None of it is enjoyed. All of it is meant to be believed.

Recommendation letter writing for a recommender who will sign sentences they did not write
and honestly would rather not oversell.

Nothing in a recommendation letter is enjoyed. Every sentence is written to produce a belief in the reader, and that is a different craft from writing to be read.

Every claim traceable to something the recommender personally witnessed. Claims stated narrowly and firmly, so an honest letter is not inflated. More weight on the comparable claim than on the adjective. One line that makes the letter checkable. And a list of what they are signing, handed over before it goes out.

WHAT THE WORK IS 01 / 04

A lukewarm letter is worse than no letter, because it actively says something.

  • 01 The claimsOnly what was witnessed.
  • 02 The comparisonMeasured against what?
  • 03 The signingWhat they answer for.
Written to be believed, not read Specific rather than emphatic Narrow claims, firmly stated Recommender signs knowingly

The Writely difference

Every other page here is read as writing. This one is read as evidence, and that is why a beautiful letter is a bad sign.

Nothing here needs to be admired. It needs to be believed, and a reader with no time to spare is not going to admire it first.

Build the letter out of what was personally witnessed

The line that governs everything, and it is firmer than anything on the ghostwriting pages in this set because here the recommender's name is on the letter and they will be asked about it afterwards. So the claim list is assembled first and everything else is argued against it: which project, which cohort, which year, what the person did in a room, how they behaved on the week it went wrong. What nobody witnessed stays out even when leaving it out makes the letter shorter, because a claim that cannot be sourced to a moment is a claim the recommender cannot stand behind if anybody asks.

Treat specificity as the mechanism, not the decoration

On every other page in this set the specific detail is a matter of craft. Here it is the entire thing, because it is the only part of a recommendation a reader can act on or check. She is hard-working is worth nothing and has been read several hundred times in a single morning by somebody with two hundred letters and no time. Which six she was the best of, which board he has recommended candidates to, what she carried when the supervisor left: those can be believed because they are the kind of claim somebody could confirm. And the reader is thinking about credibility the whole way through, because they know the name on the letter was attached to sentences somebody else composed.

Make a modest letter do real work instead of inflating it

The hardest case here is not a strong candidate. It is a genuinely good one whose recommender does not want to be effusive, believes saying more would be dishonest, and is uncomfortable being asked at all. Which is common and is not solvable by enthusiasm. Inflating an honest letter converts it into a letter nobody believes, and it also puts the recommender in a position they will not describe afterwards. What works is narrower claims stated firmly: fewer adjectives, more particulars, and the one thing this person did that the reader will still remember two thirds of the way down the page.

Hand the recommender the list of what they are signing

The stranger triangle at the centre of this document, and the part most services skip. The candidate is not the client and cannot be, because they have no standing over what is claimed about them. The recommender owns the claims but did not choose the wording. The writer owns the sentences and is answerable to neither. So the drafting is done to be signed, and what comes with it is a plain list of every claim in the finished letter, each one traced to the moment it came from, so the person whose name goes on it knows precisely what they are agreeing to. A recommender who has read that list will almost never ask to change anything.

Find out what the recommender can actually stand behind

Eight things a recommendation letter has to settle, and the one everybody skips is what the recommender is signing.

None of this is about making the candidate look better than they are. It is about making the reader able to believe the recommender.

Write for the reader's belief rather than for the reader's pleasure

The difference that governs the whole document and that every piece of advice about this genre gets wrong. A description is judged on how well it describes; a caption on how it reads; an opinion piece on whether it moves anybody. A recommendation letter has only one question behind it, which is whether to believe it, and a reader who has two hundred of them is not going to admire a good sentence. So the writing is not there to be enjoyed, and any sentence whose job is to sound good rather than to carry a claim is occupying space that a real claim needed.

Replace every adjective with something that can be checked

The single most useful edit on any letter and the one that most often turns a weak one into a strong one. Diligent becomes the week the deadline moved and what she did about it. Reliable becomes the number of times she was left holding a piece of work and the number of times it needed rescuing. Outstanding becomes the named thing she did that nobody else on the team was doing. None of this is embellishment, because the claims are all already true; it is the difference between a claim that can be believed and one that cannot, and the reader has been trained by two hundred letters to notice.

