# The Resume Term Currency Standard, Graded Against 20 Dated Postings

*Grade every term on your resume against 20 dated postings with written count thresholds, so a second grader reaches the same verdict on every term.*

- Canonical URL: https://www.refolk.ai/candidates/guides/resume-term-currency-standard
- Pillar: Reading the market
- Format: Standard
- Published: 2026-10-26
- Last reviewed: 2026-10-26
- Reading time: 23 min
- Keywords: outdated skills on resume how to tell, renamed software which name to put on resume, how to list expired certification on resume, check resume keywords against current job postings, resume terminology out of date recruiters

## Key takeaways

- Grade resume terms against 20 target postings pulled on one date, because two postings for the same job title share a median of only 25% of their named requirements.
- Write the demand floor as a count, 3 of 20, not as a percentage: each posting in a 20-posting sample is worth 5 points, so a 2% research floor cannot be measured.
- In Refolk's index of US profiles, Azure Active Directory still outnumbers Microsoft Entra ID 13,275 to 11,311, 39 months after the rename, which is why both names often belong on one line.
- Switching to the newest vendor name is not automatically safe: Looker Studio reached only 16.6% of combined profile mentions in Refolk's index before Google reverted the name to Data Studio.
- A retired Microsoft credential can stay in the transcript's Active section for up to two years, so credential status must come from the issuer's retirement page and expiry date, not the transcript.
- A resume passes only when every term carries one of five verdicts, every verdict has been applied, and a second grader working blind from the same saved sample agrees on all of them.

This standard is for job seekers about to send another batch of applications from a base resume who need to know that every tool, method, credential, and skill name on it is the name employers use now. It gives you a fixed sample of 20 dated postings, written count thresholds, five verdicts (pass, rename, dual-list, label, cut), and a checklist. Two people grading the same resume against the same saved sample should reach the same verdict on every term.

## What counts as current vocabulary on a resume?

A term on your resume is current when that exact name appears in at least 3 of 20 target postings pulled on one date, and no other name for the same thing also reaches 3. A credential is current only when the issuer's own record shows it active, earned on a current exam, and not retired.

Several terms in that definition need fixed meanings before anyone grades anything:

- **Term**: any tool, method, credential, or skill name on the resume, wherever it appears (skills section, summary, bullets, certifications).
- **Posting sample**: the 20 postings for your target titles, pulled on one date and saved. Every count in this standard comes from it.
- **Canonical name**: the single display name a skills taxonomy gives a skill. Lightcast's taxonomy covers over 33,000 skills and gives each one display name, but it also uses "aliases, acronyms, abbreviations, and historic names" to find the skill in text.
- **Alias**: any other name for the same thing, including the name it had before a rename.
- **Name count**: how many of the 20 postings use one particular name, counted once per posting.
- **Concept count**: how many of the 20 postings mention the thing under any of its names.
- **Demand floor**: the minimum count below which a name or concept is treated as absent from your market. Here it is 3 of 20.

### The five verdicts

Every term gets exactly one verdict.

| Verdict | Applies when | What changes |
|---|---|---|
| Pass | Resume name in 3+ of 20, no other name in 3+ | Nothing |
| Rename | Resume name under 3, another name in 3+ | Swap to the other name |
| Dual-list | Two names each in 3+, or concept in 3+ with no name at 3 | Both names on one line |
| Label | Credential expired, retired, in progress, or on a retired exam code | Issuer status and dates added |
| Cut | Concept under 3 of 20, or a lapsed credential under 3 | Removed from lists |

A cut takes the term out of the skills section, the summary, and the certifications block. It can stay inside an experience bullet only where the bullet would be false without it, such as a migration you ran away from that tool.

This standard does not decide which skills to add, which job titles to use, or how to repoint a resume toward adjacent roles. It grades only the names already on the page.

## Why can't one posting tell you whether a term is current?

Because postings for the same job title disagree more than they agree. In a study of 3,910 postings, two postings for the same title shared a median of just 25% of their named requirements. A match against one posting mostly measures one employer's wording.

That is the weak point of the usual advice to check resume keywords against current job postings. Keyword tools score a resume against one description at a time. Labour-economics studies do it the other way round and measure how often a skill appears across a whole corpus first. To decide whether a word is current, the corpus approach is the right one, so this standard fixes the sample at 20.

This is not an argument against tailoring to each posting. [Refolk](/candidates) tailors your resume to each posting you apply to and scores how well you actually fit, and that step works on one posting because it is meant to. The base it starts from needs the wider test, because a stale name in the base gets copied into every tailored version.

