Security, Compliance and Data Privacy in IT Projects

Security used to be something handled by a separate team just before go-live. That approach produces expensive surprises: a penetration test in the final week finds a design flaw that would take months to fix, or a privacy review discovers that the project collects data it is not allowed to keep. Today, security, compliance and data privacy are part of project planning from the first week. A project manager does not need to be a security specialist, but must know what questions to ask, who must be involved and where to build security work into the plan. This lesson is general education and not legal advice; always confirm legal obligations with your organisation’s legal, compliance and security teams.

Three Related Concerns

ConcernCore questionTypical owners
SecurityIs the system protected against misuse, attack and data loss?Security team, architects, engineers
ComplianceDoes the project meet the laws, regulations, standards and contracts that apply?Compliance, legal, risk and audit teams
Data privacyIs personal data collected, used, stored and shared lawfully and fairly?Data protection officer, legal, product owners

They overlap: a data breach is a security failure, a privacy incident and often a compliance breach at once.

The Confidentiality, Integrity and Availability Triad

Information security is often described using three goals. Confidentiality means only authorised people can see the data. Integrity means data is accurate and cannot be changed improperly. Availability means authorised users can access systems when needed. Requirements for each should appear in the project’s non-functional requirements (see Requirements Gathering for IT Projects).

Frameworks and Regulations You May Meet

Which apply depends on your industry, customers and countries. Examples:

  • ISO/IEC 27001: an international standard for information security management systems.
  • SOC 2: an audit report on controls for service organisations, often requested by customers.
  • PCI DSS: security standard for organisations that handle payment card data.
  • GDPR: the European Union’s data protection regulation, which can apply to organisations outside Europe that handle EU residents’ data.
  • India’s Digital Personal Data Protection Act, 2023: governs processing of digital personal data in India. Check the current rules and timelines with your legal team, as implementation details develop over time.
  • HIPAA: United States rules for health information.
  • Sector regulations: for example banking and financial regulators’ requirements on outsourcing, cyber resilience and data localisation.

The project manager’s task is to find out early which apply, who the expert owner is and what evidence will be needed.

Build Security Into the Project Lifecycle

PhaseSecurity and privacy activities
InitiationClassify the data involved; identify applicable regulations; involve security and legal stakeholders; note security risks in the charter
PlanningAdd security requirements, tasks and budget (testing, tooling, audits); plan a privacy impact assessment if personal data is involved
DesignSecurity architecture review; threat modelling; decide authentication, authorisation, encryption and logging approach
BuildSecure coding practices; code review; dependency and secrets scanning; least-privilege access
TestVulnerability scanning; penetration testing; access control testing; data masking in test environments
Go-liveSecurity sign-off as an entry criterion; monitoring and incident response ready (see Go-Live, Deployment and Cutover Planning)
Operations and closureHandover of security responsibilities; retention and deletion rules; decommissioning that securely removes data

Moving security activities earlier is often called “shifting left”. It costs less to fix a design weakness during design than after release.

Core Practices a Project Manager Should Expect

  • Data classification: label data by sensitivity (for example public, internal, confidential, restricted) so that controls match the risk.
  • Least privilege and role-based access: people get only the access their role needs, and access is reviewed.
  • Encryption: protecting data in transit and at rest.
  • Logging and monitoring: to detect and investigate suspicious activity.
  • Threat modelling: a structured session asking “what could go wrong and how do we prevent it?”
  • Secure configuration and patching of platforms and components.
  • Backup and recovery, tested periodically.
  • Incident response plan: who does what if a breach is suspected, including notification obligations.

Privacy by Design

For projects that process personal data, ask early:

  1. What personal data do we collect, and is every item necessary?
  2. What is the lawful basis or consent for using it?
  3. Who can access it, and where is it stored, including which country?
  4. How long do we keep it, and how is it deleted?
  5. Do we share it with vendors or other parties, and under what agreement?
  6. How do individuals exercise their rights, such as access or correction?

A privacy or data protection impact assessment is a structured way of answering these questions for higher-risk processing. Collecting less data is the simplest privacy control: data you never collected cannot leak.

Vendors and Third Parties

Your project’s security is only as strong as its suppliers. During procurement, ask vendors about certifications, security testing, incident history, data location, subcontractors and breach notification. Include security and privacy obligations in contracts and SLAs (see Vendor and Procurement Management). Review access given to vendor staff and remove it at the end of the engagement.

Managing Security Risk in the Plan

  • Add security risks to the risk register with owners and responses (see Project Risk Management).
  • Schedule security reviews and testing explicitly, including time to fix findings. A penetration test with no time to remediate is a formality.
  • Define security acceptance criteria: for example, no open critical or high vulnerabilities at go-live without a documented exception approved by the right authority.
  • Include compliance evidence collection as tasks, because auditors need documents, not intentions.

Worked Example: Customer Portal Launch

A retailer plans a customer portal storing names, addresses and order history. At initiation, the project manager meets the security lead and the data protection officer. They classify the data as confidential personal data, agree that payment details will be handled by an external payment provider so the portal never stores card numbers, and schedule a threat-modelling workshop during design. A pre-launch penetration test is planned three weeks before go-live, with a week reserved for fixes. The test finds a weakness in password reset. Because time was planned, the team fixes and retests it before launch, and the project meets its date with a documented security sign-off.

Common Mistakes

  • Leaving security reviews until the last month.
  • Using real customer data in test environments without protection.
  • Treating compliance as the security team’s problem alone.
  • Giving vendors broad access and never removing it.
  • Planning testing but not the time to fix what it finds.
Key takeaways
  • Involve security, legal and privacy experts from initiation, not just before go-live.
  • Classify data, apply least privilege and design privacy in from the start.
  • Add security tasks, testing, remediation time and acceptance criteria to the plan.
  • Manage vendor risk and keep an incident response plan ready.

Related Lessons