---
name: student-progress-and-milestone-planning
description: >
  Builds realistic research milestone plans backwards from the submission
  deadline, including the dependencies students and supervisors systematically
  underestimate, and identifies the student at risk of stalling early enough to
  intervene. Use for "build a timeline for my student's thesis", "is this project
  deliverable in the time left", "my student has gone quiet", "the scope has grown
  and we need to cut it", "we missed a milestone, what now", "my student keeps
  rewriting chapter one", "document progress for the annual review".
category: 15 Academic University Research
ref: "15.23"
tier: 3
inherits: [K2, K3, K4, K5]
---

# Student Progress and Milestone Planning

## 1. One-line description

A planning and monitoring method that builds a research timeline backwards from the submission deadline with the systematically underestimated dependencies included, expresses milestones as deliverables rather than activities, detects the early signals of a stalling student, and supports the scope, recovery and deliverability conversations that follow.

## 2. What this skill is used for

**Read this before you start.** Planning and progress monitoring involve a named student's personal data, and frequently sensitive circumstances: health, caring responsibilities, finances, visa status. Whether an AI system may be used on this material, and what may be recorded where, is governed by your institution's data protection and student records policy, and by its policy on AI in supervision. Settle both before the student's file is opened. Section 5 makes the policy check a required input; Section 12 sets the operating rules. Where the position is unknown, the conservative default is that no identifying or personal circumstance material is uploaded, and planning is done on the project rather than on the person.

**The research problem it solves.** Research projects overrun for reasons known in advance and planned around anyway. Timelines are built forwards from today, which produces a plan that ends when the work ends rather than when the deadline is. They are built out of activities rather than deliverables, so "working on the literature review" occupies four months and produces nothing checkable. And they omit the same dependencies every time, because those dependencies belong to other people: ethics approval, recruitment, data access, transcription and the supervisor's own reading time are not in the student's control, and they are where the months go. Meanwhile the student who is quietly failing looks, for a long time, like a student who is working: reading widely, refining chapter one, expanding the scope. By the time the absence of output is undeniable, the recoverable period has passed. This skill builds the plan backwards, forces milestones to be artefacts, names the signals early, and separates the three kinds of blocked student, who need three different interventions.

**Where it sits in the research lifecycle.** At the start of a supervised project, at every formal progress point, and at every moment when something has slipped. It is the skill that runs continuously in the background of a supervision relationship.

**Typical use cases.**
- Building a defensible timeline at the start of a thesis or dissertation.
- Re-planning after ethics approval or recruitment has taken twice as long as assumed.
- Recognising, naming and acting on the early signals of a stalling student.
- Having the scope-reduction conversation before it becomes a deliverability conversation.
- Producing a recovery plan after a missed milestone that is honest about what has to give.
- Assessing whether the project as currently defined can be delivered in the time left, and what the alternatives are.
- Documenting progress and interventions for an annual review, progression panel or extension request.

**Who uses it.** Supervisors and co-supervisors; directors of postgraduate study and programme leads monitoring a cohort; progression and annual review panels; research degree administrators preparing review documentation. It is also usable by a student with their supervisor, and it is more effective when built jointly than when handed over.

## 3. When to use it

- A project is starting and a timeline is needed that will survive contact with reality.
- Ethics, access or recruitment has taken longer than assumed and the whole plan needs rebuilding.
- A student's contact has become sporadic, or their updates have stopped containing specific questions.
- The project has grown: new research questions, an added case, a second method, a wider population.
- A milestone has been missed and the response so far has been to move it.
- A student is rewriting an early chapter for the third time and producing nothing new.
- A progression review, annual review or upgrade is approaching and progress has to be evidenced.
- You suspect the project is no longer deliverable in the time remaining and need to establish whether that is true before saying it.