### What the filtering evidence shows, and what it does not

Recruiters filter, and they say skills are what they filter on most. In Jobscan's survey of 384 recruiters, over 99.7% use filters in their applicant tracking system (ATS, the software that stores and searches applications).

| Filter | % of recruiters using it |
|---|---|
| Skills | 76.4 |
| Education | 59.7 |
| Job title | 55.3 |
| Certifications/licenses | 50.6 |
| Years of experience | 44.3 |
| Location | 43.4 |

Three limits stop this evidence from proving more than it does:

- **What recruiters say versus what they do.** A reported analysis of more than 100,000 searches on the sourcing tool Pin found job titles in 98.3% of searches. That figure reaches us through a secondary source, but it shows that what recruiters report and what they actually search for can differ.
- **Alias handling is undocumented.** I found no public documentation of how Workday, Greenhouse, iCIMS, or other ATS parsers treat aliases. The claim that ATS ignore synonyms appears only in unsourced blog posts. You cannot count on a parser to map an old name to a new one, and you cannot prove that it won't.
- **The cost of one stale term is unmeasured.** In the Harvard Business School and Accenture Hidden Workers study, 88% of employers agreed that qualified high-skills candidates are screened out for not matching the exact criteria. That shows exact-match screening exists. It does not put a price on one outdated word.

The practical conclusion is that the only name you can be sure will be found is the literal name your target postings use. Jobscan detected an ATS at 97.4% of Fortune 500 companies in its latest count, so the literal match matters at most large employers.

## What are the sample rules, counting rules, and demand floor?

The sample is exactly 20 postings for your target titles, pulled on one date and saved as text, with no more than two from any one employer. The demand floor is 3 of 20. It applies to every name and every concept, and it is always written as a count, never as a percentage.

### Sample rules

1. Exactly 20 postings, all for titles you will actually apply to in the next batch.
2. All pulled on one date. Record the date and each URL.
3. No more than two postings from any one employer.
4. Save the full text of each posting. Postings get edited and taken down, and the second grader needs the same words you counted.
5. Once saved, the sample is frozen. Adding a posting later means pulling a new sample and regrading.

Jobscan's advice to "start from the job description rather than your old resume" applies here. Pull the sample before you look at your term list, so the list cannot influence which postings get in.

### Counting rules

1. A name counts once per posting, however many times it appears.
2. Count the whole posting text except the employer's "about us" section and legal boilerplate.
3. Acronyms, abbreviations, and vendor-prefix variants of one name count as that one name. "Azure AD" and "Azure Active Directory" are one name. So are "Google Data Studio" and "Data Studio".
4. A renamed product counts its old and new names separately: "Microsoft Entra ID" is counted apart from "Azure AD".
5. The concept count is the number of postings that use any name for the thing. It is not the sum of the name counts. A posting that says "Microsoft Entra ID (formerly Azure AD)" adds 1 to each name count and 1 to the concept count.

### Why the floor is 3 of 20

Floors from research studies don't scale down to one person's sample. One study of 1.14 million postings kept only skills with at least 2% average prevalence. Another required a skill to appear at least 5 times in a year. In a 20-posting sample, each posting is worth 5 percentage points, so a 2% floor cannot even be measured.

The floor of 3 is a policy choice, and it rests on one fact about the sample. With at most two postings per employer, any name that reaches 3 has been used by at least two different employers. One mention can be one employer's in-house vocabulary. Three cannot.

Be clear about what the floor does not do. At an observed rate of 25% (5 of 20), the 95% interval on the true rate is roughly plus or minus 19 points. A term at 3 of 20 and a term at 6 of 20 do not reliably differ in demand. The floor sorts terms into present and absent the same way every time. It does not rank them.

> **Rule:** Thresholds are counts out of 20
>
> Every threshold in this standard is written as a whole-number count of postings, such as "3 of 20", never as a percentage. Two graders can disagree about a percentage; they cannot disagree about a count taken from the same saved text.