Give the comparable claim more weight than the emphatic one

Because a comparison is checkable and an adjective is not. Best of the six students I have supervised since 2019 carries the recommender's own standard inside it, which is the part that makes it persuasive, and it tells the reader exactly what they are being measured against. Enthusiastic and outstanding tell the reader nothing except that the writer had a good weekend. Where a real comparator exists it should carry the paragraph, and where one does not exist it must not be manufactured, because best of six is only available to somebody who supervised six and a fabricated one is a claim the recommender cannot answer for.

Include one line that makes the letter findable

Which sounds trivial and is the most practically valuable sentence on the page in this market. An institution, a department, a project, a cohort, a year, a title. It does nothing whatever for the reader's impression and it decides whether the letter is usable at all, because a reader who cannot verify anything quickly treats the whole document as unverified. It is also the line most often left out, almost always because the recommender assumes everybody already knows who they are, which nobody does.

Do not repeat what the reader already has

The letter sits between a curriculum vitae and an application form, both of which the reader is holding, and a paragraph restating the candidate's grades tells them nothing they have not read thirty seconds earlier. What the letter alone can carry is what a form cannot hold: how this person works, how they behaved when something went wrong, what they were like to work beside, and what they did when nobody was checking. So the repetition goes and the judgement takes its place, which is the whole of what the recommender is uniquely placed to supply.

Say something about weaknesses, or do not mention them at all

Neither half of that is popular and both are worse than they look. A letter with no weakness in it reads as a letter that was not read properly, because everybody knows what a real one contains. A letter that turns a small weakness into a strength is the same error in a different disguise, and readers are now trained to spot it. What works is one concrete thing, stated once, in proportion, with what came of it, and then no further hedging. A reader who sees one honest limit in a page of praise believes the rest of the page.

Give the recommender a way out that is not a refusal

Practical and frequently the reason a letter ends up lukewarm. Somebody who does not think they can write a strong letter has three honest options and only one of them is usually offered: write it anyway, write nothing, or write something carefully balanced that nobody will act on. The third is where nearly all of them land. What helps is agreeing in advance what the document will say if the honest answer is that the candidate is capable but not distinguished, which is a perfectly good letter and a common one, and considerably better than the hedged version everybody writes instead.

What this service cannot do

Four limits. A recommendation letter does not claim anything the recommender has not personally witnessed, and that is firmer than anything on the ghostwriting pages here because the name on the letter is theirs. It does not manufacture comparators, because best of six is only available to somebody who supervised six. It does not do the candidate’s own work, so if the curriculum vitae and the personal statement are weak, that is a different conversation and this service will say so. And it does not promise an outcome, which belongs to somebody else’s decision.

Read the recommender’s existing writing before you start. One page of anything they have written is worth more about how firmly they are willing to speak than any question you can ask them about it.

A better brief

Send the moments, not the praise.

The brief that stalls a recommendation letter is the one that sends a list of adjectives about the candidate, because those are the words the letter is trying to replace. Send instead: the specific moments. The project, the deadline, the thing that went wrong, the person the recommender turned down in order to recommend this one. Along with the recommender’s own position, how well they know the candidate, and anything they have already said no to. That last item saves more trouble than any other part of the brief.

  • The moments, not the adjectives.
  • How well they know the candidate.
  • What they will not claim.
  • What they have already refused.
  • The deadline and the form.
  • The candidate’s own material.
Share your brief
LETTER CHECK Before drafting

Six things before the first sentence.

Claims traced
Comparator real
Adjectives cut
Position named
Weakness stated
Signer briefed

How the work is done

From what the recommender personally witnessed, to what they are willing to sign.

Claims, comparison, draft, signing. The first step removes anything unwitnessed and the last one hands the finished list to the person whose name goes on the letter.

01 / THE CLAIMS

Find everything the recommender personally witnessed, and drop the rest

First, and it is mostly listening rather than writing. Which projects, which cohort, which year, what this person actually did in a room, how they behaved on the week it went wrong, what they did when nobody was checking. Everything asserted but unwitnessed comes out at this stage rather than later, because that is the moment it is quickest to lose. The output is a claim list with a moment attached to each entry, and where the list is thin the honest conclusion is that this recommender cannot write a strong letter, which is useful to know before anybody drafts anything.