## 4. When NOT to use it

- **The problem is the quality of the work, not the pace of it.** A student producing work on time that is not good enough needs developmental feedback, and a schedule will not supply it. Go to **15.20 Supervision Feedback and Development**. Re-planning in response to a quality problem produces a student who delivers the same inadequate work sooner.
- **The student's difficulty is not academic.** Illness, bereavement, caring responsibilities, financial crisis, harassment, visa or immigration pressure and mental health difficulty all present as missed milestones and avoidance of contact, and they are the more likely explanation for a sudden change in a previously reliable student. A milestone plan applied to a non-academic problem adds pressure to someone who needs support and delays the referral. `RESEARCHER DECISION REQUIRED` (K5 §2.4): the supervisor judges whether this is a planning situation, and where there is any doubt, routes to the institution's support services first. This skill does not assess a student's circumstances and must never be used to build a case about them.
- **The decision is a formal progression, termination or extension decision.** These are institutional processes with defined criteria, evidence requirements and rights of appeal. This skill produces the evidence and the plan; it does not make the decision, recommend an outcome, or supply language for a decision the academic has not independently reached. See §12.1.
- **The methodological problem is being treated as a scheduling problem.** A study that cannot answer its question does not become deliverable by being given more time. Where the design is the issue, go to **15.08 Research Design and Methodology Chapter**, or audit it with **15.22 Research Rigour Audit**.
- **The plan is being built without the student.** A timeline handed to a student is a document they will nod at and not use. The dependencies that matter, their other commitments, their access constraints, their working pattern, are things only they know. Build it with them, or do not build it.
- **The deadline is not actually fixed.** Where an extension is genuinely available, planning against the old date produces false urgency and a plan nobody believes. Establish the real endpoint first, including whether the registration period and the funded period are the same date, which they frequently are not.
- **The project has already failed on the timeline and what is needed is a decision, not a plan.** Where no honest plan reaches the deadline, producing an optimistic one is the least kind available option. Say so, and move to the alternatives conversation in Step 11.

## 5. Required inputs

**Required.**
- **The real submission deadline, and what sets it.** Registration end, funding end, visa expiry, a maximum period of study, or a programme date. These are often different dates and the binding one is not always the latest. Never assume.
- **The current state of the project, expressed as artefacts that exist.** Not "the literature review is underway". A word count, a draft, a completed ethics submission, a recruited sample.
- **The regulations governing the degree**: the required milestones, the progression points, what extensions or suspensions exist, and their conditions.
- **Institutional policy on AI use in supervision and on student data.** If unknown, resolve before proceeding.
- **The student's participation.** See §4.

**Optional, and what each one adds.**
- **The student's other commitments.** Teaching, employment, part-time status, caring responsibilities, clinical or professional practice. Converts an abstract number of months into actual available working weeks, which is usually a much smaller number and the single most common reason plans fail.
- **Institutional turnaround times.** Ethics committee meeting dates and typical review cycles, data access approval periods, examiner nomination lead times, and the formal submission and examination windows. These are knowable in advance and are the difference between a plan and a wish.
- **The supervisor's own availability.** Leave, sabbatical, conference periods, and honest reading turnaround. A supervisor who takes three weeks to read a chapter and plans on one is the schedule's hidden critical path.
- **The history of previous milestones.** The best predictor of how long the next chapter takes is how long the last one took. A student's own track record beats any generic estimate.
- **Comparable completed projects in the group.** Gives realistic durations for recruitment, fieldwork and transcription in this specific setting, which generic estimates get badly wrong.
- **The student's own account of where they are stuck.** Frequently the fastest route to the correct diagnosis in Step 8, and routinely more accurate than inference from the schedule.

## 6. Questions to ask before starting

1. **What is the real deadline and what sets it?** Determines everything. Default: identify every candidate date, take the earliest binding one, and state which it is in the plan.
2. **How many working weeks does this student actually have?** Not calendar months. Subtract teaching, employment, known leave, and the periods when institutional processes are closed. Default: assume a part-time student has substantially fewer usable weeks than the pro-rata figure, and ask rather than assume for anyone.
3. **What has to be approved, granted or provided by someone else?** Ethics, data access, gatekeeper permission, equipment, licences, translation, a co-supervisor's sign-off. Default: list every external dependency and assign each a realistic elapsed time including one failed round, because first submissions are frequently returned.
4. **What degree structure and examination model applies?** Whether there is an upgrade or transfer point, whether a viva follows submission, how long examiner nomination and appointment take, and whether corrections time sits inside or outside the registration period. These vary by country, institution and degree, and they change where the plan ends. Default: ask, and if unavailable, plan to submission only and state that the post-submission period is unmodelled.
5. **Is the scope currently deliverable, on the evidence so far?** Ask this at the start, not only when it is already too late. Default: test the current scope against the available weeks explicitly, and record the answer.
6. **What is the earliest signal we would agree counts as a warning?** Agreeing this with the student in advance converts a later difficult conversation into a pre-agreed trigger. Default: propose two missed checkpoints, or six weeks without a deliverable.
7. **What happens if the worst dependency fails?** Recruitment does not reach target, ethics is refused, access is withdrawn. Default: name the fallback for the single riskiest dependency at planning time, not when it fails.