#### The order verdicts are assigned in

1. **Credential check** - If the term is a credential, read status from the issuer; active, current, and not retired passes
2. **Lapsed credential** - Expired, retired, in progress, or old exam code gets label at 3+ of 20, cut under 3
3. **Concept floor** - Any other term whose concept count is under 3 of 20 is cut
4. **Name split** - Compare name counts to choose pass, rename, or dual-list
5. **Record** - Write the counts, the verdict, and the applied form on the grading sheet

*Credential status is settled first, the concept floor second, and the split between names last, so every term reaches exactly one verdict.*

#### Choosing between pass, rename, and dual-list

Horizontal axis runs from Resume name under 3 of 20 to Resume name 3+ of 20. Vertical axis runs from Other name under 3 of 20 to Other name 3+ of 20.

| Quadrant | What it means |
| --- | --- |
| Rename | Replace your name with the one the postings use |
| Dual-list | Put both names on one line, higher count first |
| Check concept count | Dual-list the top two names if the concept is 3+, cut if under 3 |
| Pass | Keep the name as written |

*Once a concept clears the floor, the count for your resume's name and the count for the strongest other name decide the verdict.*

## How are renamed products graded?

Count every name separately across the sample and let the counts choose between pass, rename, and dual-list. The vendor's rename date decides nothing on its own, because vendors, employers, and candidates switch names at different speeds.

Refolk's index of professional profiles shows how slowly the candidate side moves.

| Product | Old-name profiles | New-name profiles | New-name share |
|---|---|---|---|
| Azure AD / Entra ID | 13,275 | 11,311 | 46.0% |
| Google Data Studio / Looker Studio | 8,392 | 1,668 | 16.6% |

US profiles in Refolk's index. Counts were taken 39 months after the Entra rename and 42 months after the Looker Studio rename. The shares are derived from the counts.

**1.17x - US profiles listing Azure Active Directory for every one listing Microsoft Entra ID**

In Refolk's index, 39 months after Microsoft's rename, the old name is still the more common one on candidate profiles.

The table shows two things, and there is one thing it does not show.

First, old names outlive renames by years. Microsoft announced the rename of Azure AD to Microsoft Entra ID and relabelled its own products in phases. Thirty-nine months later, more US profiles in Refolk's index still say Azure Active Directory than Microsoft Entra ID. People rarely edit their profiles, so old skill tags pile up, and a recruiter searching only for the new name misses everyone who never updated. That is the case for the dual-list verdict. When your sample shows both names at 3 or more, writing both on one line is the only form that matches either search.

Second, the newest name is not automatically the safe one. Google renamed Data Studio to Looker Studio. Looker Studio reached only 16.6% of combined profile mentions in Refolk's index, and then Google changed the product's name back to Data Studio.

**16.6% - Share of combined US profile mentions that used "Looker Studio" before Google reverted the name**

In Refolk's index, 8,392 profiles still said Google Data Studio against 1,668 that said Looker Studio.

Anyone who rushed to put the newer name on their resume ended up with the abandoned one. The answer to "which name do I put for renamed software" is whichever name your postings use, checked against the vendor's current naming.

What the table does not do is grade your resume. These are counts of what candidates write, not what employers write. Grade only against your posting sample.

### The dual-list format

When the verdict is dual-list, put the higher-count name first and the other in parentheses on the same line:

- Microsoft Entra ID (formerly Azure AD)
- Data Studio (formerly Looker Studio)

If the counts tie, lead with the vendor's current name as shown in its own release notes. Taxonomies keep historic names on purpose, which is why a Lightcast-style extractor may still match an old name. That protection only exists where the parser's rules are documented, and no ATS publishes its alias rules, so do not rely on it.

> **Watch out:** Renames can go both ways
>
> Anyone who changed "Data Studio" to "Looker Studio" now has the outdated name again. Confirm every product's current name on the vendor's own release notes, not on a third-party explainer, before you rename anything.

## How are credentials graded?

A credential is graded on the issuer's record first and on posting demand second. If the issuer's verification portal shows it active, earned on a current exam, and not retired, it passes. Anything else gets a label, or a cut if fewer than 3 of 20 postings mention it.

Write each credential under its exact issuer name. The rename test does not apply to credentials, and active credentials are exempt from the floor, because whether to list a valid credential at all is a positioning question outside this standard. Credentials still matter for screening: 50.6% of recruiters in Jobscan's survey filter on certifications or licenses.

| Credential class | Validity | After retirement |
|---|---|---|
| Microsoft role-based/specialty | 1 year, renewable in final 6 months | Active until expiry |
| Microsoft Fundamentals | Does not expire | Active for 2 years after retirement |
| Microsoft MOS/MCE | 5 years | n/a |
| CompTIA CE (2011 onward) | 3 years | Credential not invalidated |
| CompTIA pre-2011 | Lifetime | n/a |

