Project delivery
Tailoring decision record
What the standard says, what this project does instead, the reason, who agreed it and when to revisit. Includes a section on what was deliberately not tailored, which is what separates a decision from an omission.
What it is
Tailoring is deciding what a particular project actually does out of everything the method offers, and at what formality. The output is not a lighter process. It is a stated decision with reasoning that somebody can challenge later, and this record is that output.
The reason column carries the whole thing. An undocumented departure is indistinguishable from an omission and will be read as one, whereas a recorded decision lets a sponsor who wants the full standard applied be answered with something better than preference.
The section on what was deliberately not tailored matters as much as the changes. It shows the exercise distinguished between a practice and a control rather than removing whatever was inconvenient, and the test is who is harmed by the omission. A report nobody reads harms nobody. A design review that exists so a second qualified person sees the work protects whoever uses the output.
When to use it
- A standard method is being applied to a project it does not fit.
- A project management office has asked why you are not following the standard.
- Artefacts were dropped as overhead and something has now fallen through the gap.
- A project has components with genuinely different needs and one process is serving none of them.
What is in it
- What this project is
- The decisions
- What was not tailored, and why
- The test used
- Revisit
- How to use this
The template
Project. Settlement rebuild · Owner. Project manager · Agreed with. Project management office, 4 February
What this project is
Two or three lines establishing the context every decision below is fitted to. Size, duration, team, novelty, regulatory position and how costly a mistake is.
Twelve months, a blended team of fourteen across three locations, replacing a process that already exists so the outcome is well understood. Regulated, with a fixed November reporting deadline and an external certification on the critical path. Two of the three components have been delivered in similar form before.
The decisions
| Element | Standard says | This project does | Reason | Agreed with | Revisit |
|---|---|---|---|---|---|
| Life cycle | One approach per project | Predictive for the migration and certification, adaptive for the exception queue | The migration has a fixed sequence and an external gate. The queue has requirements operations cannot describe until they use it | Sponsor, 4 February | At phase two planning |
| Stage gates | Four formal gates with a full pack each | Two gates, plus continuous assurance on the adaptive component | A gate exists to decide whether to keep committing. Applying that to a fortnightly increment costs more than the review is worth | Project management office, 4 February | July |
| Change control | All changes to a board meeting monthly | Project manager decides below 25000 and below any milestone effect. Above either goes to the board | Monthly board against fortnightly delivery produced five week decision latency on the last programme | Sponsor, 4 February | At phase two planning |
| Estimating | Three point estimates per activity | Three point for the migration, relative sizing for the queue | The team delivering the queue is stable and has history to forecast from. The migration is priced by a contractor | Delivery lead, 4 February | Not planned |
| Status reporting | Weekly written report to all stakeholders | Per audience, per the communications plan. Board pack monthly | One report to forty people was simultaneously too much and not enough on the last programme | Project management office, 4 February | June |
| Design review | Second qualified engineer reviews before build | Unchanged | This is a control rather than a practice. Whoever uses the output is harmed if it is skipped | Not tailored | Never |
| Compliance evidence | Assembled for the annual audit | Built as each release is made | Evidence assembled retrospectively records what people can reconstruct, not that the control operated | Regulatory affairs, 4 February | Never |
What was not tailored, and why
Naming these matters as much as naming the changes, because it shows the exercise distinguished between practice and control rather than removing whatever was inconvenient.
The design review, the compliance evidence requirement, the health and safety induction for site work, and the procurement thresholds all stay as the standard sets them. Each is an obligation owned by somebody outside the project, and the route to changing one is a waiver from that owner rather than a tailoring decision here.
The test used
For each element, what job was this practice doing, and is that job needed here?
A practice goes when the job is not needed on this project, or when something lighter does the same job. Where a practice was dropped, the row above says what now does its work. A practice that went because it was tedious would leave the job undone, which is under tailoring rather than tailoring.
Revisit
A tailoring decision taken at initiation was taken with the least information anybody on this project will ever have. The Revisit column puts each one against a review point already in the calendar rather than creating new ones.
Uncertainty falls as work proceeds, so a component needing short cycles at the start may be understood well enough by phase two to be planned out, and one assumed to be routine may turn out to need the loop after all.
How to use this
The record is the deliverable. The output of tailoring is not a lighter process, it is a stated decision with reasoning that somebody can challenge later. An undocumented departure is indistinguishable from an omission and will be read as one.
Agree it rather than announce it. The organisation sets what is fixed, the team supplies most of the knowledge about what the work needs, and governance approves departures. What the project manager owns is the reasoning and this record of it.
Adjust elements together. Dropping a report while leaving the process that produced it and the governance that expected it in place creates a gap somebody finds at the review, and the usual conclusion is that tailoring does not work.
Keep it to one page. A record covering every practice in the method is a copy of the method. Cover what was changed, what was deliberately not changed, and why.
Questions people ask
- Is tailoring always about doing less?
- No. A project with three suppliers and a regulator needs more traceability than the standard assumes, and deciding that is tailoring just as much as dropping a report.
- Who approves a tailoring decision?
- The organisation sets what is fixed, the team supplies most of the knowledge about what the work needs, and governance approves departures from a standard. What the project manager owns is the reasoning and the record.
- What is under tailoring?
- Removing a practice without replacing the job it was doing. The overhead disappears, the project feels lighter for a few weeks, and then somebody asks a reasonable question nobody can answer.