to four weeks. That gives time to agree scope, collect diagrams and accounts, ask providers for permission and brief staff. Complex environments, or those with several third parties, need longer.
Starting early also gives you room to fix obvious problems first, such as unsupported software or default passwords. Testers will find these anyway, and it is better to spend their time on subtler issues.
↑ Back to ContentsWho inside the business should be involved?
Involve IT, a senior sponsor, whoever owns the systems being tested and, where relevant, legal or compliance. The sponsor approves scope and risk, IT supports the testers and owners deal with findings.
Name a deputy for each role. Tests that stall because one person is away waste time and money. Our IT support teams often coordinate this on behalf of clients.
↑ Back to ContentsHow do you prepare web applications for testing?
Provide a test environment if you can, with realistic data but no real personal records. Prepare accounts for each user role, a list of features and any known weak points. Tell testers about integrations such as payment gateways, which may need separate permission.
If testing the live system, schedule it outside peak trading hours, make sure logging is on and agree what to do if something breaks. A clear picture of how the application works saves testers hours of discovery.
↑ Back to ContentsHow do you prepare network and cloud environments?
Provide diagrams, address ranges and a list of services exposed to the internet. For cloud environments, share the account structure and confirm which roles the testers will use. Check that logging and alerting are working so you can see the activity.
Remove surprises: undocumented systems, forgotten test servers and unmanaged devices will be found by testers, and it is better that you know about them in advance. Our IT support teams can help audit what is connected.
↑ Back to ContentsWhat happens if the test causes an outage?
Agree the plan beforehand: who stops testing, who restores service and who informs customers. Make sure backups are recent and that the restore procedure has been tested. Testers should pause immediately when told.
After any disruption, review what happened and adjust the scope or approach. Occasional minor disruption does occur, which is why preparation matters. Our penetration testing service follows an agreed escalation process, and we support businesses in Manchester and Salford.
↑ Back to ContentsHow do you manage third-party applications during a test?
Software-as-a-service tools such as email, CRM and accounting platforms belong to their vendors and are usually out of scope. Test your configuration, such as sign-in settings and permissions, rather than the vendor’s systems.
Check each vendor’s policy before including anything. Some allow limited testing, others prohibit it entirely, and unauthorised testing could breach your contract.
↑ Back to ContentsWhat documentation should you gather in advance?
Collect network diagrams, asset lists, lists of internet-facing services, application architecture, previous test reports and policies on password management and remote access. Even imperfect documents help testers focus.
Note known issues you have already accepted or scheduled to fix. It avoids wasting time on them, and it shows testers where to look for related weaknesses.
↑ Back to ContentsHow do you keep staff informed without causing alarm?
Send a short note to IT and the helpdesk with dates, source addresses and a contact. For the wider business, a brief message that security testing is planned, without detail, avoids confusion if someone notices unusual activity.
If social engineering is in scope, agree who knows in advance and how staff will be debriefed. Done well, it becomes a learning exercise. Our IT support team can coordinate communication.
↑ Back to ContentsHow do you get value from the report afterwards?
Read the management summary first, then work through findings by severity. Assign each a named owner, a target date and a note of how the fix will be verified. Group related findings, because a single change often closes several issues at once.
Book the retest before you finish the first review, while the findings are fresh. If you want a partner for the fixes, our penetration testing team can advise, and our phishing simulator shows how staff respond to the human side of attack. Our guide to penetration testing explained covers what a good report contains. We support businesses in Manchester, Salford and Warrington.
↑ Back to ContentsFrequently Asked Questions (FAQs)
How long does a penetration test take?
From a few days to several weeks depending on scope.
Will a penetration test disrupt our systems?
Testers take care to avoid disruption, but there is some risk. Agree testing windows and backups first.
Do we need to tell our hosting provider?
Often yes. Check the provider’s policy before testing begins.
What access does a tester need?
It depends on the test type. Some use only public access, while others need accounts to test what a user could do.
Should staff know about the test?
Key technical staff and suppliers should. For some tests, you may keep wider staff unaware to test detection.