How to Ask for a Letter of Recommendation

Almost every disappointing reference has the same two causes: the referee was asked too late, and they were given nothing to work with. Both are entirely within your control, and fixing them is probably the highest-return hour in the whole application.

Understanding what a reference is for makes the rest obvious. Assessors do not read references for praise — praise is universal and therefore carries no information. They read for two things a candidate cannot supply about themselves: comparison against a peer group, and corroboration of specific claims. A letter that says you are excellent adds nothing. A letter that says you were among the strongest in a defined group, and describes the project you claimed, is decisive.

Choosing referees

Rank candidates by this order, not by seniority:

Someone who supervised your actual work A project supervisor, a lab PI, a dissertation adviser, a manager who gave you a real task. They can describe how you work, not just what you scored.
Someone who taught you in a small setting A seminar leader who has read your writing and watched you argue beats a lecturer from a 300-person course who knows you as a grade.
Someone whose role matches what is being assessed For a scholarship with a leadership criterion, the person who watched you lead something is more useful than a professor who did not.
Someone senior who barely knows you Last. A famous name producing four vague sentences is actively harmful — it signals to the reader that nobody who knew you well was willing.

Ask whether they can write a strong letter, not whether they will write one

Phrase the request so a no is easy: “Would you be able to write me a strong letter of recommendation for X?” A referee who has reservations, or who simply does not remember you well enough, is given a graceful exit. Take a hesitant answer as a no and thank them. A lukewarm letter is much worse than a different referee.

How much notice

Four to six weeks is the working minimum for a considered academic letter, and more is better if your deadline falls at a bad time in the academic calendar. December and January deadlines collide with end-of-term marking; September requests land during vacation. If you must ask with less notice, say so explicitly and apologise for it rather than hiding the date at the bottom of the email.

Ask in person or by video where you reasonably can, and follow up in writing the same day with the brief below. The in-person ask gets a real answer; the written brief gets a real letter.

The brief: what to send, in one email

This is the part almost nobody does, and it is the difference between a generic letter and a specific one. Send a single message with everything, clearly organised:

1. The logistics, at the top

Every programme or scheme, with its deadline and time zone, as a short list.
How the letter is submitted — portal link, upload, email, sealed envelope — and whether the system emails them a link and when.
Any form or rating grid they must complete alongside the letter.
Whether a specific letterhead, signature or institutional email is required.

2. The criteria

Attach or link the scheme’s published assessment criteria. Most referees have never seen them. A referee who knows that “commitment to improving the lives of others” is a scored criterion will write to it; one who does not will write a standard academic reference that scores zero on it.

3. The reminder pack

Your CV.
Your draft personal statement or research proposal — so the letter corroborates rather than contradicts it.
A short note on what you did with them specifically: the module and year, the project title, what you produced, your grade if it helps, and one or two things that went wrong and how you handled them. This last item is what turns a letter from adequate to memorable.

4. The optional draft points

Not a draft letter — that is presumptuous and, in some systems, prohibited. Instead: “If it helps, the points I think are most relevant for this scheme are (a), (b), (c).” Busy referees frequently appreciate this, and none is obliged to use it.

If a referee asks you to write it yourself

This happens, particularly with overloaded academics and in some professional settings. It puts you in an awkward position, and the honest way through it is to write a factual account of your work with them in the third person, in plain prose, and send it as material for them to use, adapt or discard. Do not sign anything on their behalf, and never submit a letter they have not read and approved. Some application systems require the referee to submit directly, which resolves the problem for you.

Waiving your right to see the letter

US systems commonly ask whether you waive access to your recommendations. Admissions readers consider unwaived letters less candid and therefore less informative, so waiving is the conventional and generally stronger choice. The corollary is that you must trust the referee, which is another reason to choose people who know your work rather than people with impressive titles.

Following up without nagging

Two weeks before the deadline A short, warm message: “Just checking the portal link reached you — no rush, the deadline is the 15th.”
Five days before If it is still outstanding, message again with the exact deadline including time zone. Most systems show you the submission status.
Two days before If still outstanding, contact them directly by whatever channel actually reaches them, and simultaneously check whether the scheme has a late-reference policy. Some accept references shortly after the deadline; many do not.
Have a reserve referee identified from the start Not approached and pressured, but identified, so that a late failure is recoverable.

Missing references are one of the most common reasons an otherwise complete application is marked incomplete and never assessed. Confirming submission is your responsibility, not theirs.

Afterwards

Thank them properly, in writing, soon after submission.
Tell them the outcome, whichever way it goes. Referees rarely find out, and the ones who do are far more willing next time.
Keep a record of who wrote for what and when, in your document system. You will need references again for a job, a graduate programme or a professional registration, and asking the same person for the fourth time is much easier if you have not lost track.
Stay in occasional contact. A referee who has heard nothing for three years writes a much weaker letter than one who received two updates.

What a strong letter contains

So you can judge whether you have chosen well, these are the elements assessors weight:

A defined comparison group — “of the roughly forty final-year students I have supervised…”
Specific corroborated instances matching what you claimed in your own statement.
Assessment of independence — whether your work was directed or self-generated.
Direct engagement with the scheme’s criteria where they exist.
An honest limitation, handled fairly. Counter-intuitively, a letter that notes a real weakness raises the credibility of everything else in it.

You cannot dictate any of this. You can make all of it possible by asking the right person, early, with the criteria and the material in front of them.

Next: how to write a strong personal statement and how to organise academic documents.