## 7. Step-by-step methodology

**Step 1. Settle the policy and data position, and fix the endpoint.**
Record the AI and data position before opening the student's file. Then establish the binding deadline: list every candidate date, registration end, funding end, any visa or maximum-period constraint, and the programme's submission window, and identify which binds. A plan built against the wrong date is not a plan with an error in it, it is a different plan. *Correct result: one date, one source, recorded and agreed with the student.*

**Step 2. Work backwards from the endpoint, not forwards from today.**
Start at the deadline and move left. The last block is not the last chapter: it is formatting, proofreading, the final reference check, figure and table production, printing or upload, and the institution's own submission process, which together take longer than anyone plans and cannot be compressed because they are mechanical. Place that block first, hard against the deadline. Then the full-draft-to-supervisor block, the supervisor's reading time, the student's response time, the chapters, and so on leftwards. Forward planning consumes the buffer invisibly, because each task expands into the space available; backward planning makes the shortage visible now, which is when it can be acted on. *Correct result: a backwards schedule that reaches a start date, which may be in the past, and if it is, that is the plan's first finding.*

**Step 3. Build the dependency map, with the systematically underestimated items named.**
Every project has dependencies outside the student's control, and the same ones are underestimated every time because they belong to other people and stay invisible until they bind.

| Dependency | What is systematically underestimated |
|---|---|
| **Ethics approval** | The committee's meeting cycle, and that a first submission is commonly returned with conditions. Budget one revision round and a full cycle for it. |
| **Recruitment** | That the achievable rate is a fraction of the assumed rate, that it is slowest at the start, and that a fallback strategy must be planned before it is needed. |
| **Data access** | Approvals, agreements, gatekeeper permissions, secure environment onboarding, and the gap between agreement in principle and data arriving. |
| **Transcription and data preparation** | The ratio of preparation to collection time, high for qualitative data and non-trivial otherwise. Cleaning, coding and checking are work, not overhead. |
| **Supervisor reading time** | Actual turnaround including leave and term-time load, that the student cannot proceed during it, and that there is more than one draft. |
| **Final production** | Formatting, proofreading, references, figures, appendices, and the institution's submission and, where applicable, binding process. |
| **Examination logistics** | Examiner nomination, appointment approval, reading time and viva scheduling, where these fall inside the period being planned. |

For each, record the realistic elapsed time, who controls it, whether it can run in parallel, and the fallback if it fails. Then identify the critical path: the chain where any slippage moves the deadline. *Correct result: a dependency table with elapsed times, owners, parallelism and fallbacks, and a marked critical path.*

**Step 4. Convert every activity into a deliverable.**
A milestone is an artefact that exists or does not exist, on a date, that someone other than the student can inspect. "Working on the literature review" is an activity and can absorb any amount of time without producing evidence. "A 4,000-word literature chapter draft with a stated organising argument, sent to the supervisor" is a milestone: binary, checkable, and unmistakable in its absence. Apply the test to every line: could a third party tell, on the date, whether this happened? If not, rewrite it until they could. This single conversion does more to prevent silent drift than any amount of monitoring, because drift is only visible against something that was supposed to appear. *Correct result: every milestone is a noun with a form, a size and a recipient, not a verb.*

**Step 5. Place the buffer explicitly and defend it.**
Add contingency as a named block, not as padding distributed invisibly inside estimates. Padded estimates get spent silently; a named buffer has to be consciously drawn down, and the drawdown is itself a progress signal. Size it against the risk in the dependency map, and place most of it before the critical dependencies, because a buffer at the end is a buffer nobody reaches. Then state the rule for spending it: whose decision, and what gets recorded. *Correct result: a named contingency block with a size, a position and a spending rule.*

**Step 6. Distinguish real progress from reading.**
Reading is necessary, it is not progress, and it is the most comfortable place in a research degree to spend a year. Progress is artefacts moving through states: data collected, chapters drafted, analysis completed, decisions made and recorded. Monitor those, with a standing check that no period longer than the agreed interval passes without an artefact changing state. This is not a judgement about diligence; a student reading twelve hours a day may be making no progress at all, and noticing that is the supervisor's job, not theirs. *Correct result: a monitoring schedule expressed as artefact state changes with dates, not as meeting frequency.*

**Step 7. Watch for the four early signals of stalling.**
These appear well before the missed deadline, and each has a specific meaning.