02 / THE COMPARISON

Work out what this person is being measured against, and by whom

Then the claim that carries the most weight, because it is the only one in the genre that carries the recommender’s own standard inside it. Which six they supervised. Which board they have put candidates in front of. Which cohort, and what happened to the others. A comparison tells the reader not only what the candidate is but what good looks like to the person writing, which is twice the information and is checkable. Where a real comparator exists it takes the paragraph, and where one does not it is not manufactured.

03 / THE DRAFT

Write it to be believed, with every claim traceable

Now the drafting, which is straightforward once the claim list exists and is bounded by it. Adjectives out, particulars in. The repetition of the curriculum vitae out and the judgement in, since that is the only thing the recommender is placed to supply and the reader already has the rest. One weakness stated once and in proportion. And one line making the letter findable. Then it is read against the claim list line by line, and anything that cannot be traced comes out, however much the paragraph is improved by it.

04 / THE SIGNING

Hand the recommender the list of what they are signing, and a way to decline any of it

Finally the part that makes the document hold up afterwards. The finished letter goes out with every claim listed and traced to the moment it came from, so the person whose name is on it knows exactly what they are agreeing to and can strike any line before signing. That is usually the step that turns a hesitant recommender into a willing one, because the risk stops being theoretical. What is delivered is the letter, the claim list, and a short note on anything the recommender may want to soften.

What actually gets handed over

Five documents. The claim list is what makes the letter defensible after it has been signed.

A recommendation letter is short and the work behind it is not. Four of these five are inputs, and only one is the document anybody asked for.

The claim list

Every claim the recommender can personally vouch for, each with the moment attached: which project, which cohort, which year, what this person did in a room, how they behaved when something went wrong, what they did when nobody was checking. This is the document the letter is written from and the one it is checked against before it goes anywhere. Where the list is thin the honest conclusion is that this recommender cannot write a strong letter, which is far better to establish before a draft exists than after one has been promised and cannot be written honestly.

The comparator note

What this person is actually being measured against, and by whom. Which six the recommender supervised, which board they have put candidates in front of, which cohort and what happened to the rest of it. A comparison is the only claim in the genre that carries the recommender’s own standard inside it, so it tells the reader both what the candidate is and what good looks like to the person writing. It is also the claim most often invented, which is why the note records whether a real one exists rather than assuming that one does.

The draft, checked against the list

Adjectives out and particulars in. The repetition of the curriculum vitae out and the judgement in, since that is the only thing the recommender is placed to supply and the reader is already holding the rest. One weakness stated once and in proportion. One line making the letter findable. Then every sentence is read back against the claim list and anything that cannot be traced comes out, however much the paragraph improves for having it. This is the only page in the set where the deliverable is constrained on purpose by an input document.

The signing note

The finished letter with every claim listed and traced, so the person whose name goes on it knows exactly what they are agreeing to and can strike any line before signing. This is usually the step that turns a hesitant recommender into a willing one, because the risk stops being theoretical and becomes a list they have read. It is also the honest answer to the awkwardness of being asked, since being asked to approve specific sentences is a different request from being asked to write something.

The exclusions note

Everything the recommender refused, and what was dropped as unwitnessed rather than as unflattering. The distinction matters: a letter improved by removing an adjective is a better letter, while a letter improved by removing an inconvenient fact is a dishonest one, and nobody reading the finished document could tell which has happened. This note is what lets anybody tell afterwards, and it is also the point at which the candidate’s own material gets looked at properly rather than assumed.

Four of these five documents exist so the letter can be defended afterwards. Nobody asks for them and everybody needs them when a reader rings the recommender to check.

What a recommendation letter has to settle

Six things, and the letter nobody checked has settled none of them.

Each can be settled before a word is drafted, and the first one settles whether this recommender can write this letter at all.

  • Every claim personally witnessed.
  • A real comparator, or its absence.
  • Adjectives replaced with particulars.
  • The repeat of the CV removed.
  • One weakness, stated in proportion.
  • The signer shown the claim list.
THE LETTER Six things

Six things a letter must settle.