#### What sits under the word "active"

1. **Resume line** - What you wrote, which is the weakest evidence
2. **Transcript or certificate** - Can show a retired credential as Active, or print a superseded exam code
3. **Expiry date** - 1 year for Microsoft role-based, 3 years for CompTIA CE, none for Fundamentals
4. **Retirement and exam version** - Whether the credential can still be earned and which exam is current

*A credential's status has four layers, and only the issuer's expiry and retirement records settle whether the line on your resume is accurate.*

### What "active" can hide

- **Microsoft role-based and specialty certifications** are valid for one year from the date you complete all requirements, and you can renew them in the final six months. Fundamentals certifications do not expire.
- **Retired Microsoft credentials** stay in your transcript's Active section until they expire. If they have no expiry date, they stay there for two years after retirement. The transcript alone cannot tell you whether the credential can still be earned, so check it against the credential retirement page.
- **CompTIA certifications earned under the continuing education (CE) program** expire every three years and need 20 continuing education units to renew. Certifications earned before the CE program began are lifetime.
- **A retired exam does not invalidate a CE credential.** Security+ exam SY0-601 has been retired and replaced by SY0-701. Someone who earned Security+ on SY0-601 and kept up CE renewal still holds a valid Security+. As one verification guide puts it, "trust the portal over the version listed on a physical certificate".

The label verdict is the answer to how to list an expired certification on a resume, but the exact wording is a judgement call. No one has published a recruiter survey on which status labels employers consider accurate. Resume-advice sources do agree that listing an expired certification without saying so is misleading. The formats below are fixed so that two graders produce the same line.

**Credential label formats**

```
Active, current exam (pass, no label):
<Issuer credential name>

Active, earned on a retired exam, kept valid by CE:
<Issuer credential name> - active through <Mon YYYY>, renewed via continuing education

Expired:
<Issuer credential name> - earned <Mon YYYY>, expired <Mon YYYY>

Retired by issuer, still in date:
<Issuer credential name> - active through <Mon YYYY>; credential retired by issuer

In progress:
<Issuer credential name> - in progress, exam booked for <Mon YYYY>
```

*Replace the angle-bracket fields with the issuer's exact credential name and the month and year from the issuer's portal; never print a retired exam code.*

> **Watch out:** The Active section is not a status check
>
> A retired Microsoft credential can sit in the transcript's Active section for up to two years. Compare the expiry date with the retirement page before grading any Microsoft credential as pass.

## Baseline skills employers assume and no longer list

Grade baseline skills exactly like any other term. If fewer than 3 of 20 postings name one, cut it from the skills list, and if it matters, show it inside a bullet as demonstrated evidence.

Lightcast classes these as Common Skills, "sometimes referred to as baseline skills", and gives "Communication" or "Microsoft Excel" as examples. They are the classic case of outdated skills on a resume: not wrong, just so assumed that postings no longer bother to name them.

The employer evidence points toward evidence over lists. In NACE's employer surveys, screening by GPA fell from 73% of employers to 42%, while skills-based hiring rose to 70%. And 73% of employers that do not screen by GPA, and 67% of those that do, consider whether a candidate has demonstrated proficiency in key competencies. Employers rely on skills more, but they look for proof of them.

There is a limit here. No one has published a single figure for how far employers prefer examples over a bare skills list. So a cut is not a claim that the skill is worthless. It is a claim that a bare list entry for a skill your sample does not name adds nothing, and that the skills section is where current names belong. If 3 or more postings do name the skill, it passes like any other term.

## The grading procedure, step by step

Seven steps take one grader from a base resume to a verdict on every term, then hand the work to a second grader. The timed steps add up to roughly two and a half to three hours, plus about 15 minutes per credential.

#### Grading a base resume for term currency

1. **Freeze a dated 20-posting sample** - Pull 20 postings for your target titles on one date, no more than two per employer, and save each URL, the pull date, and the full text. Done when the saved set cannot change (about 45 minutes).
2. **List every term on the resume** - Put each tool, method, credential, and skill on its own row of the grading sheet, from every section. Done when no name on the page is missing (about 20 minutes).
3. **Normalize each term to a canonical name** - Use Lightcast's Skills Extractor or its alias logic to record each term's canonical name, aliases, and historic names. Done when every row lists every name you will count (about 30 minutes).
4. **Count every name across all 20 postings** - Count each name once per posting, then count the postings that mention the concept under any name. Done when every name and concept has a count out of 20 (60 to 90 minutes).
5. **Check every credential at the issuer** - Use the CompTIA verification portal or the Microsoft Learn transcript and retirement page, and record status, expiry date, and exam version. Done when no status rests on the resume or a certificate (about 15 minutes each).
6. **Assign a verdict to every term** - Apply the decision order: credential status, then the concept floor, then the name split. Done when every row carries exactly one verdict.
7. **Have a second grader re-grade blind** - Give a second person the saved sample and the term list without your verdicts. Done when both agree on every term and on whether the resume passes.