- **Avoidance of contact.** Meetings postponed, emails answered slowly or not at all, updates arriving only when prompted. The mechanism is almost always shame about having nothing to show, and it is self-reinforcing.
- **Expanding scope.** New questions, an additional case, a further method, a wider literature. Frequently presented as ambition and frequently a way of postponing the point at which existing work must be finished and judged.
- **Perfectionist rewriting of an early chapter.** The introduction or the literature review, revised for the fourth time, always improving and never done. It substitutes visible effort for forward movement, and it is the hardest signal to name because the student is working hard and the work is getting better.
- **The disappearance of specific questions.** The most reliable signal and the least noticed. An engaged student asks narrow, technical, answerable questions. When the questions become general ("do you think this is going okay?") or stop, engagement with the problem has usually stopped too.

Any one of these is a conversation. Two together is an intervention. Record the observation, dated, with what was seen, and treat it as a signal about the project's state, never as a characterisation of the student (§12.5). *Correct result: a dated observation naming the signal and the evidence for it.*

**Step 8. Diagnose the block before intervening, because there are three kinds.**
Interventions fail mostly because the wrong one is applied.

| Block | How it presents | The intervention |
|---|---|---|
| **Technical skill** | Work stops at a specific operation: a statistical method, a software environment, an interview technique, a piece of equipment. The student can usually name it if asked directly. | Training, a specialist session, a worked example, or a different method. Fast to resolve once identified. |
| **Decision** | The student is circling. Multiple options, none chosen, work continuing in all directions. Often presents as more reading. | Force the decision to a date, list the options with their consequences, and say plainly that a defensible choice now beats a perfect choice never. The supervisor frequently has to take part of it. |
| **Motivation or confidence** | Avoidance, silence, the perfectionist rewrite, or a stated loss of belief in the project. Slowest to resolve and most often misdiagnosed as one of the other two. | Reduce the next step until it is small enough to complete, restore a short feedback cycle, and consider whether something non-academic is driving it. |

Diagnose by asking rather than inferring: "what is the specific thing you would do next if you sat down tomorrow?" separates the three quickly. A skill block produces a named operation, a decision block produces two options, a motivation block produces a pause. Never diagnose a motivation or confidence block from the schedule alone, and where the possible cause is a personal circumstance, route to institutional support rather than into the plan (§4). *Correct result: a named block type with its evidence, and a matched intervention.*

**Step 9. Have the scope conversation, and have it early.**
Scope reduction is the most common intervention a supervisor makes and the one most often left too late, because it feels like a loss and because it is easier to hope. Structure it. List every component of the project and mark each **core** (the thesis fails without it), **supporting** (it strengthens the argument), or **additional** (it is interesting and it is a separate project). Cost each in weeks against the available weeks from Step 2, cut from the bottom until the plan fits, and say plainly what was cut and why. Two framings make the conversation survivable and both are true: a smaller completed thesis passes and a larger incomplete one does not, and the cut material is the next paper rather than wasted work. Where a student resists, ask them to produce the schedule in which everything fits; that settles it faster than arguing about ambition. `RESEARCHER SIGN-OFF REQUIRED` (K5 §2.5, §2.8): the judgement that a component is not core is academic, it affects examinability, and it belongs to the supervisor. *Correct result: a scoped component list, marked and costed, with cuts recorded and reasoning stated.*

**Step 10. Plan the recovery when a milestone is missed, rather than moving it.**
Moving a missed milestone and changing nothing else is the most common response and it does not work, because the conditions that caused the miss are still present. Recovery has four parts: establish factually what happened, without attributing it; re-baseline, recalculating from today against the unchanged deadline and showing what no longer fits; change one condition, whether more supervision contact, reduced scope, a resolved skill block or a smaller next milestone; and set a checkpoint two to three weeks out rather than at the next scheduled meeting. State the arithmetic explicitly. A student who can see that the remaining chapters require a pace they have never achieved will engage with the scope conversation; one shown a plan that quietly assumes doubled output will not. *Correct result: a re-baselined plan, one named changed condition, and a checkpoint within three weeks.*

