The short answer
Decide what to test, set up test accounts, pick a testing window and name one person the testers can reach. With those four things ready, a pentest starts faster and finds more.
Decide what to test
Start with what holds or touches your customers' data. For most companies, that is the web app, its API and the cloud behind them. Write down the web addresses, API docs and accounts that are in scope. Not sure? See our six kinds of test.
Set up access
- A test account for each kind of user, like a normal user and an admin
- API docs, an OpenAPI file or a Postman collection
- A read-only role, for a cloud review
- Two test customers, if many customers share your app, so the testers can check that one cannot reach the other
A test copy, or the real thing?
A test copy of your app is safer, as long as it matches the real one closely. Testing the real app shows exactly what an attacker would see, but it needs more care. Either works. Decide on the scope call.
Tell the right people
- Your own team, so nobody worries about odd traffic.
- Your cloud provider, only if their rules ask for it. The big ones let you test what you own without asking first, as long as you follow their rules.
- Any outside service in scope, if their terms say so.
During the test
- Name one person the testers can reach quickly.
- Expect a call the same day if something serious turns up.
- Try not to make big changes to what is being tested until the testing is done.
Plan time to fix
Most teams need a few weeks to fix what is found. If you have an audit or a customer deadline, count back from it and leave room for a retest. See how often to test for timing.