A few notes on the harder steps:

- **Step 3.** Lightcast refreshes its taxonomy every few weeks, and sources disagree on whether it is every two or every four. Check canonical names when you grade rather than relying on a list you made months ago. O*NET and ESCO are other public taxonomies you can cross-check a canonical name against.
- **Step 4.** Search each saved posting for every name on the row, including acronyms. Mark hits per posting, not per occurrence.
- **Step 5.** Record the exam version even when the credential is active, because a retired exam code printed on the resume triggers a label.

**Grading sheet**

```
Sample folder: <path> | Pull date: <YYYY-MM-DD> | Postings: 20 | Max per employer: 2 | Floor: 3 of 20
Grader: <name> | Blind second grader: <name>

Term as written | Type (tool/method/credential/skill) | Canonical name | Other names counted | Resume-name count /20 | Top other-name count /20 | Concept count /20 | Issuer status, expiry, exam version | Verdict | Applied form on resume
Azure Active Directory | tool | Microsoft Entra ID | Microsoft Entra ID; Entra ID | _/20 | _/20 | _/20 | n/a | |
CompTIA Security+ | credential | CompTIA Security+ | Security+; SY0-701 | _/20 | _/20 | _/20 | <from portal> | |
```

*One row per term; copy the header into a spreadsheet and keep the sample folder path with it so the second grader opens the same text.*

## How does this standard go wrong?

It usually goes wrong by measuring the wrong thing with confidence. Seven failure modes cause most bad verdicts, and each has a specific check that catches it.

### 1. The one-posting match score

**What it looks like:** a high match percentage against one description, taken as proof the vocabulary is current. **Why it lies:** two same-title postings share a median of only 25% of their named requirements, so a strong match on one can pass terms that most of your target postings never use. **Check:** every count comes from the full 20.

### 2. A floor finer than the sample can measure

**What it looks like:** a threshold copied from research, such as "keep skills above 2%". **Why it lies:** each posting is worth 5 points in a sample of 20, so 2% falls between possible values. At a 25% observed rate the 95% interval is roughly plus or minus 19 points, so small differences in count are noise. **Check:** the threshold is written as 3 of 20, and no verdict depends on comparing two counts that both clear the floor.

### 3. Profile counts treated as posting demand

**What it looks like:** "most people still say Azure Active Directory, so I will too." **Why it lies:** Refolk's index counts what candidates write, not what employers ask for, and candidates update more slowly than employers do. **Check:** grade only against the dated posting sample. Use profile counts to understand why old names persist, never to decide a verdict.

### 4. Round-trip renames

**What it looks like:** a rename verdict based on the last name change you remember. **Why it lies:** Data Studio became Looker Studio and then Data Studio again, so an early update left people with the outdated name. **Check:** confirm the current name on the vendor's own release notes, then let the posting counts decide.

### 5. "Active" read off the transcript

**What it looks like:** a Microsoft credential graded pass because it sits in the Active section. **Why it lies:** retired credentials stay there until they expire, or for up to two years. **Check:** compare the expiry date with the credential retirement page.

### 6. Exam version mistaken for status

**What it looks like:** a CompTIA credential graded as lapsed because the certificate shows a retired exam code, or graded pass because the code is printed next to the name. **Why it lies:** an older exam version is still valid if CE renewal was kept up, and a paper certificate does not show renewal. **Check:** trust the portal. Label with the "active through" format and drop the retired code.

### 7. Survey answers taken as observed behaviour

**What it looks like:** "76.4% of recruiters filter by skills, so every stale term costs me interviews." **Why it lies:** that is a survey answer, while one reported analysis of search logs found titles in 98.3% of searches. **Check:** do not rest the cost of a stale term on the survey alone. The standard fixes names because the literal match is the only one you can verify, not because the survey puts a figure on the loss.

> A term is current when the postings you are about to answer use it, not when the vendor says so.