**Step 11. Ask honestly whether the project is deliverable, and name the alternatives.**
At some point the question stops being how to recover and becomes whether the project as defined can be delivered at all. Ask it deliberately rather than letting it arrive. The test is arithmetic first: what remains, at the pace this student has actually achieved over the last six months, against the weeks available. Where the answer is no, state the alternatives as a set rather than springing them one at a time, and name only those the regulations offer: a reduced scope that is examinable; a change of design to a feasible one, for example from primary collection to secondary analysis; an extension or suspension where the grounds exist; a change of degree or exit award where the structure provides one; or discontinuation. Be explicit about which are the institution's decisions rather than yours. `RESEARCHER SIGN-OFF REQUIRED` (K5 §2.5, §2.8). Deferring removes the options that exist only while time remains. *Correct result: a dated written statement of the arithmetic, the options available, and what each requires.*

**Step 12. Document progress, interventions and decisions as you go.**
Every review, extension request, appeal and change of supervisor requires a record, and reconstructing one afterwards is laborious and unconvincing. Keep a running log: milestones set, met and missed with dates; signals observed with evidence; interventions and their effect; scope changes with reasoning; and the student's own account where they gave one. Two disciplines make it defensible. Record observations rather than characterisations: "no draft received between March and June, three meetings postponed by the student" is a fact; "the student is disengaged" is a judgement a panel cannot check and a student may reasonably contest. And record what the supervisor did as well as what the student did, because the first question at any review is whether the support was there (K2 §5). *Correct result: a dated log of facts and actions a review panel could read without you present.*

## 8. Analytical framework

Planning runs backwards and monitoring runs forwards, and the two use the same objects.

    Deadline → Final production → Full draft → Chapters → Analysis
      → Data collection → Approvals and access → Design → Today

Built in that direction, the plan reveals its own infeasibility at the start rather than at the end. Built forwards, it always fits, because the last task simply lands wherever it lands.

Monitoring then runs the other way, on three tracks that must be read together:

    Artefacts:     what exists that did not exist last time
    Dependencies:  what is waiting on someone else, and for how long
    Signals:       contact, scope, rewriting, and the specificity of questions

A plan tracked only on artefacts misses the student who is stalling but not yet behind. A plan tracked only on signals produces anxiety without evidence. Read together, they distinguish the three situations that look identical from the outside: a student who is behind because a dependency is stuck, a student who is behind because the scope is too large, and a student who is behind because they have stopped working on it. Each needs a different response, and the intervention that fits one makes the others worse.

## 9. Output format

**1. Plan header.** Student, project, degree, binding deadline and its source, the date the plan was made, and the AI and data position.

**2. Available time.** Calendar period, minus known unavailability, expressed in working weeks. This number is usually smaller than expected and is the plan's most important line.

**3. Milestone schedule.**

| # | Milestone (deliverable) | Form and size | Recipient | Date | Depends on | Owner |
|---|---|---|---|---|---|---|

**4. Dependency register.** Each external dependency with realistic elapsed time, who controls it, parallelism, fallback, and whether it sits on the critical path.

**5. Contingency.** The named buffer, its size, its position, and the rule for spending it.

**6. Scope statement.** Components marked core, supporting or additional, with weeks costed against each, and anything already cut recorded with the reason.

**7. Monitoring and triggers.** The agreed checkpoint interval, and the pre-agreed signals that trigger a conversation.

**8. Progress log** (maintained, not written once). Dated entries: milestones met and missed, signals observed with evidence, interventions and effects, scope changes and reasoning.

**When the plan does not fit, the format must show that rather than absorb it** (K4 §1). Do not compress estimates until the schedule reaches the deadline. The correct output of a backwards plan that starts in the past is a plan that says so, with the shortfall stated in weeks, and a scope statement showing what would have to be cut to close it. That is a successful output, and producing it early is the whole point of the method.

## 10. Quality checks

Run before the plan is agreed, and at every review.

1. Is the binding deadline correct, and is its source stated?
2. Was the plan built backwards from the deadline?
3. Is available time expressed in working weeks rather than calendar months?
4. Is every milestone a deliverable that a third party could verify on the date?
5. Does every external dependency have a realistic elapsed time, an owner and a fallback?
6. Does the ethics estimate include at least one revision round and the committee's cycle?
7. Is the supervisor's own reading turnaround in the plan, honestly, and multiplied by the number of drafts?
8. Is final production, submission and, where applicable, examination logistics allowed for as its own block?
9. Is the contingency named, positioned and governed by a stated rule?
10. Does the plan fit, and if it does not, is the shortfall stated in weeks rather than absorbed into optimistic estimates?
11. Was the plan built with the student, and do they have a copy they can act on?
12. Are the monitoring triggers pre-agreed, so a later conversation is a trigger rather than a surprise?
13. Does the progress log record observations rather than characterisations of the student?
14. Does the log record what the supervisor did as well as what the student did?
15. Where a conversation about deliverability is due, has it been had, or is it being deferred?

