Penetration testing

Penetration testing: find the way in first

For owners and IT leads who've been asked "when was your last pen test?" by a client, an insurer or a tender. An independent specialist, one we bring in and manage or one you engage directly, tries to break in with your written permission and shows you exactly how. You get a report your board can read, our engineers fix what's found, and a retest proves it's closed.

See how three small gaps become one way in
Penetration test report, page 1Example data
Scope
Internet-facing systems, Microsoft 365, office network
Overall risk
High, from one chain of three findings
Findings
1 critical, 2 high, 4 medium, 6 low
Fix first
Multi-factor sign-in on remote access

Retest: critical and high findings closed.

Next: the medium findings, in the fix plan agreed with you.

Example report summary. The business and findings are made up.

How three ordinary gaps become one way in

Scanners list problems one at a time. A tester does what an attacker does: joins them up. Here is one illustrative path, step by step, with what each finding looks like alone and what it means once it's chained.

Illustrative scenario, based on situations we see often. Not a single client.

  1. Finds a staff password in old breach data

    What the tester did

    Searches public breach lists for your staff email addresses and finds one with a password from a site breached years ago.

    Rated on its ownMedium
    What it means for you

    Someone reused a work password on another site. On its own, it's only a risk if it works somewhere.

  2. Tries it on remote access

    What the tester did

    The password still works on the office remote access portal, which asks for nothing else.

    Rated on its ownHigh
    What it means for you

    Anyone with that password can reach your office network from home, from anywhere.

  3. Looks through the shared drive

    What the tester did

    Inside, the shared drive is open to every staff account. A folder of old setup scripts holds an administrator password in plain text.

    Rated on its ownMedium
    What it means for you

    One staff login can read things only IT should see.

  4. Signs in as the administrator

    What the tester did

    The administrator password still works. The tester now controls every computer and account on the network, and stops there.

    Rated on its ownCritical, as a chain
    What it means for you

    This is the path a ransomware gang would take. Each step alone looked fixable later. Together, they're the way in.

What gets fixed, and in what order

  • NowMulti-factor sign-in on remote access. The chain breaks at step two.
  • NowChange the administrator password and remove it from the scripts folder.
  • NextTighten the shared drive so staff see only what their job needs.
  • NextReset the reused password and turn on alerts for staff logins found in breaches.

Then the retest

Retest: the tester tries the same path again. It stops at step two. That's the evidence your client or insurer asks for.

A penetration test shows how your gaps join up, so you fix the one that breaks the chain first.

Work out what you should test

What should you actually test?

Three questions. You get a suggested scope in priority order, written so you can take it to any tester, including the specialist we'd bring in.

What do you run? Pick all that apply.
Why are you testing now?
Is multi-factor sign-in on every account, and is patching on a schedule?

Your suggested scope

3 tests, in this order

  1. External test. Everything an attacker can reach from the internet, and whether any of it lets them in.
  2. Cloud configuration review. Sign-in rules, sharing settings and admin access in Microsoft 365 or Google Workspace.
  3. Phishing simulation. Who would hand over a password, and how fast someone reports it.
  • Start with the external test and the cloud review. They cover the most common ways in.

Your answers stay in your browser. Nothing is saved or sent.

With the defaults (internet-facing systems and Microsoft 365, testing to see where you stand, basics in place), the suggested scope is an external test, a cloud configuration review and a phishing simulation, in that order.

Compare a pen test with a scan or a gap check

Pen test, scan, gap check or audit: which answers your question?

They're often sold as the same thing. They aren't. Pick by the question you need answered.

Four ways to check your security, compared
ComparedVulnerability scanEssential Eight gap checkPenetration testSecurity audit
The question it answersWhich known weaknesses are showing?How far are we from ASD's baseline?Could someone actually get in, and how far?Are our policies, controls and records in order?
Who does itSoftware, checked by our engineersOur engineers, with youAn independent specialist testerOur engineers, or an independent auditor
How oftenContinuously, on a scheduleOnce, then reviewedAt least yearly, and after big changesYearly, or for a certification
What you getA list of issues to patchA gap list against each controlA report of real attack paths, ranked, and a retestFindings against a standard
How it's paid forIncluded on every planQuoted before it startsQuoted for your scopeQuoted before it starts

Vulnerability scanning is on every Real Bytes plan. Penetration testing is quoted, because the work depends on scope. See our Pricing page and our security audit page.

See what the report looks like

What you get: a report written for two audiences

Page one is for the owner and the board. The rest is for whoever fixes things. Here's the shape of it.

Example data. Names, addresses and findings are made up.

