How do you manage a PhD like a project? A step-by-step method

What are the steps to manage a PhD like a project?

A doctorate has everything a project has: a start date, a funding end, formal gates, deliverables, risks and a small team of advisers. What it usually lacks is the routine that makes projects steerable. These seven steps add that routine without turning research into paperwork.

• 1. Copy every formal gate from your regulations: progression or confirmation review, annual reviews, notice to submit, submission, examination.
• 2. Add your outputs (papers, chapters, datasets) and the enablers behind them (ethics approval, data plan, validated methods, field campaigns).
• 3. Forecast every open milestone once a month and record the slip in days against the original plan.
• 4. List the work left as packages of four to twenty weeks, each with optimistic, most likely and pessimistic estimates.
• 5. Calculate your runway buffer: available working weeks minus the estimated weeks of work.
• 6. Agree a one-page charter with your supervisors and run every meeting on an agenda, decisions and a same-day summary.
• 7. Review weekly for ten minutes, monthly for the time log and backups, and quarterly for the runway.

Why do PhD plans fail in year two?

Most doctoral plans are drawn early, in good faith, and never compared with the calendar again. Year one feels open, so drifting by a few weeks seems harmless. Then year two brings fieldwork or data collection, a first paper, teaching, a formal review and a visible funding clock at the same time.

The usual pattern is not laziness. It is three quiet drifts: enablers take longer than planned (an ethics amendment, a protocol that fails, access to a site), writing time gets squeezed by everything that has a deadline, and nobody has a number that says whether the remaining work still fits. A plan that is never tested against the weeks left becomes a hope.

The fix is simple: keep the original plan, add a forecast, and measure the gap. The gap shows how good your estimates are and when to raise a decision with your supervisors.

Time is the other blind spot. Teaching, marking, training and admin all arrive on other people's deadlines, and they are rarely counted. A monthly time log takes five minutes: hours in the lab or field, reading, writing, teaching and admin. In the worked example used in this article, a first review year came to 1,532 hours, of which 259 were writing (17%) and 82 were teaching (5%). Writing stayed below 20 hours a month for eight months and the literature review slipped in exactly those months. Once two fixed writing blocks a week went into the calendar, writing rose to between 30 and 46 hours a month. Numbers like these turn a vague feeling of being busy into a decision you can discuss.

How do you calculate a PhD runway? A worked example

A runway compares the working weeks you have before your submission target with the weeks of work left. Estimate each work package with three numbers and combine them with the PERT formula, a technique from large engineering projects: expected weeks = (optimistic + 4 x most likely + pessimistic) / 6.

Here is an illustrative example for a researcher in month 14 of a four-year funded PhD, with a submission target 138.7 weeks away, six weeks of leave a year, and 85% of working weeks available for research work packages (the rest goes to teaching, training and admin).

• Estimate in working weeks, not calendar weeks, and keep packages between four and twenty weeks so each estimate stays meaningful.
• Use your own past slip to calibrate: if done milestones ran 11 days late on average, your most likely estimates are optimistic.
• A negative buffer is not a failure. It is the moment to choose between narrowing scope, changing the order of work or moving the target date, with your supervisors, while all three options are still open.
• Recalculate quarterly and before every annual review. The buffer moves as you learn more about your methods and data.

What should a PhD supervisor agreement include?

Supervision is the working relationship that matters most, and it often runs on unspoken assumptions. A one-page charter, filled in together in the first month and revisited after each annual review, records them. It never replaces your university's code of practice, which sets minimums and complaint routes.

Keep it short enough that people reread it. The meeting habit matters as much as the document: send an agenda of no more than three decisions 48 hours ahead, keep the meeting on those decisions, and send a summary email the same day listing decisions and actions with owners and dates.

