# The Send-Ready Cover Letter Standard, Graded Before You Attach It

*You will be able to grade any cover letter draft against explicit pass/fail gates and know exactly which lines to fix before you attach it.*

- Canonical URL: https://www.refolk.ai/candidates/guides/send-ready-cover-letter-standard
- Pillar: Positioning and materials
- Format: Standard
- Published: 2026-09-18
- Last reviewed: 2026-09-18
- Reading time: 16 min
- Keywords: cover letter checklist before sending, cover letter pass fail criteria, is my cover letter good enough to send, what a good cover letter must include, grade my cover letter

## Key takeaways

- In a 2024 ResumeGo survey of 418 US hiring professionals, 66% spent under 30 seconds on a cover letter, so the pass/fail test has to weight the first two sentences, not the body.
- The swap test is the only binary personalization criterion in the evidence: replace the company with a named competitor, and if a body sentence stays true, that sentence fails.
- Single-typo rejection rates range from 40% (Accountemps) to 77% (CareerBuilder) across different surveys, but the range still supports one policy: treat any typo or a wrong company name as automatic fail.
- Named recipient and personalized are two independent gates: a letter can name the manager and still fail the swap test, or pass the swap test while opening with Dear Hiring Manager.
- AI writes perfectly-formatted generic prose, so a clean format now proves less and swap-test-surviving detail proves more.
- In Refolk's index there are 138,692 US recruiters against 9,777 in the UK, a 14.2x gap, so a US-framed grading policy reaches the largest grader population.

Deciding whether a cover letter draft is send-ready is a judgement most people make on feel, which means two reviewers rarely agree and the same draft passes on Monday and fails on Friday. This guide is for anyone writing or rewriting a cover letter, especially when the draft came out of an AI tool and reads fine but says nothing. It converts "does this seem okay?" into eight pass/fail gates, so two people grading the same letter reach the same verdict and both know exactly which lines to fix.

The standard matters more now, not less. AI writes perfectly-formatted, grammatically clean prose that is completely generic, so the old proxies - neat layout, correct spelling - prove less than they used to. What survives is specificity that no template can fake. This document anchors the whole grade on one binary test that produces the same answer from any reviewer.

## What a good cover letter must include

A send-ready cover letter has a fixed set of parts, and every one is checkable. University career centers converge on the same components, which makes them a defensible definition of done rather than one office's taste.

The load-bearing components, where all authoritative sources agree:

- **Named-recipient salutation** with a colon after it, addressed to a real person by full name, or the fallback "Dear Hiring Manager:" when no name exists.
- **Position and company identification**: exactly which role you are applying for and how you found the opening.
- **Specific skill-to-job evidence**: concrete examples of your skills tied to the components of the job description.
- **Error-free text** that is neat and correctly formatted.

Optional and context-dependent parts include a referral mention and an explanation of a gap. These help in some cases and are absent in others, so they do not gate the grade.

Case Western's checklist requires addressing a specific named person, confirming the spelling of that name, a colon after the salutation, stating exactly which position and how you found it, stating why you want the role and organization, and giving specific skill examples. The University of Georgia specifies the header and contact block formatted like the resume, the date, the recipient's title and full name, and the greeting. Texas A&M defines the body as an introductory paragraph, a bridge paragraph tying your skills to the job description, and a closing paragraph restating interest and contact information.

> **Rule:** Structure is a gate, not a preference
>
> Every load-bearing component - named salutation with colon, position and company named, specific skill evidence, error-free text - must be physically present. A missing component is an automatic fail before any judgement of quality begins.

Note one genuine disagreement in the sources: they agree on the components but differ on the order of the contact block versus the date. Either order passes. Do not fail a letter on that alone.

## The format passing range: the definition of done for layout

The format range is tight and published, so it grades cleanly: a value is either inside the range or it is not. This is the easiest gate to pass and the one most people get wrong by a hair.