Penetration test report, example

Findings, ranked

  • CriticalChainReused staff password, remote access without multi-factor sign-in and an exposed administrator password give full control of the network.
  • HighRemote accessRemote access portal accepts a password alone.
  • HighMicrosoft 365Two staff accounts excluded from the sign-in rules that block risky locations.
  • MediumShared driveSetup scripts readable by every staff account.
Example findings page. The business and findings are made up.

What's in the report

  1. Summary. Scope, overall risk, and the few findings that matter most, in plain words.
  2. Scope and rules. What was tested and what wasn't, the testing window and the rules agreed.
  3. Findings. Each one rated by severity, with evidence and the fix.
  4. Attack paths. How findings chain together into a real way in.
  5. Fix plan. Now, next and later, so you can plan the work and the spend.
  6. Retest. What was checked again and the result, added to the same report.
See how a test runs, start to finish

How a test runs, and who does what

An independent specialist does the testing, either one we bring in and manage, or one you engage directly. Either way we help write the scope and our engineers fix what's found.

Why the testers aren't our engineers: the people who look after your systems shouldn't mark their own work. PCI DSS, for one, asks for organisational independence of the tester.1

  1. Before

    Scope call

    What you run, what worries you and what your client or insurer needs. We write the scope with you.

    What we need from you: Thirty minutes, and the questionnaire if someone sent you one.

  2. Before

    Written authorisation

    Testing without the owner's written permission is unauthorised access. Microsoft, for example, requires explicit written authorisation before a third party tests your cloud resources.2

    What we need from you: A signature from someone who can authorise it.

  3. Testing window

    Specialists test

    On dates and times agreed with you. Denial of service testing is off the table: Microsoft's rules prohibit it on its cloud.2

    What we need from you: A contact who answers the phone if something looks wrong.

  4. After

    Report and walk-through

    We go through the report with you in plain words: what matters, what can wait, what it costs to fix.

    What we need from you: An hour with whoever makes the call on fixes.

  5. After

    Our engineers fix it

    Fixes in the order agreed, usually starting with whatever breaks the worst chain.

    What we need from you: A yes on the fix plan.

  6. After

    Retest

    The testers check the fixes and the result goes into the same report. PCI DSS expects testing to be repeated to verify corrections.1

    What we need from you: Nothing. You get the updated report.

Questions to ask any tester

  • Who exactly will test, and are they independent of who runs our IT?

    They should be. The testing is done by an independent specialist, either one we bring in and manage or one you engage directly. Our engineers help write the scope and fix what's found, but never test their own work.

  • What accreditation do the testers hold?

    Ask for it in writing. CREST, for example, independently assesses accredited companies against recognised standards.3 When we bring the tester in, we give you their details in the scope document.

  • Is the retest included?

    When we bring the tester in, we scope the retest with the test, so the price you see covers proving the fixes worked. If you engage a tester directly, ask for the retest to be in their quote.

  • What happens if you find something serious mid-test?

    You're told straight away, not at the end. Serious findings can be fixed during the testing window if you choose.

See it from your desk

Why people book a test

The reason shapes the scope. Three common ones.

The tender asks for a penetration test in the last 12 months. We've never had one.

Illustrative scenario, based on situations we see often. Not a single client.

A bigger client sends a supplier security questionnaire. One question asks for the date and summary of your last penetration test. The tender closes in six weeks.

What you notice: one tidy answer instead of a scramble, and a report you can reuse for the next tender.

What we do

  • Read the questionnaire with you and scope exactly what they ask for, no more.
  • Book the testing window so the report arrives before the deadline.
  • Give you a summary page you can attach, and help answer the rest of the questionnaire.
Pick your scope

Small businesses reported an average cost of $55,400 per cybercrime report in 2025-26.4

Check us before you trust us with a test

Ask any provider for proof. Here's ours, with links so you can check it yourself.

SMB1001 Diamond
Independently audited by BSI. Check our certificate on CyberCert Verify.6
ASD Network Partner
A partner in ASD's Cyber Security Partnership Program.
4.8 from 40+
Google reviews. Read what clients say.
Since 2013
In business since 2013. Real Bytes as it is today formed in 2017.

Our stack: Guardz with SentinelOne, Huntress, Check Point, Microsoft, Ubiquiti UniFi, ConnectWise, Pax8. We also support, when a client already has them: CrowdStrike, Sophos, Fortinet, Meraki, Veeam.