## 11. Common failure modes

| Failure | How to recognise it | How to prevent it |
|---|---|---|
| **Forward planning** | The schedule always fits, and the last task lands near the deadline by coincidence | Build backwards from the deadline (Step 2) |
| **Activity milestones** | Plan lines beginning with "working on" or "continuing" | Every milestone is an inspectable artefact (Step 4) |
| **Calendar months as capacity** | Twelve months treated as twelve months of work | Convert to working weeks after subtractions (Step 2) |
| **The invisible ethics cycle** | Ethics allowed two weeks | Committee cycle plus one revision round (Step 3) |
| **Recruitment optimism** | A target rate assumed rather than observed | Plan the fallback before recruitment starts |
| **Supervisor as hidden critical path** | Drafts wait weeks and the plan assumes days | Put honest turnaround and leave into the plan |
| **Distributed padding** | Every estimate is slightly generous, and the buffer is invisible | One named contingency block with a spending rule (Step 5) |
| **Reading counted as progress** | Months pass, much is read, nothing exists | Monitor artefact state changes (Step 6) |
| **The moved milestone** | A missed date is simply rescheduled | Re-baseline and change one condition (Step 10) |
| **Late scope conversation** | Scope discussed only once the deadline is unreachable | Test scope against weeks at planning time and at each review (Step 9) |
| **Misdiagnosed block** | Training offered to a student who is actually stuck on a decision | Ask what they would do tomorrow (Step 8) |
| **Silence read as progress** | A quiet student assumed to be working | Treat disappearing specific questions as a signal (Step 7) |
| **Characterisation in the record** | Log entries describing the student rather than events | Record observations with evidence (Step 12) |
| **AI-generated schedule detached from institution** | A plausible timeline using milestones and periods this institution does not have | Regulations and dates supplied, never inferred (§12.4) |

## 12. AI guardrails

Skill-specific. The universal prohibitions in K4 apply in full and are not repeated.

1. **These skills assist an academic's judgement, they do not replace it.** Assessment, examination and progression decisions are the responsibility of the named academic and cannot be delegated to an AI system. The skill never assigns a mark, never produces feedback to be passed to a student unread, and never makes a progression or examination recommendation the academic has not independently reached. Institutional regulations and marking criteria take precedence over anything in this skill, and where a student's work may have integrity concerns, that is an institutional process and not a judgement this skill makes.

2. **Settle the policy and data position before the student's file is opened.** Progress material contains personal data and frequently sensitive circumstances. Institutional data protection and AI policy govern what may be processed and recorded. Where the position is unknown, do not upload identifying or personal circumstance material, and plan the project rather than the person.

3. **Never assess or infer a student's personal circumstances, health, capability, motivation or commitment.** Missed milestones are evidence about a project's state. They are not evidence about a person, and the most likely causes are frequently invisible in the schedule. Where a pattern suggests a non-academic cause, the output is a prompt to the supervisor to consider support routes, never a diagnosis and never a record about the student.

4. **Never invent institutional structures, milestones, periods or regulations.** Progression points, upgrade or transfer vivas, maximum registration periods, extension and suspension grounds, exit awards and examination timelines vary by country, institution and degree. Where they are not supplied, say they are required. A plan built on an assumed structure is confidently wrong in exactly the places that matter.

5. **Never characterise the student in the record.** Log observations with evidence and dates. "Three meetings postponed by the student between March and June, no draft received" is a fact a panel can check and a student can contest fairly. "The student is disengaged" is neither.

6. **Never recommend or imply a progression, termination, extension or exit outcome.** The skill lays out the arithmetic and the options the regulations provide. The decision, and the framing of it to the student, belong to the named academic and the institution (K5 §2.8).

7. **Never produce an optimistic plan to make a schedule fit.** Where a backwards plan does not reach the deadline, the output states the shortfall in weeks (K4 §1). Compressing estimates to close a gap manufactures a plan that will fail later, when fewer options remain.