The University of North Carolina requires one page, business-letter format, and a standard font at 10 to 12pt matching the resume. The University of Michigan adds the layout specifics: 10 to 12pt in the same font as the resume, 0.5 to 1 inch margins, left justified, beginning no more than 2 inches from the top. The word-count consensus across commercial sources is 250 to 400 words across three to four concise paragraphs, with one page treated as a hard ceiling that should not be exceeded unless the posting requests more.

| Spec | University career centers | Commercial consensus |
|---|---|---|
| Page count | One page (UNC) | One page hard ceiling |
| Word count | 3 to 4 paragraphs (WWU) | 250 to 400 words |
| Font size | 10 to 12pt matching resume | 11 to 12pt |
| Margins | 0.5 to 1 inch | 1 inch, down to 0.75 inch |

The standard fonts named by UNC are Aptos, Arial, Calibri, Palatino, and Times New Roman. Matching the resume matters: a letter in a different typeface reads as assembled from parts.

> **Note:** Format compliance now proves less
>
> With 49% of hiring managers dismissing AI-generated resumes and AI producing flawless generic prose, a clean format no longer signals effort. It is table stakes. Passing the format gate earns you nothing; failing it costs you the read.

## Why the opening two sentences carry the grade

The first pass is a fast skim, so the pass/fail test must weight the opener far more than the body. In a 2024 ResumeGo survey of 418 US hiring professionals, 66% spent under 30 seconds reviewing a cover letter. On that pass, most of the body goes unread.

**66% - recruiters who spend under 30 seconds on a cover letter**

From a 2024 ResumeGo survey of 418 US hiring professionals, both recruiters and hiring managers.

A separate Resume Genius survey found a longer distribution - 48% spend 30 seconds to 2 minutes and 15% spend over 2 minutes - so sources disagree on duration. They agree on shape: the first pass is a skim, and the opening is where the only guaranteed impression forms. That is exactly why the most-flagged AI tell is an opener, not a mid-letter phrase.

The grading rule follows directly. Read only the first two sentences, in isolation, and ask whether they say something specific before a reader would move on. If the opener is a name-announcement ("My name is...") or the phrase "I am writing to express," it fails regardless of how strong the body is, because the body may never be read.

> A generic opener wastes the only impression 66% of recruiters are guaranteed to form.

The comparison point is instructive: a 2018 Ladders eye-tracking study found recruiters skim resumes for an average of 7.4 seconds. Cover letters get slightly more, but the same discipline applies - front-load the specific, or lose it.

## The swap test: the one criterion two graders can agree on

The swap test is the only personalization criterion in the evidence that is genuinely binary, which is what makes two-grader agreement possible. Replace every mention of the company and role with a named competitor; if a body sentence still reads as true and complete, that sentence is too generic and fails.

"Specific examples" is subjective - two graders will judge the same paragraph differently on feel. "Does this sentence stay true when the company name is a competitor" produces the same yes or no from any reviewer. Anchoring the whole standard on it is what separates this from a tips listicle.

#### How the swap test runs

1. **Identify** - Mark every company and role mention in the body
2. **Swap** - Replace each with a named competitor in the same market
3. **Reread** - Read each body sentence with the new name in place
4. **Verdict** - If a sentence stays true, it fails; the letter passes only when at least one becomes false

*A binary personalization check that any two graders execute the same way.*

The underlying principle is one career centers state plainly: generic templates are easily identified and signal a lack of interest in the specific opportunity. The swap test operationalizes that into something gradeable.

Two subtleties keep the test honest. First, swap the claim, not just the noun. "I've long admired [Company]" survives a naive swap because the grader only replaced the name; ask whether the underlying reason is still true of the competitor, and if it is, fail it. Second, personalization is not a summary of the company's About page. Reciting what the company does reads tailored but adds nothing. The detail has to connect to your own evidence, not just describe the employer.

Tailoring is not cosmetic. In a Resume.io study, applicants who tailored their cover letter received offers at 35.8% versus 21.2% for those who never tailored, and 40% of hiring managers named clear alignment with company needs as the single stand-out factor. Getting to that alignment across many postings by hand is slow, which is where a tool that drafts and tailors each letter from your own history earns its place. [Refolk](/candidates) writes the letter from your actual background and tailors it to each posting, so the swap-test-surviving detail is present before you start grading rather than something you retrofit.

