A penetration test checklist is the difference between a test that tells you something useful and one that burns five days of consultant time arguing about IP ranges, waiting for VPN accounts and tripping over a firewall nobody remembered to whitelist. Most of the value of a pen test is decided before the tester sends a single packet: in the scoping call, the rules of engagement, the asset inventory and the people you line up to respond when something interesting turns up.
This guide is a practical, UK-focused pre-test checklist for preparing for a pen test. It walks through what to decide, what to document, who to tell and what to switch on (or deliberately leave alone) in the weeks before testing starts. It covers pen test scoping, legal authorisation, CHECK and CREST considerations, cloud provider rules, backups and change freezes, and how to set up the penetration testing process so the final report turns into fixes rather than a PDF that sits in a shared drive. Whether you are commissioning your first external infrastructure test or your fifth annual web application assessment, the same preparation steps apply.
What is penetration test preparation, and why does it matter?
Penetration testing is an authorised, simulated attack on your systems by a qualified tester who tries to find and exploit weaknesses the way a real attacker would. A vulnerability scan tells you that a server is missing a patch; a penetration test tells you whether that missing patch, combined with a weak service account and a flat network, lets someone reach your finance share from the guest Wi-Fi. The tester’s time is the expensive part — typically £850 to £1,400 per day in the UK for a CREST-accredited consultant — so every hour lost to logistics is an hour not spent finding problems.
Preparation is everything that happens between “we should get a pen test” and the first day of testing. It includes deciding why you are testing (an insurer’s questionnaire, a client’s supplier assurance demand, PCI DSS, Cyber Essentials Plus groundwork, a new application going live), defining exactly what is in and out of scope, getting written authorisation from everyone who owns the assets, preparing accounts and access, and agreeing how findings will be communicated. Skipping any of these steps rarely stops a test from happening; it just quietly reduces what you get from it.
In our experience, the single most common outcome of poor preparation is not a disaster — it is a bland report. The tester could not get credentials until day three, the staging environment did not match production, half the IP ranges turned out to belong to a former hosting provider, and the findings that did appear were never assigned an owner. A tight vulnerability assessment checklist and pre-test plan fixes all of those. It is the same discipline we recommend for end-of-life estates in our guide to Cyber Essentials and unsupported software: know what you own before anyone tests it.
Write a one-paragraph “test objective” before you request quotes. “Find out whether an unauthenticated attacker on the internet can reach customer data in our CRM” produces far better quotes — and far better reports — than “we need a pen test for the insurance renewal”.
Pen test types compared: black-box vs grey-box preparation
The first scoping decision shapes almost every item on your penetration test checklist: how much will the tester know, and how much access will they be given? Black-box testing simulates an outsider with no inside knowledge. Grey-box testing gives the tester credentials, architecture diagrams and sometimes source code, so they can spend their days finding weaknesses rather than rediscovering your network map. For most UK SMEs with a fixed budget, grey-box gives considerably more coverage per pound.
Black-box test
Tester starts with only a company name or IP range
Grey-box test
Tester receives credentials, diagrams and context
There is also white-box testing, where the tester has full source code and configuration access. It is common for bespoke applications and code-heavy environments, and it demands the most preparation: repository access, build instructions and a developer on call. Whichever model you choose, write it into the scope document explicitly. Disagreements about “how much the tester was supposed to know” are one of the most frequent sources of friction in the final report review.
A useful rule of thumb: if your main worry is an attacker who has phished one member of staff — which, according to the UK Government’s Cyber Security Breaches Survey, is by far the most common route into UK organisations — then a grey-box internal or assumed-breach test is closer to reality than a black-box external test. Phishing is the starting point; the pen test should tell you how far it goes from there.
Pen test readiness scoring — where most UK businesses sit today
Before you book, it is worth an honest self-assessment. The cards below show the preparation areas we most often see in good shape, the areas that usually need work, and the ones that regularly derail tests on day one. Use them to spot where your own pre-test effort should go.
None of the high-risk items are technically difficult. They are coordination problems: getting a signature from a hosting provider, telling a managed SOC not to block a particular IP, or creating a test user that can actually pass Conditional Access. That is why they get left until the last minute — and why a written checklist matters.