8. **Never estimate a dependency's duration as though it were known when it is not.** Ethics cycles, access approvals, recruitment rates and examination timelines are institution- and setting-specific. Supply the structure and the questions; mark unsupplied durations as requiring local figures rather than filling them with plausible ones.

9. **Cap confidence on pace projections at the evidence available** (K3 §3.2, §3.6). A projection from two milestones is a hypothesis, not a forecast, and must be labelled as one. State the achieved pace it is based on, and the period it covers.

10. **Never let a plan substitute for a support referral.** Where the signals are consistent with a non-academic difficulty, say so and stop planning until the supervisor has decided which situation this is.

## 13. Best-practice principles

- **Plan backwards or do not plan.** Forward plans always fit, which is exactly why they are worthless.
- **Working weeks, not months.** The gap between the two is where most overruns live, and it is visible in ten minutes of subtraction.
- **A milestone is a noun.** If a third party cannot tell on the date whether it happened, it is not a milestone.
- **The dependencies that sink projects belong to other people.** Ethics committees, gatekeepers, data controllers, transcribers and supervisors are on the critical path, and none of them is managed by the student.
- **Assume ethics comes back once.** It is the single most reliable planning correction available, and it is free.
- **The supervisor is part of the schedule.** Honest reading turnaround, multiplied by the number of drafts, and placed in the plan where the student can see it.
- **Buffer must be named to survive.** Padding inside estimates is spent invisibly and always.
- **Reading is not progress, and a student can read for a year in good faith.** Noticing that is the supervisor's job.
- **The specificity of a student's questions is the best available progress instrument.** It changes before the output does.
- **Cut scope early, from a marked list, with the reasoning stated.** A smaller completed thesis passes; a larger incomplete one does not, and the cut material is the next paper.
- **A missed milestone that is only moved will be missed again.** Change a condition, or expect the same outcome.
- **Have the deliverability conversation while options still exist.** Every month of deferral removes one of them, and the last month has none left.
- **Write the record as you go, in facts.** It is what protects the student as much as the institution.

## 14. Worked example

Generic fictional scenario, academic, professional doctorate in education.

**INPUT**

A part-time professional doctorate student, employed full-time as a teacher, is eighteen months into a project on assessment practice across a group of schools. The registration end date is thirty months away. The supervisor has noticed that the last two meetings were rescheduled by the student, and that the most recent update said the literature review was "coming along" and asked no questions. The student has recently proposed adding a second phase of classroom observation to the planned interview study.

**PROCESS**

*Step 1.* Policy position recorded: no personal circumstance material processed externally; planning on the project. Candidate deadlines listed. Registration ends in thirty months; the institution's regulations require submission four months before that to allow the examination process inside the registration period. The binding date is therefore twenty-six months away, not thirty, and neither supervisor nor student had noticed.

*Step 2.* Backwards plan. Final production and submission: four weeks. Full draft, supervisor reading and student response: ten weeks. Discussion and conclusion: eight weeks. Analysis and findings: sixteen weeks. Fieldwork and transcription: to be built in Step 3. Available time is then converted to working weeks: a part-time student teaching full-time, around school terms, has roughly eighteen usable weeks a year, not fifty-two, giving about thirty-nine working weeks in twenty-six months. The backwards plan already consumes thirty-eight of them before fieldwork.

*Step 3.* Dependency map. Ethics has not been submitted, requires gatekeeper consent from four schools, and the committee meets six times a year: realistic elapsed time with one revision round is five months. Transcription of twenty interviews, which the student intends to do themselves, adds roughly six working weeks the plan had not accounted for. Supervisor reading turnaround is honestly three weeks in term time. Ethics is on the critical path and nothing else can start behind it.

*Steps 4 to 6.* Milestones rewritten as deliverables: "continue literature review" becomes "5,000-word literature chapter with an organising argument, to supervisor, by a named date". Six such milestones replace nine activity lines. A named eight-week contingency is placed immediately after ethics approval, where the risk concentrates. Monitoring is set on artefact state changes at six-week intervals.

*Step 7 and the judgement call.* Three of the four stalling signals are present: postponed contact, expanding scope (the proposed observation phase), and the disappearance of specific questions. The supervisor's instinct is to read the proposed second phase as ambition and the postponed meetings as workload, and to respond by encouraging the student and agreeing a new date. The signal set argues otherwise, and so does Step 3: the student is proposing to add a phase to a project whose ethics submission has not been written, eighteen months in.