## The automatic-fail phrase list

Some phrases fail a letter on sight because recruiters have seen them thousands of times and read them as a signal that the rest is generic too. These are automatic fails, not style notes.

The single most identifiable AI opener is: "I am writing to express my strong interest in the [Position] role at [Company]." It signals immediately that the letter will be generic. The word "delve" has become such a reliable AI marker that many recruiters flag any letter containing it. Level-mismatched vocabulary is another documented tell: an entry-level applicant writing about "spearheading transformative initiatives" or "leveraging synergistic opportunities" reads as machine-assembled.

Older human clichés predate AI and fail for the same reason:

- "To Whom It May Concern"
- "I am writing to express my interest"
- "I'm the best" or "uniquely qualified candidate"
- "I'm a fast learner"

> **Watch out:** The scan misses paraphrases
>
> Searching for the exact word "delve" misses "dive deep into." Grade the vocabulary-level mismatch, not only the banned strings. If an entry-level writer is using C-suite verbs, fail it whether or not the specific flagged word appears.

The deeper point: because AI produces clean, well-formatted, generic prose, the load-bearing signal has shifted from format to specificity that a model cannot fabricate without your inputs. The phrase list catches the model's fingerprints. The swap test catches the emptiness underneath them.

## How harsh is a single typo

Treat any typo, or a wrong company name, as an automatic fail. The exact rejection rate varies by survey, but every source points the same way, and the range is defensible where a single cherry-picked figure is not.

| Survey source | Reported rejection rate | Fault measured |
|---|---|---|
| CareerBuilder | 77% | single typo or grammar mistake |
| CareerBuilder | 61% | typo spotted |
| Accountemps (150 execs) | 40% | one typo |

These rates come from different vendors and years, so they are not poolable - present them as a range, 40 to 77%. UNC states the rule as policy: any misspelling will automatically disqualify a cover letter. Given that spread, the honest conclusion is not a precise probability; it is that the downside of one typo is severe enough that the only safe posture is zero defects.

The correct-but-wrong error is the dangerous one. Spell-check passes "dear sir or madman" and a correctly-spelled but wrong company name. A green checker result is a false positive. The misnamed-company fault is the classic recycling tell - forgetting to swap the company or person's name is one of the faults HR professionals flag most. A second human has to read the letter aloud and verify the company name against the posting.

**40 to 77% - hiring managers who would reject on a single typo**

The range spans an Accountemps survey of 150 executives (40%) and a CareerBuilder survey (77%). Different vendors, same direction.

## The grading procedure

Run these eight steps in order. The first three are structural and fast; the middle three carry the real judgement; the last two make the verdict reproducible. Steps 4 through 6 are best done by a second person, because the author cannot un-see their own intent.

#### Grade the draft before you attach it

1. **Check structural completeness** - Confirm every required part is present - header and contact block, date, recipient block, named salutation with a colon, opening, evidence body, closing with a call to action. Done when every component from the standard checklists is on the page.
2. **Run the format and spec pass** - Verify one page, 250 to 400 words, a 10 to 12pt standard font matching the resume, 0.5 to 1 inch margins, left justified. Done when every value falls inside the published range.
3. **Verify the salutation and recipient** - Confirm a real named person or the fallback Dear Hiring Manager, never To Whom It May Concern, spelled correctly, with a colon. Done when the name is confirmed against a source or you have documented that none could be found.
4. **Apply the opening-lines skim test** - Read only the first two sentences in isolation, because the first pass is under 30 seconds. Done when the opener is specific and non-generic without a name-announcement or I am writing to express.
5. **Run the swap test for personalization** - Replace every company and role mention with a named competitor and reread each body sentence. Done when at least one body sentence becomes false after the swap; rewrite any that stay true.
6. **Scan for AI tells and cliches** - Search for the documented AI and cliche phrases and for level-mismatched vocabulary. Done when zero automatic-fail phrases remain and no C-suite verbs are used by an entry-level writer.
7. **Do a zero-defect proofread** - Read aloud, run a checker, and have a second human confirm, because spell-check passes correctly-spelled wrong words. Done when there are zero typos, the company name is correct, and tenses are consistent.
8. **Reach a two-grader verdict** - Have two independent reviewers apply the same checklist and compare pass/fail line by line. Done when both reach the same overall send or no-send verdict.