The engineers who fix what the test finds

  • Philip Verner, Director at Real BytesPhilip VernerDirector
  • Blake Bath, General Manager at Real BytesBlake BathGeneral Manager
  • Mason Wise, MSP Practice Lead at Real BytesMason WiseMSP Practice Lead
  • Mitchell Morgan, Automation & Platform Lead at Real BytesMitchell MorganAutomation & Platform Lead
  • Travis Jacob, Operations Manager at Real BytesTravis JacobOperations Manager
  • Bill Nobbs-O'Brien, Service Desk & Field Engineer at Real BytesBill Nobbs-O'BrienService Desk & Field Engineer

Photos: Real Bytes. Meet our engineers

When a penetration test isn't the next step

A test is worth the money when it can tell you something new. Sometimes something else should come first.

What a test can't do

  • Protect you on its own. It's a snapshot of one window in time, not monitoring.
  • Fix anything. It finds and explains. The fixing is separate work.
  • Replace the basics. Missing multi-factor sign-in or patching will fill the report.
  • Certify you. It's evidence for a certification, not the certification itself.

What we need from you

  • A clear reason for the test, so the scope fits it.
  • Written authorisation from someone who can give it.
  • A contact who answers during the testing window.
  • A decision on the fix plan once you've seen the report.
Read the questions people ask

What people ask about penetration testing

What is a penetration test?

An authorised, controlled attempt to break into your systems the way a real attacker would. It shows which weaknesses can actually be used and how far an attacker could get, not just a list of possible problems.

Will testing disrupt our business?

It shouldn't. Scope, dates and times are agreed before anything starts, and you have a contact to call. Denial of service testing is excluded; Microsoft prohibits it on its cloud in any case.2

How often should we test?

At least once a year, and after a big change such as a new system going live or a move to the cloud. PCI DSS sets that pace for card environments: at least once every 12 months and after any significant change.1

Who actually does the testing?

An independent specialist. We can bring one in and manage the engagement, or work with a tester you engage directly. Either way we help write the scope and our engineers fix what's found.

Can Microsoft 365 and the cloud be tested?

Yes, within each provider's rules. Microsoft doesn't need to approve testing of your own Azure resources, but testers must follow its rules of engagement and have your explicit written authorisation.2 For Microsoft 365, a configuration review checks sign-in rules, sharing and admin access.

Are we too small for a penetration test?

Maybe. If multi-factor sign-in and patching aren't done yet, start there: an Essential Eight gap check costs less and tells you more. Once the basics are in place, an external test and a cloud review are a sensible first test for a small business.

Do we need one for SMB1001 or insurance?

SMB1001 adds penetration testing at Diamond, its top level. Insurers and client questionnaires vary: some ask for a recent external test, others for evidence of controls. Send us the questionnaire and we'll tell you what it actually asks for.

How much does it cost?

Penetration testing is quoted, because the work depends on scope: how many systems, apps and people. When we bring the tester in, you see a fixed quote for the scope you choose, with the retest included, before anything is booked. If you engage a tester directly, they quote you for the test and we quote our part, the scoping and the fixes. It isn't part of the monthly plans, which start at A$150 a person a month ex GST. Our Pricing page lists it with the other quoted services.

Want to know if someone could get in?

Tell us what you run and why you're asking. One of our engineers will suggest what's worth testing, what can wait and what to fix first, before you spend anything on a test.

Three months, then month to month on 60 days' notice. No exit fees. Onboarding is quoted before you commit.

Already a client? Log a job in your portal or email helpdesk@realbytes.com.au.

What happens next

  1. You pick a time that suits you.
  2. One of our engineers calls you. It takes 30 minutes and it's free. There's no slide deck.
  3. If we can help, we send a written plan. If we can't, we'll tell you who can.
  4. Nothing happens until you say yes.

Sources

Figures and sources checked 22 November 2026. The small numbers on this page link here.

  1. Microsoft Learn, Microsoft Entra ID and PCI DSS Requirement 11. Requirements 11.4.2 to 11.4.4, read 22 November 2026
  2. Microsoft Learn, Penetration testing (Azure). Summary of the Microsoft Cloud Unified Penetration Testing Rules of Engagement, read 22 November 2026
  3. CREST International, Cyber security procurement guide for Australia and New Zealand. CREST, page updated 1 July 2026
  4. Australian Government, business.gov.au: Annual Cyber Threat Report 2025-26. Summarising ASD's Annual Cyber Threat Report 2025-26, 28 October 2026
  5. COSBOA Cyber Wardens Research Report 2024. Survey of 2,098 small business owners and employees, March 2024, page 15
  6. CyberCert Verify. Search Real Bytes to check our SMB1001 certificate