The most common way a project manager gets caught out in front of a board is not a wrong number. It is two right numbers. The Status Report says amber. The RAID log, in the appendix, has a red issue nobody mentioned. A board member reads both and asks which one is true. That question is hard to answer well in the room. Here is how to stop it.
Why the three documents drift apart
Three honest reasons, none of them about competence:
- Different days. The RAID log is updated on Tuesday after the risk review. The milestone report is cut from the plan on Thursday. The status report is written on Friday night from memory of both.
- Different people. The PMO analyst owns the log. The planner owns the milestones. The PM owns the narrative. Each is right about their own document.
- Copy and paste. Last month’s status report is the template for this month’s. A date that did not get changed is now wrong, and it looks exactly like the ones that did.
The fix is not to tell people to be more careful. It is to change where the facts live and when the documents are cut.
Rule one: one source, one period
Every fact that appears in more than one document gets exactly one home. Dates live in the plan. Risks and issues live in the RAID log. Money lives in the budget tracker. The status report and the SteerCo pack summarise those sources; they never hold a fact of their own. If a date in the status report differs from the plan, the status report is wrong by definition, whatever the PM remembers.
Then cut all three from their sources in the same window. Freeze the plan and the log on the same day, build the reports from the frozen copies, and do not touch the sources again until the pack is out. If something changes after the freeze, it goes in next period, or it goes in as a dated late-change note on page one. Never silently.
Rule two: the five facts that must match
Put the three documents side by side and check these, in this order:
- The overall colour. The status report RAG is not kinder than the worst open item in the RAID log or the worst milestone in the milestone report, unless the report says why, with a name against the reason.
- The go-live date. One date, in all three documents, and the same as the plan. If the forecast differs from the baseline, both are shown, and the gap has a reason.
- The top risks and issues. The three to five the status report talks about are the top rows of the RAID log, by rating. Not the ones the PM finds most interesting.
- The milestone dates. Every milestone named in the status narrative carries the date the milestone report shows. Not last month’s date.
- The budget line. Spend to date and forecast at completion are the same figures, to the pound, as the budget tracker. Not rounded differently in each.
This takes about ten minutes. It is checks one and five from What checks should I run before I issue a project report?, done properly.
Rule three: when they disagree, decide and record
You will find mismatches. That is the point. When you do:
- Go to the source. The document closest to the fact wins. The plan beats the narrative. The log beats the summary slide.
- Fix the other document. Do not split the difference and do not leave both.
- If you cannot fix it in time, say so in the pack. “RAID item 14 is rated red; the overall status is held at amber because the mitigation lands on 3 November. Decision: J. Patel, 29 October.” A board will accept a reasoned exception. It will not accept a surprise.
- Write down who decided. A name and a date, in the pack or in a log next to it. Six months later, when someone asks why the colour was amber, you want the answer in writing, not in someone’s memory.
The RAID log carries the most weight here, because it is the document most packs get wrong. The rules for keeping it honest are in What is a RAID log, and how do I keep it honest?
A real example
A programme we took over had three documents and three go-live dates. The plan said 14 March. The milestone report, cut two weeks earlier, said 28 February. The status report, copied forward from January, said “end of February”. All three had been issued to the same board in the same pack. Nobody was lying. Nobody had compared them.
We did three things. We made the plan the only home for dates. We moved the freeze for all three documents to the same Wednesday. And we added a ten-minute side-by-side to the issue checklist, with a tick box and a name. The next pack had one date. The one after that had a slip, shown as baseline and forecast, with the reason beside it. The board noticed the slip. It also noticed that the pack had stopped contradicting itself, and said so.
Where PMOpack fits
PMOpack builds the Status Report, RAID log, Milestone Report and the rest of the pack from the one file you upload, in one period, so rule one is the default rather than a discipline. Then fixed checks, which are not AI, run across the whole pack at generation time and compare the artefacts in the same period with each other: the overall colour against the items beneath it, dates across artefacts, an amber or red without a path to green, a placeholder owner on a decision. A critical finding blocks publish until a person fixes the data or accepts it with a written reason, and the acceptance is stored and cannot be edited later. That stored reason is rule three, kept for you. A check skips, and says so, when the partner artefact is missing. It checks figures for agreement with each other, not against your finance system, and it does not promise to catch every error. A person still decides which document is right.
Questions people ask
- Why do my status report and RAID log disagree?
- Usually because they were updated on different days by different people from different sources. Neither is wrong. They describe different moments. Fix the process, not the person.
- Which document wins when two reports disagree?
- The one closest to the source. For dates, the plan. For risks and issues, the RAID log. For money, the budget tracker. The status report summarises those; it never overrides them.
- How long does a side-by-side check take?
- About ten minutes for a standard monthly pack, if you check the same five facts in the same order every time.
- Should the overall RAG match the worst item in the RAID log?
- The overall colour should never be kinder than the worst open item without a written reason. If the log holds a red issue and the report says amber, say why in the report, with a name.
- Can a tool check that my reports agree?
- The mechanical part, yes: same colour, same dates, same top risks, roll-ups that add up. A tool can flag the mismatch. Deciding which document is right is still a person's job.