The whole procedure takes about 35 minutes with a peer, most of it in steps 5 through 7. If you are grading alone, do the swap test on a printed copy the next morning; distance substitutes imperfectly for a second reader.

## How the grade goes wrong

This is the part of the standard that matters most, because every failure mode below is a case where a letter passes a naive check while failing the real criterion. A standard is only as good as its false-positive defenses.

**Word count passes but the opener fails.** A 300-word letter can still open with "I am writing to express." It is format-legal and skim-fatal. Check: read only sentences one and two in isolation; if generic, fail regardless of length.

**Swap-test false positive.** "I've long admired [Company]" survives a naive swap because the grader only replaced the noun. Check: swap the name and ask whether the underlying reason is still true of the competitor. If yes, fail.

**Spell-check clean but still wrong.** The checker passes "dear sir or madman" and a correctly-spelled wrong company name. Check: a second human reads aloud and verifies the company name against the posting.

**AI-tell scan misses paraphrases.** Searching "delve" misses "dive deep into." Check: grade the vocabulary-level mismatch, not just the exact banned words.

**Named-recipient theater.** "Dear Hiring Manager" passes the "has a salutation" box but fails the "named person" criterion. Check: require an actual name or documented evidence that none could be found.

**Personalization that is website summary.** A paragraph reciting the About page reads tailored but adds nothing. Check: does the detail connect to your own evidence, or just describe the company?

**Two graders diverge on "specific enough."** Without the swap test, "specific examples" is subjective and reviewers split. Check: force every grader to run the binary swap test rather than judge feel.

**Restates the resume.** The letter passes all format checks while duplicating bullet points. Check: does each body claim add context the resume lacks?

```figure
kind: matrix
title: Named recipient and personalized are two independent gates
caption: A letter can pass one gate and fail the other, so grade them separately.
x: Generic body (fails swap test) :: Specific body (passes swap test)
y: Dear Hiring Manager :: Named real recipient
x: Fallback greeting, generic body :: Fix the greeting or document no name exists
x: Named recipient, generic body :: Rewrite the body until the swap test bites
x: Both gates fail, do not send :: Send-ready on these two gates
```

Score named recipient and personalization as two independent gates. A letter can name the manager yet fail the swap test, or pass the swap test with "Dear Hiring Manager." Collapsing them into one "is it personal" impression is how graders disagree.

## Who is grading, and at what scale

The population applying this standard is enormous and US-concentrated, which is worth knowing because it tells you whose skim time you are writing for. In Refolk's index there are 138,692 US profiles holding the titles Recruiter, Technical Recruiter, or Talent Acquisition Specialist - the people grading cover letters at scale.

| Segment | US count | UK count | US:UK ratio |
|---|---|---|---|
| Recruiter / Tech Recruiter / TA Specialist | 138,692 | 9,777 | 14.2x |

The advisory side that teaches these criteria is smaller but real: Refolk's index returns 5,407 US profiles for Career Coach or Resume Writer. They are the multiplier who turn a personal checklist into team policy. If you want to see who is actually reading your openers in a given market, you can query the population directly.

Ask me this: `Technical recruiters in Austin who screen software engineering applications at staffing firms` - [run the search](https://www.refolk.ai/start?q=Technical%20recruiters%20in%20Austin%20who%20screen%20software%20engineering%20applications%20at%20staffing%20firms).

*Returns the recruiters who apply the under-30-second skim, so you can see the profile of the person your opener has to survive.*

## The send-ready checklist

Run this immediately before you attach the file. It is the compressed form of the eight steps and the failure-mode defenses. Every item is a checkable statement, not a topic.