The penetration testing process by the numbers
Penetration testing in the UK is a mature market with fairly consistent pricing and durations. The figures below are typical ranges for SMEs and mid-market organisations commissioning tests from CREST-accredited suppliers in 2026. They are indicative rather than quotes — scope drives the final number — but they help set expectations when planning your preparation timeline.
The lead time is the figure most organisations underestimate. Good testing firms book up several weeks ahead, particularly in the run-up to financial year ends and insurance renewal seasons. If an insurer or a client has given you a deadline, work backwards: report needed by the end of the month means testing in the second week, which means scope signed off and access ready at least two weeks before that.
The reporting gap matters too. Most suppliers will flag critical findings by phone or email during the test, but the formal report — with risk ratings, evidence and remediation advice — usually lands one to two weeks after testing finishes. Build that into any deadline you have been given.
Where pre-test preparation time actually goes
When we track preparation effort across engagements, the distribution is remarkably consistent. Scoping and asset discovery dominate, followed by access set-up and third-party authorisation. Technical preparation on the systems themselves is a relatively small slice — most organisations should not be “hardening for the test” anyway, because that hides the very weaknesses you are paying to find.
The practical lesson is that the person running preparation does not need to be your most senior engineer. They need to be organised, able to chase signatures, and empowered to ask awkward questions such as “who actually owns this subdomain?” An office manager or operations lead paired with your IT provider is often the most effective combination.
Asset inventory work also pays off long after the test. The list of domains, IP ranges, cloud subscriptions and SaaS tenants you assemble for scoping is exactly what you need for Cyber Essentials, insurance questionnaires and supplier due diligence. Keep it as a living document rather than a one-off spreadsheet.
Pen test scoping benchmarks: how prepared are UK organisations?
The progress bars below reflect the proportion of organisations we see arriving at a scoping call with each item already in place. They are drawn from our own pre-test reviews of UK SMEs and mid-market firms rather than a national survey, but they line up with what CREST member firms report anecdotally: the documentation and coordination items lag well behind the technical ones.
Pre-test items in place at the first scoping call
Two numbers stand out. Only around a third of organisations can show that findings from their previous test were closed — which means the new test frequently rediscovers the same problems at full day-rate. And only about a quarter set aside budget for remediation before the test begins. A pen test with no remediation budget is an expensive way to produce a list of problems you cannot afford to fix.
The cloud inventory gap is also widening. Many SMEs now run a mix of Microsoft 365, an Azure subscription or two, a few SaaS platforms with their own admin portals and a website hosted somewhere else again. Each of those is a potential attack surface, and each has its own rules about who may test it. If you recently migrated workloads — for example as described in our guide to moving Windows Server 2016 workloads to Azure — make sure the new resources, and any old on-premises IPs you have retired, are reflected in scope.
Preparing for a pen test: a realistic timeline
For a typical SME engagement — an external infrastructure test, an internal network test and one web application — the preparation window runs to about six weeks from the decision to test to the start of remediation. The timeline below shows how the work usually falls. Compress it if you must, but do not drop stages.
If your deadline is tighter than six weeks, the stages you can safely compress are supplier shortlisting (use a provider you already trust) and the gap between contract and testing. The stage you should never squeeze is third-party authorisation — testing a hosted asset without the host’s permission can breach their terms and, in some circumstances, the Computer Misuse Act 1990.
Pen test readiness gauge: a decision framework
Once you have worked through the scoping and access items, it helps to score your overall readiness before committing to dates. We use a simple 100-point framework across five areas, each worth 20 points: objective and scope, authorisation, access and environment, communications, and remediation readiness. The gauge below shows the median score we see across UK SMEs at the point they first ask for a quote.
How to read your own score:
- 80–100: book the test. Remaining gaps can be closed during the lead time.
- 60–79: book provisionally, but make authorisation and access the priority for the next fortnight. Most organisations land here once they have done the inventory.
- 40–59: fix scope and authorisation first. Testing now is likely to waste days on logistics.
- Below 40: consider a vulnerability assessment or a Cyber Essentials gap review first. A full penetration test is unlikely to be the best use of budget until the basics are in place.
The point of the gauge is not precision; it is to force a conversation about the weakest of the five areas before money is committed. A business scoring 18/20 on scope but 4/20 on remediation readiness is about to commission a report it will struggle to act on.