But the supervisor does not diagnose from the pattern. Step 8 says ask. The question put is: what is the specific thing you would do if you sat down tomorrow? The answer is that the student does not know whether the study is interviews about assessment practice or an evaluation of one assessment approach, and has been reading around both for four months. That is a decision block, not motivation and not skill, and it also explains the proposed observation phase, which is a way of keeping both options open rather than choosing.

*Steps 9 and 10.* Two interventions matched to the diagnosis. The decision is forced to a date two weeks out, with the options written down against their consequences for design, ethics and feasibility, and the supervisor states plainly that they will make the call if the student does not. The scope conversation then runs against the working-weeks arithmetic: the observation phase is marked additional, costed at eleven working weeks the project does not have, and cut. The interview study is core. The literature review is bounded to the chosen question, converting four months of undirected reading into a chapter.

*Steps 11 and 12.* Deliverability is asked explicitly and answered: with the decision made, the scope cut and ethics submitted within six weeks, the project fits with the eight-week contingency intact. It does not fit if ethics slips past two cycles, so that trigger is recorded now with a fallback of two schools rather than four. The log records the deadline correction, the three signals with their evidence and dates, the diagnosis and the question that produced it, both interventions, the scope decision with its reasoning, the ethics trigger, and the supervisor's own three-week term-time turnaround, because a later review will reasonably ask.

**OUTPUT**

A revised twenty-six-month backwards plan with six deliverable milestones, a dependency register with ethics on the critical path, a named contingency block, a scope statement recording the cut phase and its eleven-week cost, a two-week decision deadline, and a recorded trigger with a fallback. `RESEARCHER SIGN-OFF REQUIRED` on the scope cut and on the deliverability judgement (K5 §2.5, §2.8).

## 15. Advanced usage

**Cohort-level monitoring.** Where a director of postgraduate study oversees many students, run the signal set rather than the schedule across the cohort, because schedules are individual and signals are comparable. Two or more signals present, or no artefact state change in a defined period, generates a check with the supervisor rather than the student. This catches the students whose supervisors are themselves overloaded, which is a common and invisible failure mode, and it is also where a cohort-level pattern (an ethics committee running slowly, a data source that has become inaccessible) becomes visible when no individual case would show it.

**Planning under a hard immovable constraint.** Where a visa, a funding end or a maximum registration period sets a date that cannot move at all, invert the method: fix the deliverable set to what fits, and treat scope as the only variable. Agree at the outset what the minimum examinable thesis is, and treat everything beyond it as optional from day one. This is uncomfortable and it is far kinder than discovering the constraint at month thirty.

**Using the student's own achieved pace.** After two or three milestones, replace generic estimates with this student's demonstrated rate, per K3 §3.2. It is the only estimate grounded in evidence, and it is usually the conversation that ends the argument about whether the remaining plan is realistic. Label it as a projection from a small number of observations, not a forecast.

**Where the standard approach does not fit.** For projects dependent on seasonal or event-bound data collection, an academic year, a harvest, a clinical cycle, an election, a missed window costs a full period rather than a few weeks, so the buffer goes before the window and the fallback must be a different data source rather than a later date. For team-based or lab-based doctorates, the student's plan is a dependent of the group's, and the group's schedule belongs in the dependency register. For students supervised at a distance or across institutions, add the co-ordination cost explicitly: joint decisions take longer than either supervisor assumes, and unagreed advice is a common source of stalling.

## 16. Skill chain

**Recommended previous skills:**
- **15.04 Research Proposal Writing.** Hands over the scope, objectives and intended method that the plan is built around, and against which scope growth is later measured.
- **15.09 Ethics Clearance Application.** Determines the realistic ethics dependency, which is usually the critical path at the start of a project.

**Recommended next skills:**
- **15.20 Supervision Feedback and Development.** Takes the milestone that has arrived and produces the feedback on it, prioritised to the time before the next milestone.
- **15.21 Thesis Examination and Marking.** Where the plan reaches submission, and where a corrections outcome generates a new short plan against the regulations' period.
- **15.18 Thesis Architecture and Chapter Coherence.** Where a scope cut requires the thesis structure to be rebuilt around what remains.

**Runs well alongside:**
- **15.22 Research Rigour Audit**, where the question is whether a schedule problem is actually a design problem.
- **15.19 Examiner-Proofing and Defence Strategy**, which is scheduled as a deliverable milestone in the final months rather than left to the last weeks.

---
A Yazi Supplied Skill and resource.