Cybersecurity / Scoping and Retest
How we count scope, and what a retest commits us to
Two questions decide whether you can compare quotes at all: what counts as one thing being tested, and what happens after the findings land. Here are our answers, in writing, before you talk to us.
This page is about counting, not pricing
We do not publish rates. Effort follows attack surface, and a number quoted before anyone has seen your environment is a guess wearing a price tag.
What we can publish is how we count. When you hold two proposals side by side, the difference is usually not the rate. It is that each vendor counted your estate differently and neither said so. If you know our unit, you can normalize any quote you receive, including ours.
What counts as one target
A target is one thing being tested. Most engagements are quoted as a number of targets plus the factors below.
One web application
A single application, including the roles and tenants you ask us to test within it. A second application is a second target, even when it shares a codebase or a login.
One mobile app, per platform
The iOS and Android builds count separately. They ship different binaries, different platform controls, and usually different defects, so they are tested as two targets.
One API set
The endpoints behind one product or service, up to an endpoint count confirmed at scoping. Undocumented endpoints your consumers use are still in scope if you want them tested, so tell us they exist.
One network range
One external or internal range, up to a live-host count confirmed at scoping, from the perspective you need tested. Outside and inside are different positions and are counted separately.
Endpoint and host thresholds are confirmed on the scoping call, because an API with forty endpoints and one with four hundred are not the same job. What this answers up front is the question buyers actually ask: three applications count as three targets, and an iOS and Android pair counts as two.
What drives effort beyond the count
Two applications of similar size can differ by a week of testing. These are the factors that move it, and they are the same ones the scoping checklist asks you to prepare.
- Roles and tenant boundaries
- How many user roles you want exercised, and whether cross-tenant isolation is in scope. A multi-tenant product tested across roles takes longer than a single-role application of the same size.
- The position testing starts from
- Unauthenticated, authenticated, or both. Authenticated testing needs working accounts for each role, and it usually finds more, because most business logic sits behind a login.
- Environments and test data
- Production or a staging replica, and how test data is handled. A replica that drifts from production changes what the results are worth, so we agree this before testing starts.
- Integrations and data flows
- Third-party services, internal systems, and the flows between them. Where an integration carries authorization decisions, it becomes part of the surface.
- Constraints and exclusions
- Change freezes, testing windows, and the systems a tester must not touch. Constraints do not reduce rigor, but they do shape sequencing and effort.
- The evidence you need out
- A framework-driven test carries reporting obligations a general assessment does not. Tell us which audit or questionnaire is behind the request and the reporting is built for it.
What a retest entitles you to
Every vendor says retesting is included. Fewer say how many, by when, or what happens if the fix does not hold. Ours is stated here and committed in the engagement agreement.
- Included
- One retest cycle per engagement.
- Window
- Claimable within 60 days of report delivery, so remediation and verification sit inside the same audit or release cycle.
- Covers
- The findings reported in that engagement, verified against the fixes you applied.
- Result
- Findings that pass are updated to remediated and retested, the wording auditors expect, and the report you hand on reflects closure rather than open risk.
- If a fix fails
- The finding is re-reported with what we saw, and a second cycle is quoted. We do not mark something closed because a change was attempted.
What happens after the retest
A retest closes a finding. It does not stop the finding coming back in a later release, which is the part a retest entitlement on its own never covers.
That is what the assurance loop is for. A proven finding becomes an automated security regression test in your suite, and its indicators become a detection rule mapped to MITRE ATT&CK, so the same weakness is caught by your pipeline and your monitoring rather than by the next assessment.
See what each stage of the loop produces →
The phases an engagement runs through are on the methodology page, and what the resulting report contains is walked through in inside a penetration-testing report.
Ready to scope the work?
A 30-minute call with the engineers who will do the testing, not a sales gate.