Penetration test cost breakdown by test type
Pen test scoping and cost are tightly linked: almost every pound in a quote maps to a tester-day, and every tester-day maps to an item in scope. The table below shows typical 2026 UK price ranges for common SME test types. Use it to sense-check quotes and to decide which tests are worth the budget this year.
| Test type | Typical SME days | Indicative cost (ex VAT) | Key preparation item |
|---|---|---|---|
| External infrastructure | 2–4 | £1,800–£5,000 | Accurate public IP and domain list; hosting authorisation |
| Internal network / assumed breach | 3–6 | £3,000–£8,000 | VPN or on-site access; standard user account; network diagram |
| Web application (authenticated) | 3–8 | £3,000–£10,000 | Test accounts per role; stable staging; API documentation |
| Microsoft 365 / Entra ID configuration review | 2–4 | £2,000–£5,000 | Read-only Global Reader account; Conditional Access exclusions agreed |
| Wireless assessment | 1–3 | £1,200–£4,000 | Site access; SSID list; guest and corporate network details |
Most suppliers add a day for reporting and project management, and on-site work may add travel. Retests of specific findings are usually charged at half a day to a day. Beware quotes that are dramatically cheaper than this range: they frequently describe an automated scan with a branded report rather than a manual test.
The cheapest way to reduce cost is not haggling on day rate — it is tightening scope. A flat list of 40 IP addresses, half of which you no longer use, costs more to test than a curated list of the 15 that matter. Likewise, giving a web application tester working credentials on day one can save a full day of effort compared with making them fight through registration flows.
How UK penetration tests usually end: findings by severity
One of the best ways to prepare for a pen test is to prepare for its results. The donut below shows the share of UK SME tests in our recent sample that produced at least one finding rated high or critical. In other words: expect to find something serious, and plan your remediation capacity accordingly.
The usual suspects are depressingly consistent: missing patches on internet-facing services, legacy authentication protocols still enabled, local administrator passwords reused across machines, excessive permissions on file shares, SMB signing not enforced, and web applications that trust client-side input. None of these is exotic. Most map directly onto Cyber Essentials controls — secure configuration, user access control, malware protection, security update management and firewalls — which is why passing Cyber Essentials first often makes a pen test both cheaper and more interesting.
Plan now for who will fix what. A critical finding on a line-of-business application usually needs the software vendor; a Microsoft 365 misconfiguration needs your IT provider; a weak password policy needs HR and leadership buy-in. If you have recently worked through identity hygiene — for example the steps in our guide to Microsoft 365 offboarding and leaver accounts — you will already have removed some of the stale accounts testers love to find.
The penetration test checklist — 12 pre-test essentials
This is the core checklist. Work through it in order; each item depends on the ones above it. Tick every item before the kick-off call, and keep the evidence (emails, signed forms, account lists) in one folder you can share with the tester.
- Write the test objective. One paragraph describing the attacker you are simulating, the assets you care about and the decision the report will support. Attach any external requirement (insurer wording, client security schedule, PCI DSS requirement 11.4).
- Build the asset inventory. List every in-scope domain, subdomain, public IP range, cloud subscription, SaaS tenant, application URL, office site and Wi-Fi network. Mark each one as in scope, out of scope or “ask first”.
- Confirm ownership of every asset. Check WHOIS, cloud billing and hosting contracts. IP ranges you think are yours sometimes belong to a former provider; testing them is testing someone else’s network.
- Choose an accredited supplier. Look for CREST membership, testers holding CREST CRT/CCT or equivalent qualifications, and NCSC CHECK status if you are a public sector body or supply one. Ask for a sample report and professional indemnity insurance details.
- Agree the rules of engagement. Testing windows, permitted techniques (is social engineering in? denial-of-service checks out?), data handling, how to treat personal data encountered, and the stop-test procedure. Get it signed by someone with authority.
- Obtain third-party authorisation. Read the test policies of AWS, Azure, Google Cloud and your web host; most major clouds no longer require pre-approval for standard tests on your own resources, but many smaller hosts and managed service providers still do. Get written permission where needed.
- Prepare accounts and access. Create named test accounts for each role (standard user, manager, administrator where in scope), sort out MFA enrolment, and confirm they work through Conditional Access. Issue VPN, VDI or on-site access and test it before day one.