Claims witnessed
Comparator real
Particulars in
CV repeat out
Weakness stated
Signer briefed

Where this is actually written about

All eight are new to this set, and one of them is about the opposite of what this document is for.

Twenty-two candidates were tried across three rounds and eight answered, all new. They cover the form itself, the form it is constantly confused with, the commercial practice that behaves in exactly the opposite way, the closest structural sibling this genre has, and the two documents the letter sits between. Two candidates were rejected and both losses are recorded here rather than quietly dropped, including the most useful contrast in the whole set, where the substitute would have been weaker for the purpose than nothing at all. These external resources are informative references, not Writely service pages.

THE FORM ITSELF

Wikipedia: Reference Letter

The starting point, because the two forms are routinely confused and the confusion produces a weaker document every time. Its account of what a reference is for, of who is asked to provide one and what they are being asked to attest to, and of how the request is normally put is the background to the brief on this page. It is also the source of the distinction that matters most here: a reference supplies information the reader cannot get elsewhere, which is why a vague one is worse than useless rather than merely weak.

Open resource
THE FORM PEOPLE MEAN WHEN THEY SAY THE OTHER ONE

Wikipedia: Letter of Recommendation

The distinction is real and worth being pedantic about, because the two are asked for in different situations and read with different expectations. Where a reference supplies verifiable facts, a recommendation carries an endorsement and is judged on the strength of the advocate’s opinion. Its account of how each is used, of who is normally asked, and of where the conventions differ across fields makes the question worth asking before a draft begins rather than after one has been written to the wrong shape.

Open resource
THE OPPOSITE OF WHAT THIS DOCUMENT IS FOR

Wikipedia: Endorsement

The clearest contrast available anywhere, and the reason this page can argue its central claim quickly. An endorsement exists to persuade and is written for effect; it is judged on whether it moves anybody, and its whole design is to be enjoyed as rhetoric. A reference letter exists to be believed and is written as evidence, which is why the persuasive techniques that work in one actively damage the other. Read side by side the two behaviours are easy to tell apart, which is exactly the judgement a recommender needs to make.

Open resource
THE CLOSEST STRUCTURAL SIBLING

Wikipedia: Peer Review

The genre this one borrows its hardest problem from, because a referee report has the same shape: an assessment by somebody of comparable standing who did not supervise the work and cannot ask any questions about it. Its account of how reviewers are chosen, of what they are and are not able to see, and of why a report from somebody who has never met the author is trusted less than one from an equal is a direct parallel, and it is where the insistence on checkable specifics on this page comes from.

Open resource
WHAT THE READER IS ALREADY HOLDING

Wikipedia: Curriculum Vitae

Included because the strongest single edit on most letters is removing everything the reader has already got. Its account of what a curriculum vitae carries, of the conventions about length and format across fields, and of which parts are treated as claims rather than records tells you what is redundant beside it. What the letter alone can supply is how this person works and what they were like to work beside, and repeating a qualification list in a letter wastes the only page that could have carried that.

Open resource
WHERE THE LETTER SITS AND HOW IT IS READ

Wikipedia: Job Application

The practical context, and the reason this page insists the letter is not read as writing. Its account of the stages an application passes through, of how much time each decision gets, and of what is being weighed at each point explains why a reader meets a large number of letters in a single sitting and why anything that asks for attention is simply skipped. It is also where the point about a letter being unused unless it arrives finds its context, since a document that arrives late has not been written badly.

Open resource
WHAT IS ACTUALLY BEING ASSERTED

Wikipedia: Academic Degree

Behind every letter in academic settings is a claim about a qualification, and this is the reference for what that claim rests on. Its account of how qualifications are conferred and recorded, of the difference between a qualification and a certificate of attendance, and of who is competent to attest to one explains why the recommender’s own position is the single most valuable sentence in the document. It is what makes a line naming an institution, a department and a year worth more than a paragraph describing qualities.

Open resource
THE OTHER KIND OF PRAISE

Wikipedia: Testimonial

The third form people confuse this with, and it is useful precisely because it is the version of praise with no obligation attached to it. Its account of how testimonials are gathered and used, and of why they carry little weight where a reference carries a great deal, makes the contrast plain: nothing in a testimonial has been checked and nobody is standing behind it. A recommendation letter is the same kind of praise with the checking and the standing attached, which is the whole reason the specifics matter so much.