• Meeting frequency, format and length
• Feedback turnaround for papers, chapters and short sections
• Who leads on what when there are co-supervisors
• Authorship, agreed in writing before drafting
• Leave and absence, on both sides
• Teaching and other work limits
• Research integrity and AI use
• What happens when things go wrong

How do you keep reading, notebooks and data under control?

Over three to five years, memory does not scale. Give reading a queue: every paper enters with a priority, gets read with a question, gets five lines of notes in your own words, and gets linked to the paper or chapter where it will be used. Watch two numbers: queue size and the share of read papers with notes.

Write notebook entries so someone competent in your field could repeat the work from the page: date, aim, protocol version, deviations, sample IDs, settings, observations, results and where the raw data lives. Index entries weekly so you can find them in year four.

For data, write a plan before you collect anything and have your library's research data team review it. The FAIR principles, published by Wilkinson and colleagues in Scientific Data in 2016, ask that data be findable, accessible, interoperable and reusable. Keep three copies of each dataset on two kinds of storage with one away from your desk, and test a restore each month. Funder rules differ: the NIH Data Management and Sharing Policy, for example, has required a plan with applications that generate scientific data since 25 January 2023.

How do you choose a journal and manage a publication pipeline?

Put every paper you are involved in on one list with its stage, from idea to published. Two gates prevent most avoidable trouble. First, agree authorship in writing before anyone drafts. The CRediT taxonomy, now the ANSI/NISO Z39.104-2022 standard, names 14 contributor roles that make the conversation concrete, and the ICMJE criteria used by many journals require all four of a substantial contribution, drafting or critical revision, final approval and accountability.

Second, choose the journal before shaping the paper, and check that it is trustworthy. Think. Check. Submit. is a cross-sector initiative with a checklist for this. Score two or three candidates on scope fit, readers, speed, open access route, cost and data policy with your co-authors. Check your funder's open access rules: UKRI, for example, requires immediate open access with a CC BY licence for research articles submitted from 1 April 2022.

How do you prepare for a PhD annual review?

Progression, confirmation and annual reviews go by different names and rules. The Australian National University, for example, sets a thesis proposal review at nine to twelve months and annual reports at 12, 24 and 36 months for PhD candidates. Panels everywhere ask four things: what you did, what slipped and why, what comes next, and whether the plan is credible.

Start six weeks out. Pull numbers from your registers rather than memory: milestones with slip, outputs and stage, time use, the runway and the top three risks. Explain each slip with a cause and a change. Send the draft to your supervisors three weeks ahead. Afterwards, turn every panel request into a dated milestone within a month.

If your circumstances have affected progress, through illness, caring, disability or anything else, say so early and ask your graduate school about adjustments, interruptions or extensions. Rules and their effect on funding and visas differ by university, funder and country, so get the answer in writing.

Common mistakes when managing a PhD

• Planning only the next month instead of the whole programme.
• Moving planned dates instead of recording forecasts, so slip disappears.
• Leaving supervision arrangements unspoken until something goes wrong.
• Reading without a queue or notes, then rereading everything for each paper.
• Keeping the only copy of a dataset on a laptop.
• Agreeing authorship after the paper is written.
• Accepting teaching in field or writing months without checking the time log.
• Starting thesis assembly and the career plan in the final months.
• Letting an AI tool produce text you submit as your own.

Where do tools fit, including AI?

A spreadsheet, a calendar and a few templates are enough. The PhD Research Manager kit from Bodhih Training packages them: a workbook with a milestone map, registers, a time log, a runway calculator and a dashboard, plus a charter, forms and a dated country sheet.

AI assistants can help you plan and check: building search strings, turning meeting notes into an agenda, organising reviewer comments into a table, or asking practice viva questions. They must not write work you submit. Publication ethics bodies such as COPE state that AI tools cannot be authors and that use should be disclosed. Your university's policy decides what is allowed. To check your own judgement on these questions, the AssessAll assessment Second Look: Research Integrity in Practice is a useful baseline, and Jobulary can turn what you learn into an individual development plan.