- Decide on allow-listing. Agree whether the tester’s source IPs should be allow-listed in your WAF, EDR and firewall, or deliberately left blocked so you can see how defences respond. Document the choice; do not leave it to chance.
- Tell the people who need to know. Brief your MSP, SOC or MDR provider, hosting partner and a small internal group. Share the test window, source IPs and an emergency contact. Do not tell everyone — it can distort social-engineering or detection results.
- Protect production. Confirm recent, tested backups of in-scope systems, agree a change freeze for the test window, and identify fragile systems (old ERP, OT, medical or building management devices) that need gentle handling or exclusion.
- Set up the communication plan. Kick-off call, daily check-in, immediate phone escalation for critical findings, and a shared secure channel for evidence. Agree who receives the report and how it will be transmitted and stored.
- Reserve remediation capacity. Book time with your IT provider and developers for the fortnight after the report lands, set aside budget, and pencil in a retest. Make sure the previous test’s open findings are listed so the tester can confirm whether they are fixed.
UK law matters here. Under the Computer Misuse Act 1990, accessing a computer system without authorisation is a criminal offence — and “authorisation” must come from someone entitled to give it. Your signature covers systems you own and control; it does not cover your hosting provider’s shared infrastructure or a SaaS supplier’s platform. Keep the signed authorisation alongside the scope document for the life of the engagement.
Mini case study: a 45-person professional services firm in Leeds
A Leeds-based professional services firm of around 45 staff came to us after a large client added a “penetration test within the last 12 months” clause to its supplier assurance questionnaire. The firm had booked a test once before, two years earlier, and described it as “stressful and not very useful”: the tester spent the first two days waiting for VPN access, the external IP list included a range the firm had left three years before, and nobody was assigned to fix the eleven findings, four of which were still open.
This time, preparation started six weeks out. The asset inventory alone removed two stale IP ranges and added three forgotten subdomains, including an old client portal still running on an unsupported framework. The firm’s web host required written authorisation, which took eight working days to obtain — it would have pushed the test back a week had it been left until the end. Test accounts were created for a standard user and a partner-level user, with MFA enrolled and checked against Conditional Access policies before kick-off. The managed SOC was briefed with the tester’s source IPs, and a deliberate decision was taken not to allow-list them, so the firm could see whether its EDR raised alerts.
The test ran on schedule over five days. It found two high-severity issues — the forgotten client portal, and a service account with a weak password and domain administrator rights — plus a handful of medium and low findings. Because remediation time had been pre-booked, both high findings were fixed within ten working days and verified in a half-day retest. The SOC also gained a useful data point: it detected the lateral movement within 40 minutes, but had missed the initial password spraying.
“The test itself was almost the same as last time. What changed was everything around it. We knew what we were testing, the tester could start on day one, and when the report arrived there were names next to every finding. That is what the client actually wanted to see.” — Operations Director, professional services firm, Leeds
The broader lesson: the firm did not spend more on testing. It spent roughly the same number of tester-days, but redirected the two days previously lost to logistics into actual testing, and it treated the report as the start of a remediation project rather than the end of a procurement exercise.
Common pen test preparation mistakes to avoid
Most failed or disappointing penetration tests trace back to a small number of preparation errors. Recognising them in advance is cheaper than discovering them on day one.
- Hardening for the test. Rushing out patches and configuration changes the week before hides your real security posture and often introduces instability. Fix known issues on your normal schedule; let the test show you what you missed.
- Scoping by guesswork. Handing over “our IP range” from an old firewall config, without checking ownership, leads to testing assets you do not control and missing ones you do.
- Forgetting third parties. Hosted websites, managed firewalls, colocation racks and SaaS platforms each have their own rules. Testing without the provider’s agreement can breach contracts and, in some cases, the law.
- Not telling the SOC or MSP. An unannounced test that triggers incident response wastes everyone’s time, can lead to the tester’s IPs being blocked mid-test, and occasionally results in an unnecessary incident report.
- Broken test accounts. Accounts that cannot pass MFA, are blocked by Conditional Access or lack the right application roles routinely cost testers a day or more.
- Testing a staging environment that does not match production. Findings from an out-of-date staging build can be irrelevant — or worse, give false reassurance about production.
