- 01
Market
Saudi Arabia
- 02
Written for
CTO, CIO
- 03
Decision
Work out which security activity the situation actually calls for
- 04
Languages
EN · AR
Cybersecurity credentials
NCA registration
Registered cybersecurity service provider

ISO/IEC 27001
ISO/IEC 27001 certified
Choosing
What each activity answers
The right activity follows from the question being asked of you. Read the left column as the question, not as a product.
| Option | The question it answers | Finds well | Misses | Typical effort | Best for |
|---|---|---|---|---|---|
| Automated scanning | Do we have known vulnerable components and obvious patterns? | Outdated dependencies, missing headers, publicly known vulnerabilities in libraries. | Anything that requires understanding what the application is for. | Configuration, then continuous. Not an engagement. | Running permanently in the pipeline. This is a control you own, not a service you buy. |
| Design review | Is the thing we are about to build capable of being secure? | Trust boundaries in the wrong place, authorisation decided by the caller, data collected without a reason to hold it. | Everything about the implementation, because there is not one yet. | Two to four days, before code is written. | A new service, a change to how authorisation works, or a first exposure of internal data to partners. |
| Secure code review | Is the security logic correct where a request meets a decision? | Authorisation gaps, unsafe defaults, and the same pattern repeated across a codebase. | Anything that only appears once the system is deployed and configured. | Three to five days scoped to a pattern; longer for a whole codebase. | Establishing how far a known defect extends, and clearing a significant change before release. |
| Penetration test | What can an attacker actually do to the running system? | Chained defects, business logic failures, and configuration as it exists in the real environment. | Code paths not reachable in the window, and the pattern behind a single reported instance. | Eight to fifteen days for a mid-sized application, plus retest. | Answering an external party who has asked for assurance, and an annual check against a live system. |
Sequencing
The order that wastes the least money
Scanning goes first because it is cheap, continuous and finds the class of problem that should never reach a paid engagement. Paying a tester day rate to report an outdated library is the most common way to waste a budget in this category, and it is entirely avoidable.
A design review comes next when there is a design to review — a new service, a change to how authorisation works, a first exposure of internal data to partners. It is the cheapest place to find the expensive defects, because at that stage the fix is a decision rather than a rewrite. It is also the activity buyers skip most often, because nothing has been built yet and the risk does not feel present.
The penetration test goes last, against something deployed and configured, and its findings are worth more when the first two have already run — the tester spends the days on logic and chaining rather than on things a scanner would have caught. Code review is the one that moves: before release when the change is significant, and after a test when a finding turns out to be a pattern rather than an instance.
Common situations
What people are usually being asked for
- 01
"Our enterprise customer sent a security questionnaire"
Read what the questionnaire actually asks for before buying anything. It usually wants a recent test report and evidence of a remediation process, not a certification. A test plus a written vulnerability management procedure answers most of these, and the procedure is the part that is missing more often than the test.
- 02
"We are about to launch and nobody has looked at it"
A short design review plus a test against the pre-production build, in that order, and with time between them to fix what the review finds. Running the test first against an application nobody has reviewed spends the test days on defects a review would have surfaced in an afternoon.
- 03
"A test found something and we do not know how far it goes"
This is the clearest case for a code review, scoped narrowly to the pattern rather than the whole codebase. The question is not whether the reported instances are real — they are — but how many more of them exist, and only reading the code answers that.
- 04
"We inherited an application and have no documentation"
Start with a review that produces the documentation as its output — the components, the data it holds, where authorisation is decided and what it integrates with. Testing an application nobody can describe produces a report nobody can act on.
Scope boundary
What any of this establishes
What we do
- Recommend one activity and state why the other three are not the right spend for this situation.
- Say when the honest answer is that no engagement is needed yet — usually when scanning has not been turned on.
- Write findings so the pattern is separable from the instances, which is what makes a fix complete.
Where an assessment is commissioned to satisfy a national framework control, the assessor decides which activity and which evidence they accept. Establishing that first avoids buying a competent engagement that does not answer the control.
Primary sourceCommon questions
What buyers ask
How much of this can we do ourselves?+
Scanning entirely, and you should — dependency scanning and static analysis in the pipeline are engineering practices rather than a service to buy. Code review partly: your own team reads code better than an outsider does, but they share the assumptions that produced the defect, which is why an outside reader finds authorisation gaps a team walks past. Penetration testing is the one where independence carries most of the value, and where an assessor asking for evidence will usually expect it.
Do you need source code access?+
For a code review, obviously. For a penetration test, not necessarily — but access makes the test better and cheaper, because a tester who can read the authorisation logic knows where to look instead of inferring it from responses. Where the code cannot leave your environment we work on a machine you provide, and where it cannot be shown at all we test without it and say in the report what that cost in coverage.
What does an assessment cost?+
It is a day rate against a scope, and the scope is what the conversation is about. A design review of one new service is usually two to four days. A focused code review against a known pattern is three to five. A penetration test of a mid-sized application with several roles is eight to fifteen, plus retest. Anyone quoting before understanding the role count and the integration surface is quoting a template.
In this cluster
Next step
Work out which activity you need
A short scoping conversation that ends with a recommendation for one activity and a stated reason the other three are not it.
Work out which activity you need
What happens next
- 1We reply within one business day.
- 2A 30-minute scoping call with a senior consultant.
- 3A written proposal with scope, team, and timeline.
Or reach us directly