## Passing the whole resume and keeping it current

The resume passes when every term carries a verdict, every rename, dual-list, label, and cut has been applied to the resume itself, and a second grader working blind from the same saved sample reaches the same verdict on every term. Any outstanding change, or any disagreement between graders, means it fails.

### Running the blind regrade

Give the second grader the saved posting folder, the pull date, and the term list with the canonical names filled in, but no counts and no verdicts. If you disagree on a term, recount it from the saved text together. If the counts match but the verdicts still differ, a rule in your copy of this standard is ambiguous. Rewrite that rule in plain words and regrade every term it touches, not just the one that exposed it.

### When to regrade

- **You change target title.** Pull a new sample and run every step.
- **A vendor renames a tool on your resume.** Regrade that row against a fresh sample once postings have had time to pick up the new name. Expect both names to coexist for years.
- **A credential enters its renewal window or its issuer announces a retirement.** For Microsoft role-based credentials, the window is the final six months. Rerun step 5 for that credential.
- **The saved sample no longer reflects the postings you are answering.** Re-pull and recount (steps 1, 4, and 6). Your normalized names and credential records carry over.

Each regrade is cheaper than the first, because the term list and canonical names barely change. Do it before each batch rather than after a run of silence. Once the base passes, [Refolk](/candidates) carries those current names into every posting-specific version it tailors and scores your fit against each one, so you fix vocabulary once in the base instead of once per application.

#### Before the base resume goes out

- [ ] The sample has exactly 20 postings, one pull date, and no more than two from any employer.
- [ ] The full text of every posting is saved where the second grader can open it.
- [ ] Every tool, method, credential, and skill on the resume has its own row on the grading sheet.
- [ ] Every name and every concept has a count written as a number out of 20.
- [ ] Every credential's status, expiry date, and exam version came from the issuer's portal, not the transcript's Active section or a certificate.
- [ ] Every renamed product's current name was confirmed on the vendor's own release notes.
- [ ] Every term carries exactly one verdict: pass, rename, dual-list, label, or cut.
- [ ] Every rename, dual-list, label, and cut has been applied to the resume itself.
- [ ] No retired exam code appears next to any credential.
- [ ] A second grader working blind reached the same verdict on every term and the same pass or fail for the whole resume.

## Frequently asked questions

### Renamed software: which name should I put on my resume?

Put whichever name your target postings use, and put both if both are in use. Count each name separately across 20 postings pulled on one date. If only the new name reaches 3 of 20, rename. If both reach 3, write both on one line with the higher count first, such as "Microsoft Entra ID (formerly Azure AD)". Do not default to the newest vendor name: Google renamed Data Studio to Looker Studio and later reverted it.

### How do I list an expired certification on my resume?

List it with the dates you earned it and when it expired, or cut it if fewer than 3 of your 20 target postings mention it. An expired credential shown without dates reads as active, and resume advisers call that misleading. Check status on the issuer's portal first: a retired Microsoft credential can sit in the transcript's Active section for up to two years, and a CompTIA certification earned on a retired exam is still valid if continuing education was kept up.

### How can I tell if the skills on my resume are outdated?

Count them against a dated sample of 20 target postings, not against one. A skill that fewer than 3 of the 20 name, under any of its names, is below the demand floor and comes off the skills list. One posting is not enough because two postings for the same title share a median of only 25% of their named requirements, so a single match mostly reflects one employer's wording.

### Do applicant tracking systems recognise old product names as synonyms?

Nobody can say for certain, because no major ATS publishes its alias rules. Skills taxonomies such as Lightcast's deliberately keep historic names and aliases for extraction, so some systems may match an old name. The claim that ATS ignore synonyms rests only on unsourced blog posts. The only name you can be sure will match is the literal name in the posting, and that is why the dual-list verdict exists.

### Is a keyword match score against one job posting enough to check my resume?

No, not for deciding whether a term is current. A single-posting score tells you how closely you match one employer's wording, which is useful when you tailor one application. Whether a term belongs on the base resume depends on your whole target market, so count it across 20 postings and apply a floor of 3.

### Should I still list Microsoft Excel and communication as skills?

Only if at least 3 of your 20 target postings name them. Lightcast classes these as Common Skills, also called baseline skills, and many postings assume them without listing them. When they fall below the floor, cut them from the skills section and show them inside a bullet with a result. Most employers in NACE's survey weigh demonstrated proficiency, not a bare list entry.

---

*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/resume-term-currency-standard*