- Choosing on price alone. A very low quote often means an automated scan with a manual-test label. Ask about methodology, tester qualifications and how much of the time is spent on manual exploitation.
- No owner for the report. If nobody is responsible for turning findings into tickets, the report becomes evidence of known risk rather than reduced risk — which is not a comfortable position with an insurer or the ICO after an incident.
A penetration test report is a sensitive document. It describes exactly how to break into your organisation. Agree in advance how it will be transmitted (encrypted, never as a plain email attachment), who may read it, and how long the testing firm will retain your data. Treat any request to share it with a client as a decision, not a default — a summary letter of assurance is often enough.
Penetration test preparation at a glance
A one-table summary of the key facts and decisions covered in this guide, for sharing with your leadership team or pasting into a project plan.
| Item | What good looks like |
|---|---|
| Test objective | One paragraph naming the attacker, the assets and the decision the report informs |
| Lead time | Start preparation about six weeks before you need the report |
| Supplier credentials | CREST membership; CHECK for public sector work; qualified named testers |
| Test model | Grey-box for most SME internal, web and cloud tests; black-box for perimeter realism |
| Asset inventory | Domains, subdomains, IPs, cloud tenants, apps, sites and Wi-Fi — ownership verified |
| Legal authorisation | Signed by your organisation and every third party hosting in-scope assets |
| Rules of engagement | Windows, techniques, data handling, stop-test procedure, escalation contacts |
| Test accounts | One per role, MFA enrolled, checked against Conditional Access before day one |
| Allow-listing | A documented, deliberate decision for WAF, EDR and firewall |
| Notifications | MSP, SOC and host briefed; small internal circle only |
| Production safety | Tested backups, change freeze, fragile systems identified |
| Typical SME budget | £5,000–£20,000 for a combined external, internal and web test |
| Reporting | Critical findings escalated immediately; formal report 5–10 working days after testing |
| Remediation | Owner and target date per finding; capacity and budget reserved in advance |
| Retest | Targeted retest of critical and high findings, typically within 30–60 days |
How Cloudswitched helps with penetration testing
Cloudswitched works with UK organisations on the whole penetration testing process: defining objectives, building the asset inventory, coordinating accredited testers, handling third-party authorisation and test accounts, and — most importantly — turning the report into prioritised remediation work across Microsoft 365, Azure, networks and applications. Because we also manage IT estates day to day, we can align test windows with change control, brief monitoring teams and pick up the fixes without a second handover.
Planning a penetration test?
Talk to us about scoping, preparation and remediation support, so your next test produces findings you can act on.
Talk to us about Penetration TestingFrequently Asked Questions
How far in advance should I start preparing for a pen test?
Around six weeks is a comfortable window for a typical SME engagement covering external infrastructure, an internal network and one web application. That allows a week to define the objective, a week for the asset inventory and supplier shortlist, a week for scoping calls and quotes, then two to three weeks for contracts, third-party authorisation and access set-up. If you have a hard deadline from an insurer or client, remember the report usually arrives five to ten working days after testing ends, so count back from the date the report is due, not the date testing starts.
What should a penetration test checklist include?
At minimum: a written test objective, a verified asset inventory, an accredited supplier, signed rules of engagement, third-party authorisation for hosted assets, working test accounts and access, a deliberate allow-listing decision, notifications to your MSP and SOC, backups and a change freeze, a communication and escalation plan, and reserved remediation capacity. The 12-point checklist in this guide covers each of those in the order they usually need to happen, and each item leaves evidence you can hand to the tester at kick-off.
What is pen test scoping and why does it affect cost?
Pen test scoping is the process of deciding exactly which systems, applications, networks and techniques are in and out of the test, and how much the tester will know at the start. Because UK penetration testing is priced in tester-days, scope drives cost almost directly: more IP addresses, more application roles or more sites mean more days. Tight, accurate scoping usually saves more money than negotiating on day rate, and it also improves quality because the tester spends time on the assets that actually matter to your business.
Do I need permission from Microsoft, AWS or my web host before testing?
Microsoft Azure, AWS and Google Cloud publish penetration testing rules of engagement that allow customers to test their own resources without pre-approval for most standard techniques, though denial-of-service testing and attacks on the provider’s own infrastructure are prohibited. Smaller hosting companies, managed service providers and SaaS vendors often have stricter rules and may require written permission. Check each provider’s policy during scoping and keep the evidence with your authorisation documents, because your signature alone does not authorise testing of infrastructure you do not own.
