Prevention is mostly discipline

The measures that stop the most incidents cost more attention than money:

  • Patch on a schedule. Most successful attacks use known weaknesses for which a fix already exists. Decide how fast critical fixes go in, and hold to it.
  • Multi-factor authentication everywhere it matters. Email, remote access, admin consoles and cloud accounts first. A stolen password alone should never be enough.
  • Phishing-resistant habits. Short, regular awareness for staff, with simulated phishing if you can. The support desk and finance team are the usual targets.
  • Backups that restore. Keep them separated from the main environment, and test restores. Ransomware is a bad day; ransomware with untested backups is a bad year.
  • Watch your own estate. Logs and monitoring that someone actually reviews, with alerts that lead to action. Unwatched logging only helps the investigation after the damage.

Plan the bad day before it arrives

A breach plan fits on two pages and answers:

  • Who detects and classifies an incident, and who decides it is a personal data breach.
  • Who contains it, and how (isolate systems, revoke access, preserve evidence).
  • Who must be informed, in what order: your leadership, your incident-response retainer or insurer's panel, the authorities the law requires, and the people affected.
  • Who speaks, internally and externally, and what they may say.

Indian rules require prompt reporting of cyber incidents to the national incident-response authority, with very short clocks, and data protection law adds its own duty to inform the data protection authority and affected people when personal data is breached. Regulated sectors have their own reporting clocks on top. The dates-and-details page and your sector's rules hold the specifics; the point here is that these clocks are short, so the plan and the contact list must exist before the incident, not during it.

Drill it

Once or twice a year, run a tabletop exercise: a laptop with customer data is stolen, or a vendor reports a breach, or ransomware locks a file server. Walk the plan with the actual people who would execute it. Every drill finds a gap: an outdated phone number, an unclear decision, a missing approval. Finding it in a drill is free. Finding it in an incident is not.

After an incident

Contain first, then investigate, then notify as required, then learn. Preserve evidence from the start: logs, timelines, what was affected, what you did and when. That record serves the authorities, your insurer and your own review. Close with an honest post-incident note: what failed, what worked, what changes.

Where organisations go wrong

  • A plan written for an audit that the response team has never read.
  • Backups on the same network the ransomware reaches.
  • Notification duties discovered during the incident, with the clock already running.
  • Treating the insurer's helpline as the plan. Insurance pays some costs; it does not contain the breach.

Go deeper

Next step

The tooling that supports all of this, matched to budget: tools, open source and paid.