Open resource

A letter nobody can verify is a letter nobody can use

Nigeria at the centre. Credentials are checked here, and rarely at a glance.

Writely is rooted in Nigeria and works remotely with academics, applicants, professionals and institutions. Wherever the letter is going, the rule holds: the sentence that makes it findable is worth more than the paragraph that makes it sound warm.

Remote recommendation letter writing for any market

Recommendation letter writing across Nigeria

Explore recommendation letter writing for applicants, academics and professionals in Nigerian cities, communities and states. Three things are genuinely different here and all three make the checkable line more valuable than it would be elsewhere. The first is that credentials carry real weight and are checked more often, by people with less time and fewer ways of checking, so a letter that cannot be verified quickly is treated as unverified in its entirety. The second is that references here are often requested in markets where the recommender is known only slightly, which is a structural problem for the letter to solve rather than to hide, and the solvable part is naming the one context in which the recommender did witness something. The third is that a great many applications here are decided on the strength of a small number of documents read quickly, so anything that asks the reader for attention is simply passed over, which is the argument for writing for belief rather than for effect.

Lagos Abuja Port Harcourt Kano Ibadan Enugu Benin City Uyo Kaduna Calabar Jos Abeokuta Awka Akure Warri Owerri Zaria Sokoto Bauchi Maiduguri

Nigeria states and the Federal Capital Territory

Writely’s support is available remotely across Nigeria, including:

  • Abia
  • Adamawa
  • Akwa Ibom
  • Anambra
  • Bauchi
  • Bayelsa
  • Benue
  • Borno
  • Cross River
  • Delta
  • Ebonyi
  • Edo
  • Ekiti
  • Enugu
  • Gombe
  • Imo
  • Jigawa
  • Kaduna
  • Kano
  • Katsina
  • Kebbi
  • Kogi
  • Kwara
  • Lagos
  • Nasarawa
  • Niger
  • Ogun
  • Ondo
  • Osun
  • Oyo
  • Plateau
  • Rivers
  • Sokoto
  • Taraba
  • Yobe
  • Zamfara
  • Federal Capital Territory

Remote support across Africa

Country-level discovery

Writely works with applicants and institutions across Africa, including in:

Nigeria Algeria Angola Benin Botswana Burkina Faso Burundi Cabo Verde Cameroon Central African Republic Chad Comoros Côte d’Ivory Democratic Republic of the Congo Djibouti Egypt Equatorial Guinea Eritrea Eswatini Ethiopia Gabon Gambia Ghana Guinea Guinea-Bissau Kenya Lesotho Liberia Libya Madagascar Malawi Mali Mauritania Mauritius Morocco Mozambique Namibia Niger Rwanda São Tomé and Príncipe Senegal Seychelles Sierra Leone Somalia South Africa South Sudan Sudan Tanzania Togo Tunisia Uganda Western Sahara Zambia Zimbabwe

Questions, answered

Send the moments, not the praise.

If your question is not covered here, send the essentials in a chat. What the recommender actually witnessed says more than any list of adjectives.

