Technical Anshu
Tutorials Blog
⚙️ Technical Setup

Phishing Simulation Technical Setup With a Training Domain You Own

Good programmes are boring under the bonnet: a domain bought for training and owned forever, a sending setup that cannot hurt the real domain's reputation, and a mail gateway that lets the test through by deliberate, time-boxed exception - not by accident. This page is that setup.

✍️ Technical Anshu⚙️ Domains and Delivery⏱️ Test on IT before anyone else
🎥 Watch the full video - the phishing simulation setup shown step by step

Three mistakes account for most technical embarrassment in this field: sending from the real domain, letting the test get silently eaten by the spam filter, and discovering the landing page mis-records data after launch. The setup below avoids all three, in that order.

Sending setup - the training domain (illustrative)
Domain
✔ Training domain registered to the company - renewed and kept permanently
✘ Real company domain: never used as the simulation sender
SPF record on the training domain: v=spf1 mx a -all
✔ Gateway allowlist: scoped to the test sender and the launch window only
✘ Admin console access: restricted - never left open to the whole internet
The pattern to copy: everything about the sending side is deliberate, owned by you, and switched off or fenced when the round ends.
1

Register a training domain you own - and keep it

Buy a look-alike of your own company domain - the kind of near-miss spelling real attackers use against you - and register it to the organisation permanently. It exists for training only. Two firm boundaries: never register a domain imitating a bank, a courier or any other company; and never let the training domain lapse, because an expired look-alike of your own domain is a gift to real attackers. Put its renewal on the same calendar as the real domain.

2

Set up sending without endangering the real domain

Send simulations only from the training domain, through the platform's sending profile, with its DNS records - SPF and, where your platform supports it, DKIM - configured for that domain, like the SPF line in the panel above. Your real domain's reputation, deliverability and DMARC reports stay completely untouched by the exercise. Keep the platform's admin console fenced - GoPhish, for example, serves its admin interface on port 3333 by default; that interface belongs behind a VPN or an IP allowlist, never published openly.

3

Allowlist the test in your own gateway - narrowly

Your mail security will, correctly, try to block the simulation. Coordinate with whoever runs it: an allowlist entry scoped to the specific sending address and the approved launch window, logged, and removed when the round closes. A permanent, broad bypass is not a testing convenience - it is a hole in your real defences with your name on it. And take the compliment hidden in the block: if the filter stops your test cold, that is evidence it stops real phish too.

4

Run the full path on IT first, and pick the day with care

  1. Send the campaign to IT mailboxes only.
  2. Click through as a user: delivery, link, landing page - and confirm the page records a visit and collects nothing else.
  3. Check the report path end to end: a reported test message must land where real reports land.
  4. Then schedule the real round on an ordinary working day - not payroll week, not an audit, not a crisis. Realism stops where people's worst week begins.
🔴 An allowlist entry without an expiry is a vulnerability you installed yourself. Scope it to the sender, scope it to the window, write the removal date in the campaign plan - and actually remove it.

🎥 Watch the Complete Video

See the sending profile and campaign settings configured in the video, step by step.

▶️ Watch Full Video
❓ Common Questions (FAQ)

Why not simply send from our real domain for maximum realism?

Three reasons: a mishap damages the real domain's mail reputation and deliverability; staff can no longer trust genuine internal mail during tests; and if any part leaks, it looks like a real compromise. The training domain gives all the realism the lesson needs.

The spam filter blocked our test completely. Is the round a failure?

No - it is your defences working, and worth saying so to management. Allowlist narrowly for the approved window, run the round, then close the exception. The filter's behaviour is data, not defeat.

Can the platform run on a small cloud server?

Many organisations do exactly that. Treat it like any internet-facing system: patched, with the admin console restricted to your own access route, and the sending side confined to the training domain - never a shared box that also carries real company services.

This guide is based on the phishing simulation video by Technical Anshu • Stuck on a step? Ask in the video comments - you will get an answer there 🙏