Should I patch everything before the penetration test?
Keep patching on your normal schedule, but do not launch a rushed hardening sprint in the days before testing. Last-minute changes can introduce instability during the test window and they hide your real day-to-day security posture, which is the very thing you are paying to measure. If you know about a serious issue, it is reasonable to fix it, but tell the tester so they can confirm the fix. Otherwise, let the test show you what your processes are missing.
What is the difference between a vulnerability assessment and a penetration test?
A vulnerability assessment uses automated tools to identify known weaknesses, such as missing patches or insecure configurations, and lists them by severity. A penetration test goes further: a qualified tester manually attempts to exploit weaknesses and chain them together to show real-world impact, such as reaching sensitive data from a single phished account. Vulnerability assessments are cheaper and useful monthly or quarterly; penetration tests are deeper and usually annual or after significant change. A good vulnerability assessment checklist is a sensible precursor to a pen test.
Should we tell staff that a penetration test is happening?
Tell a small group who need to know: your IT lead, your managed service provider, any SOC or MDR provider, your hosting partner and a senior sponsor. Brief them on the test window, the tester’s source IPs and the emergency contact. Avoid announcing it company-wide, particularly if social engineering or detection testing is in scope, because staff behaviour and alert handling change when people know they are being watched. Agree in advance how genuine incidents during the test window will be distinguished from test activity.
What accreditation should a UK pen testing company have?
For most private-sector organisations, CREST membership is the main benchmark, backed by individual tester qualifications such as CREST Registered Tester (CRT) or CREST Certified Tester (CCT), or equivalent certifications. Public sector bodies and organisations supplying them may need an NCSC CHECK-approved company, with CHECK Team Leaders and Members named on the engagement. Also ask for a sample redacted report, details of professional indemnity insurance, and the named testers who will actually carry out your test rather than relying on the company brand alone.
Can a penetration test break production systems?
Reputable testers work carefully to avoid disruption, but some risk always remains, particularly with older or fragile systems such as legacy ERP platforms, building management controllers or medical devices. That is why the preparation checklist includes confirming recent tested backups, agreeing a change freeze, identifying fragile systems for gentle handling or exclusion, and having a technical contact available throughout the test window. Rules of engagement should also define a stop-test procedure so either side can pause testing immediately if something unexpected happens.
What happens after the pen test report arrives?
Check the draft for factual accuracy first, then hold a debrief with the tester and the people who will carry out fixes. Assign every finding an owner and target date based on severity, track progress weekly, and book a targeted retest of critical and high findings, typically within 30 to 60 days. Store the report securely and decide deliberately what you share with clients or insurers, since a letter of assurance or executive summary is often enough. Feed the lessons into next year’s scope so each test builds on the last.
How often should a UK SME carry out penetration testing?
Annual testing is the common baseline, and many insurers and client contracts specify it. You should also test after significant change: a new customer-facing application, a major cloud migration, a merger, or a substantial network redesign. Between annual tests, regular vulnerability scanning and maintaining Cyber Essentials controls keep the basics in place, so the yearly penetration test can focus on deeper issues rather than rediscovering missing patches. Smaller organisations with very simple estates may start with a combined external and Microsoft 365 test.
Related reading
These guides cover areas that often come up during pen test scoping and remediation:
- Cyber Essentials and Unsupported Software - The UK Guide to End-of-Life Devices in 2026
- Windows Server 2016 End of Support - How UK Businesses Should Move Legacy Servers to Azure Before January 2027
- Microsoft 365 Offboarding: How to Handle Leavers’ Mailboxes, Licences and Data - UK Guide 2026
- WPA3-Enterprise and 802.1X Wi-Fi Authentication - A UK Guide to Meraki and Entra ID
- Static IP Addresses for Business Broadband: A UK Guide to Costs, IPv4 Scarcity and IPv6 Readiness in 2026
Get more from your next penetration test
Cloudswitched helps UK organisations scope, prepare for and act on penetration tests, from the first asset inventory to the final retest.
Explore our Penetration Testing service