Ask your question
What is recommendation letter writing?
Writing a reference or recommendation letter from what the recommender personally witnessed, so that a reader can believe it. It is not the same craft as writing to be read: nothing in one of these documents is enjoyed, and every sentence exists to produce a belief. The work assembles a claim list with the moment attached to each entry, finds a real comparator if one exists, replaces adjectives with particulars, removes anything that repeats the curriculum vitae, and hands the recommender the finished list of what they are signing.
What does a recommender need to give you?
Moments rather than adjectives, and that is really the whole brief. Which projects, which cohort, which year, what this person did in a room, how they behaved when something went wrong, what they did when nobody was checking. Along with their own position and how well they know the candidate, and anything they have already decided they will not claim. The last of those saves the most trouble, because a recommender who has said no to something in advance is far easier to write with than one discovering the limit halfway through.
How long should a recommendation letter be?
Long enough to carry four or five real claims and short enough that a reader with two hundred letters reaches the end. In practice that is usually one page, and the discipline that keeps it there is that nothing goes in unless it is on the claim list. A letter that runs to two pages is almost always one where adjectives have been allowed in, since particulars are compact and adjectives are not. Where a form specifies a limit, that limit decides and the content is fitted to it.
What if the recommender is hesitant or lukewarm?
It is the common case rather than the exception, and it is solvable without inflating anything. Inflating an honest letter turns it into one nobody believes and puts the recommender somewhere they will not describe afterwards. What works is narrower claims stated firmly: fewer adjectives, more particulars, and the one thing this person did that the reader will remember. Where the honest answer is that the candidate is capable but not outstanding, that is agreed up front and produces a perfectly good letter, which is better than the hedged version everybody writes instead.
Can you write in the recommender's voice?
To an extent, and reading something they have already written is the most useful input available, because it shows how firmly they are willing to speak. But the voice is the easy half. The hard half is that they did not choose the sentences and they cannot approve them one at a time without being given something to approve, which is why the finished letter goes out with its claim list attached. What should not happen is anybody other than the recommender deciding how strongly they are willing to speak, and that is the conversation this service starts with.
Is it acceptable to include comparisons?
Where a real one exists, it is usually the strongest claim in the letter. Best of the six students I have supervised since 2019 carries the recommender’s own standard inside it, which tells the reader both what the candidate is and what good looks like to the person writing, and it is checkable in a way that no adjective is. What is not acceptable is manufacturing one. Best of six is only available to somebody who supervised six, and an invented comparison is a claim the recommender cannot answer for when the reader rings to check.
What if the recommender does not know the applicant well?
Then the letter should be narrow, and honest about being narrow, because a letter from a near-stranger has a completely different job from one from a supervisor. What it can still carry is the one or two things they actually witnessed, which is usually more than people expect. It cannot carry anything about character under pressure, which is exactly what a reader most wants and exactly what nobody can attest to having seen. So the structure is short, specific, and clear about the capacity in which the recommender knows the candidate.
Who writes the letter, the recommender or the candidate?
The recommender, in every sense that matters, and the arrangement on this page is that the claims are theirs while the sentences are written for them. Which is not the same as ghostwriting and is worth being clear about, because the candidate is not the client here: they have no standing over what is said about them and cannot approve a claim about themselves. In practice that means the recommender signs, the claim list is agreed with them, and the candidate is told what the letter says rather than editing it.
How specific should it be about weaknesses?
One, stated once, in proportion, with what came of it, and then no further hedging. A letter with no weakness in it reads as one nobody read properly, because everybody knows what a real one contains. A letter that turns a small weakness into a strength is the same error in different clothing and readers are now trained to spot it. The practical effect is that a single honest limit in a page of praise makes the rest of the page believable, which is the opposite of what hedging does.
Can you guarantee the outcome?
No, and it would not be worth doing if it could. The outcome belongs to a panel or a hiring manager who has the whole file in front of them, and a letter is one part of it. What this service can do is remove the reasons a letter gets set aside: a claim nobody can check, a paragraph that repeats what is already in the file, an honest recommendation turned into an obviously inflated one, or a document that arrives after the deadline. None of that is a guarantee and all of it is within reach.
How quickly can it be delivered?
Quickly, but the constraint is the recommender rather than the writing. If they are available for one conversation about what they witnessed, a letter is straightforward. If they are travelling, or awkward, or slow to answer, nothing in the writing will fix that and it is better to know on the first day than the last. Where a deadline is tight the useful move is to agree the claim list first and write from it, rather than sending a questionnaire that comes back with adjectives a week before the closing date.
Where is Writely available, and how do I start?
Writely is based in Nigeria and works remotely with applicants, academics, professionals and institutions across Nigeria and the rest of Africa. Use the Chat with us button to send whatever the recommender remembers, how well they know the candidate, and anything they have already decided not to claim. That is enough to begin, and enough to say honestly whether this recommender can write this letter.

Your next move

Send the moments, not the praise. We’ll ask the recommender what they witnessed.

Whether you need the claims traced, the adjectives replaced, or the signer shown exactly what they are agreeing to, begin a clear conversation with Writely.

Request services