OneNym.idDiscuss deployment
English

THE HARDENED STACK / Appendix B

A 30/60/90-Day Hardening Plan for a New IT Lead

When you inherit security for an organisation that never had an owner, the instinct is to fix the first alarming thing you see. This appendix offers a calmer sequence for the first ninety days, written for someone who still has a full operational workload.

This page summarises the chapter. The full text is in the book.

Two principles for the whole plan

The first is not to break the business while securing it. Prefer the safe, reversible version of each change, announce it in advance, give people a way to report problems, and keep a rollback ready. The book treats trust as a control in its own right: a loud failure in week one makes security look like the thing that stops work.

The second is to document as you go: the inventory, the administrators you find, the exceptions you grant and the reasons. That record becomes the foundation for everything afterwards and is far cheaper to write during the work than to reconstruct later.

Days 0-30: visibility and stopping the bleeding

Build a rough inventory of devices, servers, cloud accounts and SaaS applications with their owners, and record the gaps as findings. In parallel, confirm MFA for every user (administrators and finance first), confirm that backups exist with a copy out of ransomware's reach and actually test a restore, switch on audit logging in Microsoft 365 and the cloud accounts, and identify and start trimming administrators. The cyber-insurance questionnaire is a useful free gap analysis.

Days 31-60: hardening identity and endpoints

Put a conditional-access baseline in place that enforces MFA, blocks legacy authentication and reacts to risky sign-ins. Move administrators to phishing-resistant MFA and separate admin accounts, and create offline break-glass accounts. Start removing local admin rights and rotating the local admin password, handling misbehaving applications one by one. Confirm patching and EDR, enable and escrow full-disk encryption, and put SPF, DKIM and DMARC in monitoring mode.

Days 61-90: processes that keep going

Formalise joiner-mover-leaver so access is reliably revoked, centralise logging with a few trusted detections, move DMARC toward quarantine, start a scanning and patch-SLA rhythm, and write down incident basics using Appendix C. Then agree the next quarter's priorities with leadership and tell the business plainly which risks are being consciously left for later. The book also cautions against buying a new platform in the first quarter; fundamentals cost mostly time, and tooling should come later to remove a specific, felt burden.

KEY POINTS

Key points

  • In the first month, test an actual restore, not just the existence of backups.
  • Verify the list of administrators yourself; there are usually more than expected.
  • Use the cyber-insurance questionnaire as a free gap analysis.
  • Put DMARC in monitoring mode first and enforce only after legitimate senders are fixed.
  • Write down every exception and decision with its reason as you work.
  • End day 90 with an agreed plan for the next quarter and a written list of accepted risks.

QUESTIONS

Common questions

What should a new IT security lead do in the first 30 days?
Focus on visibility and the few controls that prevent the most common disasters. Build a rough asset inventory with owners, confirm MFA for all users starting with administrators and finance, confirm backups exist with one copy out of ransomware's reach and test a restore, turn on audit logging in Microsoft 365 and cloud accounts, and verify who holds administrator rights.
How do you harden security without disrupting the business?
Choose the safe, reversible form of each change, announce it before making it, give users a clear way to report problems, and keep a rollback plan ready. Make adoption easy, for example by helping people enrol in MFA, rather than granting exemptions. A disruptive change early on can cost months of goodwill and make security look like an obstacle.
What does a 30/60/90-day security plan look like for a small IT team?
Days 0-30 establish visibility and stop the biggest risks: inventory, MFA for all users, a tested restore and audit logging. Days 31-60 harden identity and endpoints with a conditional-access baseline, phishing-resistant MFA for admins, reduced local admin rights, disk encryption and DMARC in monitoring mode. Days 61-90 build lasting processes such as a reliable leaver process, trusted alerts, DMARC enforcement, a scan-and-patch rhythm and a plan for the next quarter.
Should a new IT lead buy a security platform in the first 90 days?
The Hardened Stack advises caution. A tool the team has not yet learned to operate, tune and act on will not improve security in the first quarter and adds another project. Finishing the fundamentals first costs mostly time rather than licence fees, and tooling is better bought later to remove a specific burden the team has actually experienced.
Why should IT tell leadership which security risks are not being addressed?
A small team cannot cover everything, and making the uncovered areas explicit turns them into accepted, shared decisions rather than silent gaps. This protects the IT lead from blame and lets leadership decide whether to fund a risk they would otherwise never have been told about.