Comparison

Penetration test vs vulnerability scan: which one does a small business need?

Two different tools, often confused. A penetration test is a human-led, deep, point-in-time engagement. A vulnerability scan is automated, broad, and recurring. Most small businesses need the second most of the time, and the first occasionally. Here is the honest version.

The two terms get used interchangeably in sales conversations, and that is part of the problem. A penetration test and a vulnerability scan answer different questions, at different prices, on different cadences. If you are running a small business in Australia, knowing which one you need at any given moment saves money and produces better evidence at audit.

A penetration test is a human-led engagement. A consultant (or a small team) spends days or weeks examining a defined scope, chaining low-severity findings into a real exploit path, testing business logic, and writing a bespoke report. Engagements are scoped: a single web application, a network segment, a cloud account, or sometimes a physical or social-engineering exercise. Good firms publish their methodology (OWASP WSTG, PTES, OSSTMM) and tell you what is in and out of scope before they start.

The output is a written report with proof-of-exploit screenshots, severity ratings, and remediation advice. The best pen tests find real business-logic bugs that no scanner would ever catch: an authentication bypass, an authorisation flaw, a chained vulnerability that converts a low-severity issue into a high-impact one. The value is the human judgement.

The honest tradeoffs: pen tests are expensive, slow to schedule, and capture a single point in time. The web application you tested in March is not the web application you have in September. The findings go stale, the team changes, the third-party scripts update, and the scope drifts. That is not a flaw in pen testing — it is a property of the tool. You use it when the question is “is this thing fundamentally well built” or “can a determined human get in”, not for continuous assurance.

A vulnerability scan is automated. A scanner sends requests against a target, compares the responses against a database of known issues, and produces a list of findings. Some scanners are signature-based (a list of CVEs); some include configuration checks (TLS, DNS, email authentication, HTTP headers); some do light dynamic analysis on web applications. The scan runs in minutes or hours, not days.

The output is a list of known issues with severity scores and, usually, generic remediation guidance. A scan is excellent at catching the broad category of known, publicly documented problems: expired certificates, missing security headers, an outdated WordPress plugin, a weak TLS configuration, an exposed admin path. It is poor at catching novel issues, business-logic flaws, or anything that requires a human to understand context.

The right way to think about a vulnerability scan is the same way you think about a smoke alarm. It does not detect every kind of fire, it will not put one out, and it occasionally false-alarms. It catches the broad-based, common stuff loudly enough that you can act on it, and it does so cheaply enough to run on a recurring schedule.

AttackEdge is an external vulnerability scanner with three constraints baked in: it is passive (no exploitation, no test orders, no payload injection), it scans only the public surface of your business (the website, DNS, email authentication, TLS, subdomains, public IPs), and it produces a plain-English report each cycle written for the IT person or web developer who actually fixes things. We sit in the vulnerability-scan column, not the pen-test column. We say that on every page that mentions a pen test, because it matters.

The role we play for an Australian SMB is the recurring external check, run monthly, with dated evidence ready for an insurer or a client question. We are not a substitute for a pen test when one is genuinely required. We are the layer you keep running between pen tests, or instead of them when a pen test is the wrong tool for the budget and the question.

Some genuine signals that a pen test is the right next step:

  • Large-contract procurement. An enterprise customer, government tender, or large partner is asking for a recent pen-test report as part of vendor onboarding.
  • Regulated environment. You operate in a sector where the regulator or your auditor specifically requires a manual pen test on a fixed cadence (some health, financial-services, and critical-infrastructure contexts).
  • Post-incident validation. Something went wrong, you remediated it, and you want a human to confirm the fix actually closed the hole.
  • New public app launch. You shipped a new web application that handles payments, sensitive data, or a novel workflow. A pen test before launch catches the business-logic bugs no scanner would.

Outside those situations, recurring external vulnerability scanning produces more value for a small business. It catches the common issues that account for the overwhelming majority of real-world breaches at SMB scale (weak email auth, expired TLS, exposed admin paths, abandoned subdomains, unpatched plugins) and it produces fresh dated evidence on every scheduled run.

In Australia, a competent external pen test on a single web application costs roughly A$19,000 to A$15,000. Broader scopes (a few applications, the network, social engineering) run A$15,000 to A$40,000 and up. Those are the prices for a real engagement with a published methodology and a written report — not a junior consultant running a free scanner and turning the output into a PDF.

AttackEdge sits at A$39 per month (Solo, one website and email) or A$99 per month (SMB, up to 15 assets), with a A$149 one-off Snapshot for a single baseline. The two are not in competition: pen tests answer “is this thing well built”, AttackEdge answers “is the public surface still healthy this month”.

Side by side

The honest comparison.

Three columns, seven rows. Pick the right tool for the question you are trying to answer.

AspectManual pen testAttackEdgeExternal vulnerability scanDIY scanner
Typical cost (AU)A$19,000 to A$40,000 per engagementA$39 to A$99 per month, or A$149 one-offFree to A$200 per month for an open-source or self-hosted scanner
CadenceOnce or twice a year, scoped engagementUp to you — scheduled or manual, weekly or monthly, on Solo and SMBOn-demand, only when someone remembers to run it
DepthDeep: chained exploits, business logic, authenticated testingBroad and shallow: known issues across the external surfaceDepends on the operator. Usually unauthenticated, signature-based
Who runs itA human consultant or a small team, billed by the dayAutomated scanner, reviewed by ArmoniaLabs, run from SydneyYou, your IT person, or a contractor on a shared laptop
What you getA bespoke report, often with proof-of-exploit screenshotsA monthly PDF, the same shape every cycle, plain-English summaryRaw scanner output. Usually JSON or HTML, no executive summary
Evidence for insurance or auditStrong for a single point in time. Goes stale quicklyDated evidence of recurring monitoring, fresh on every scheduled runHard to defend. No methodology, no dated evidence chain
When to chooseLarge procurement, regulated environment, post-incident validation, new public app launchOngoing assurance on the external surface between (or instead of) pen-test cyclesYou have in-house security capability and you want raw output

If a regulator, insurer, or large customer is asking specifically for a pen-test report, buy a pen test. Get recommendations, read the methodology, scope the engagement honestly, and budget for it as a once- or twice-a-year cost.

For everything else — the regular question of whether your public surface still looks reasonable this month, whether your email auth is still enforcing, whether a forgotten subdomain has appeared, whether your TLS is still valid — a recurring external vulnerability scan is the right shape and the right price. AttackEdge is one such service. There are others. Pick one and run it monthly.

See what attackers see

Ready to see what AttackEdge actually delivers?

The free check covers the headline issues on your domain in about a minute. A paid plan adds the full methodology, the evidence log, and recurring scans you can show an insurer or a customer.

Hosted in Sydney · Passive scanning only · From A$39 per month