Commissioning a security assessment for the first time raises practical questions. What should be tested? Who needs to sign what? Do testers need accounts? Should the IT team be warned? Will anything break?
This guide walks through the preparation in the order it usually happens. None of it is complicated, but each step matters. Time spent here means testers spend their time testing rather than waiting for access, and it means the report covers the systems you actually care about.
Decide why you are testing
Start with the reason. A first assessment is often prompted by something specific: a product launch, a customer security questionnaire, an investor's due diligence, an insurer's request or simply a sense that nobody has checked for a while. The reason shapes the scope.
If a customer has asked for evidence of testing, find out exactly what they expect, such as which systems, how recently tested and what they want to see. If you are launching a product, the application and its API are the obvious focus. If you are unsure what you expose at all, an external review of your internet-facing systems is a sensible first step.
Write the reason down in a sentence or two and share it with the supplier. It helps them propose the right kind of testing and it helps you judge whether the final report answers your question.
Define the scope
The scope is the written list of what will be tested and what will not. It is the most important document in the engagement, because testers must stay inside it. Anything left out will not be tested, and anything included must be something you can authorise.
- Applications: names, URLs and the environment to be tested, such as staging or production.
- APIs: the base addresses, versions and, ideally, documentation such as an OpenAPI specification.
- Networks: the IP address ranges and domains you own or control.
- Cloud: the accounts, subscriptions or projects included.
- User roles: which types of user the testers should act as, for example customer, staff member and administrator.
- Exclusions: anything explicitly off limits, such as a payment provider's hosted page or a fragile legacy system.
Choose the environment and prepare access
Where possible, test a staging environment that closely mirrors production. It lets testers work more thoroughly without any risk to customers or live data. The closer it is to production, in code version, configuration and features, the more meaningful the results will be. If staging is very different from production, say so, because it affects how far the results apply.
If testing must take place on production, agree limits: times of day, avoidance of destructive techniques, and a contact who can be reached throughout.
- Create dedicated test accounts for each role in scope. Two accounts per role is ideal, so testers can check whether one user can reach another user's data.
- Fill test accounts with realistic but fictional data, never real customer records.
- For cloud reviews, create a read-only role that you control and can remove afterwards.
- If your systems restrict access by IP address, arrange to allow the testers' addresses for the testing period.
- Share any documentation that helps, such as architecture diagrams, API specifications or a list of user roles and what each should be able to do.
Decide who to tell
Testing generates activity that can look like an attack. If your monitoring team, managed service provider or hosting provider sees it without warning, they may block the testers or start an incident response. Tell the people who need to know, and give them the testing windows and the testers' source addresses.
Some organisations deliberately keep testing quiet to see whether their monitoring notices it. That is a valid choice for a later, more mature engagement, but for a first assessment it is usually better to inform the relevant teams so the testing can proceed smoothly.
Name one contact on your side who can be reached during testing. If testers find something critical, or if anything appears to be affecting service, they need someone who can make a decision quickly.
Plan for the report before it arrives
A report is only useful if someone acts on it. Before testing begins, decide who will receive the report, who will own the fixes and how much developer time you can set aside for them. Agree with the supplier how the report will be shared securely and when the walkthrough call will be.
Check that a retest is included, so you can confirm that fixes have worked. If a customer or auditor will want evidence, ask whether the supplier can provide a short summary or letter confirming that testing took place and which issues have been resolved.
Finally, ask how engagement data will be handled: where findings and evidence are stored, who can see them, and when they will be deleted.
The short version
Preparing for a first security assessment comes down to six things: a clear reason, a written scope, signed authorisation from the right owner, a suitable environment with test accounts, the right people informed, and a plan for acting on the report. Get these in place and the assessment will spend its time where it should, finding issues in the systems that matter to you.
