In short
- People write where they are read. Adoption starts with the sponsor reading from the live record.
- An old template with no end date will outlast any platform. Name the date and hold it.
- Measure adoption by the decisions taken from the record. A login only shows that someone found the page.
The contract was signed in the spring. The system was configured over the summer, and every project manager sat through a morning of training. Go-live was announced in an email to all staff.
Three months later it is Wednesday afternoon, and the PMO analyst is building the weekly report. She opens the spreadsheet she used last year and writes to the project managers for their numbers. The platform is open in another tab. Its dashboard still shows the figures from go-live week.
Nobody decided to abandon it. Each person chose where to spend ten spare minutes, and the choices added up. I think this is the most expensive way for software to fail, because nothing breaks and so nobody is asked to fix it.
What goes wrong
A stalled rollout rarely has a dramatic cause. It has four ordinary ones, and each makes the others worse.
The platform is a second place to type. The steering committee still receives the old pack, so the old pack still has to be filled in. The platform asks for the same dates and percentages again. An owner with ten minutes on a Wednesday gives them to the document the committee reads. The platform is left for later, and later does not come.
The first screen is empty, or it belongs to someone else. A new user logs in and finds a blank dashboard, or sample projects left over from the demonstration. Neither says anything about their own work. They close the tab having learned that there is nothing here for them yet, and that lesson is hard to undo.
Nobody senior reads from it. People write where they are read. If the sponsor asks her questions from the slide pack, the slide pack is the system of record, whatever the contract says. A project manager who keeps the platform up to date and is never asked about it will soon stop, and will be right to.
Every department starts on the same day. A rollout to the whole organisation spreads the support thin. Each department meets its own snag in the same week, and none gets enough help to get past it. No single portfolio becomes complete enough for a committee to rely on. A platform used a little everywhere can be trusted nowhere.
Together these form a loop. The data is thin, so nobody senior trusts it. Nobody senior reads it, so nobody keeps it up. The old template stays because the platform cannot yet be relied on, and the platform cannot be relied on because the old template stays.
More training leaves the loop intact. It teaches people how to enter an update and gives them no reason to.
What good looks like
None of these moves depends on which platform was bought. They are decisions about the rollout, and most of them belong to the sponsor.
Start with one portfolio and its real data
Choose one portfolio whose head wants this to work. Load its real projects, owners, dates and budgets before anyone is invited in, so that each person's first screen shows their own work under their own name. Remove the sample data. Bring in the second portfolio only when the first has run a full reporting cycle without the spreadsheet.
Make it the only source of the committee's pack
Decide that the steering committee's pack is produced from the platform and from nowhere else. A project that is missing from the system is then missing from the pack, and the committee will ask why. In the room, the sponsor asks her questions from the live record. Once people see what is read, they know where to write. There is more on building the pack this way in What a PMO should stop doing by hand.
Withdraw the old template on a named date
Two places to report will run side by side for as long as they are allowed to. Announce the date on which the old template stops being accepted, give fair notice, and have the sponsor hold it. After that date, an update sent in the old format goes back to its sender with a link. The first refusal is uncomfortable. It is also the moment the rollout becomes real for everyone watching.
Measure adoption by the decisions taken from it
A login count rewards opening a page. Count what the organisation did from the record: decisions minuted with a reference to it, approvals given in it, packs produced with nothing copied by hand. Report those to the sponsor each month. If a committee met and decided nothing from the platform, that is the adoption figure for the month. The meeting itself is the subject of The weekly PMO meeting should not start with a status round.
People write where they are read.
Expect the spreadsheet to come back
The old way of reporting does not leave quietly. Three habits are worth watching for in the first quarter.
The private copy. Someone keeps a sheet "just to check" the platform. Within weeks the pack is built from the sheet again. Treat a second copy as a fault in the record, and fix the record.
The senior exception. One executive prefers the old pack, and a team is asked to keep producing it for her alone. Everyone below her has two places to type again.
The date that moves. The withdrawal date arrives in a busy week and is pushed back a month. People conclude that the next date will move too.
Take the update where the owner already works
The last move lowers the cost of writing. An owner who has to log in to a separate system to report will do it last, if at all. Take the update where the owner already is: a reply to an email, a message in the chat tool the team uses, the schedule file the planner already keeps. The system's job is to turn that into a proposed change to the record, show it to a person and wait for a yes.
We built Lodestar to work this way. It imports Microsoft Project and Primavera P6 schedules saved as XML. An owner can answer the weekly request by replying to an email, and the reply becomes a draft update for review. In Microsoft Teams, Slack or WhatsApp, the Polaris assistant shows a proposed change before and after, and nothing changes until the owner presses Apply. Reports and board packs come from the same live model.

Questions for your next review
- Which document did this committee read from last month, and where were its numbers typed?
- On what date does the old reporting template stop being accepted, and who will hold that date?
- Which portfolio is complete enough in the system today for us to run a review from it alone?
- How many decisions in last quarter's minutes can be traced to the live record?
- Where does an owner have to go to give an update, and how many times do they give it?
A platform is adopted when the people at the top of the programme read from it and the old way of reporting has an end date. Until then it has only been installed.
