Skip to content
Search Kaul Advisors
Toolkit 3 min read

A practical RAID log for ERP programmes

Most RAID logs become archives. A working log is short, owned, and tied to decisions — here is a structure that stays useful under pressure.

SK

Sunny Kaul

Principal Advisor, Kaul Advisors

Hands typing on a laptop during a focused working session

A RAID log is one of the first artefacts a programme creates and one of the first to stop being read. It grows into a long list of items with vague owners and no decisions, maintained because the methodology requires it.

A working log is different. It is short, it changes, and it is reviewed where decisions are made.

Separate the four categories, then connect them

Risks, assumptions, issues, and dependencies are different things. Mixing them into one undifferentiated list makes it hard to see what needs action now.

  • Risks are uncertain events that could affect the programme.
  • Assumptions are things believed to be true but not yet verified.
  • Issues are problems that have already happened.
  • Dependencies are work or decisions the programme needs from somewhere else.

The value comes from connecting them. An assumption that proves false becomes an issue. An issue may create a new risk. A dependency may be the trigger for both.

Give every item a trigger and a response

“Risk: data quality” is not actionable. It does not tell anyone when to act or what action to take.

A useful entry contains:

Field What it should answer
Description What could happen, specifically?
Owner Who is accountable for managing it?
Trigger What signal tells us it is happening?
Response What will we do if it does?
Date When will this be reviewed again?

The trigger is the field most often missing and most valuable. It turns a passive list into an early-warning system.

Keep the active list short

A RAID log with two hundred open items is not a control. It is a sign that the programme has stopped making choices about what matters.

We keep a working list of the items that could change a decision in the next reporting period. Everything else lives in a reference register. Items move onto the active list when their likelihood, impact, or timing changes.

This sounds like extra administration. In practice it reduces it, because the governance conversation focuses on a manageable set of live questions.

A RAID log should help a programme decide what to do next. If it only records what might go wrong, it is a diary, not a control.

Record decisions separately

Risks and issues generate decisions, but the decision itself needs its own record. A decision log should capture:

  • the decision made
  • the date and decision owner
  • the options considered
  • the rationale
  • what changes as a result

This is particularly important in ERP programmes, where the same design question can return in a different workstream three months later. Without a decision record, the organisation relitigates it from memory.

Review in the right room

The full RAID log does not belong in every meeting. Different items need different audiences.

  • Delivery stand-up: active issues and near-term dependencies.
  • Design authority: assumptions and risks that affect process or solution decisions.
  • Steering committee: risks and decisions that require senior authority or investment choices.

A short, consistent review at each level keeps the log current without turning governance into a reading exercise.

A simple test of health

At the end of each review, ask three questions:

  1. Did we close, change, or escalate at least one item?
  2. Does every active item have a named owner and a next date?
  3. Is there anything on the list that no one discussed?

If the answer to the third question is yes, the log is drifting. If the first two answers are consistently yes, it is doing its job.

Topics RAIDGovernanceToolkit