#### Verify before you attach

- [ ] All load-bearing components present - header, date, recipient block, named salutation with colon, opening, evidence body, closing with call to action
- [ ] One page, 250 to 400 words, 10 to 12pt standard font matching the resume, 0.5 to 1 inch margins, left justified
- [ ] Salutation names a real person spelled correctly, or documents that no name could be found; never To Whom It May Concern
- [ ] First two sentences read as specific in isolation, with no name-announcement and no I am writing to express
- [ ] At least one body sentence becomes false when the company name is swapped for a named competitor
- [ ] Zero automatic-fail phrases, no delve or paraphrases of it, no vocabulary mismatched to your level
- [ ] Read aloud by a second human, company name verified against the posting, zero typos and consistent tenses
- [ ] Two independent reviewers reach the same send or no-send verdict line by line

## Keeping the standard current

The mechanisms in this standard are durable; the specific numbers are not, so re-check them rather than trusting them forever. Skim-time surveys are reissued periodically - when you see a new one, note the sample size and whether it separates recruiters from hiring managers, because the 66%-under-30-seconds figure comes from 418 US professionals in a single 2024 survey and a different sample will move it. Treat the direction (a fast first-pass skim) as stable and the exact percentage as replaceable.

The automatic-fail phrase list is the part that ages fastest. As AI writing tools change their defaults, new tells appear and old ones fade. "Delve" is flagged today because a generation of models overused it; the mechanism - a phrase becomes a marker once recruiters have read it thousands of times - will outlast the specific word. Re-check the list against what hiring managers are currently writing about publicly, and add new markers as they surface. The swap test needs no updating: it depends on nothing but your letter and a competitor's name, which is exactly why it is the spine of the standard.

## Frequently asked questions

### How do I tell if my cover letter is good enough to send?

Grade it against eight gates rather than judging feel. It must be structurally complete, inside the format range, addressed to a verified name, specific in its first two sentences, free of any sentence that survives the swap test, clear of AI-default phrases, and free of a single typo. If any gate fails, fix the offending line before you attach it. A draft is send-ready only when it clears all eight.

### What is the swap test for a cover letter?

Replace every mention of the company and role with a named competitor, then reread each body sentence. If a sentence still reads as true and complete after the swap, it is too generic and fails. The test is binary, so two graders swapping the same name reach the same verdict. Your letter passes only when at least one body sentence becomes false, meaning it genuinely depended on this employer and not any employer.

### Is one typo really enough to get a cover letter rejected?

Reported rejection rates for a single typo range from 40% in an Accountemps survey of 150 executives to 77% in a CareerBuilder survey. The exact figure varies by vendor and year, so no single number is gospel. But every source points the same direction, and one university career center states any misspelling automatically disqualifies a letter. Treat any typo, or a wrong company name, as an automatic fail.

### What phrases signal an AI-written cover letter?

The single most identifiable AI opener is I am writing to express my strong interest in the [Position] role at [Company]. The word delve has become a reliable AI marker that many recruiters flag on sight. Level-mismatched vocabulary, such as an entry-level applicant claiming to be spearheading transformative initiatives, is another tell. Older human clichés like To Whom It May Concern and I am a fast learner fail for the same reason. Remove all of them.

### How long should a cover letter be to pass?

The passing range is 250 to 400 words across three to four concise paragraphs on a single page. One page is a hard ceiling, and the letter should not exceed 400 words unless the posting specifically requests more. Use a 10 to 12pt standard font that matches your resume, with 0.5 to 1 inch margins, left justified. Word count alone does not make a letter good, but falling outside the range is an automatic format fail.

### Why weight the opening two sentences so heavily?

Because in a 2024 ResumeGo survey of 418 US hiring professionals, 66% spent under 30 seconds reviewing a cover letter. On that first pass most of the body goes unread, so a generic opener wastes the only guaranteed impression. That is also why the most-flagged AI tell is an opener rather than a mid-letter phrase. Read your first two sentences in isolation and ask whether they say anything specific before a reader would move on.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/candidates/guides/send-ready-cover-letter-standard*