To screen 500 applications without losing fairness, split the work into stages: a short set of must-have questions at application, a 30 to 45 second review of each CV against the same written must-haves, a second pass on the people who pass to record evidence, and quality checks on a sample of both the yes and the no decisions. Use several reviewers who have calibrated together, keep records and tell every applicant the outcome.
Key takeaways
- Write three to five must-haves before the ad goes live, and test them on 20 real CVs.
- Use short application questions only for facts that really decide eligibility, such as right to work or required licences.
- Review every application against the same must-haves. Mark yes, no or unclear. Do not rank.
- Calibrate reviewers on 10 CVs, then sample 10 percent of the “no” decisions for a second look.
- Keep records of criteria, decisions and dates, and send every applicant an outcome.
Why volume changes the problem
With 20 applications, one person can read each CV closely. With 500, closer reading is impossible, and the risk grows that fatigue, unclear criteria or a quick glance at the wrong features decide who gets seen. People also become less consistent through the day: the 400th CV is read differently from the first. A volume process needs structure that does not depend on one reviewer’s energy.
Structure has another benefit. In many places, employers are expected to show that they used job-related criteria and applied them equally. Written criteria and records make that possible. The guidance of the US Equal Employment Opportunity Commission on prohibited employment policies and practices and the Acas page on recruitment both point to fair, job-related selection.
Stage 0: Prepare before the ad goes live
- Write the must-haves. Use the role brief. Keep to three to five must-haves that can be checked on a resume or CV.
- Test them on 20 CVs. Ask two people to read the same 20 CVs against the must-haves. Where they differ, fix the wording and add an example of what counts.
- Choose the application format. Fewer fields means more applications, and also more noise. Ask only for what you need to decide whether someone is eligible.
- Plan the reviewers and the time. Estimate: 500 applications at 45 seconds each is about six and a quarter hours of reading for the first pass. Split across three reviewers, that is about two hours each, in short sessions.
- Prepare the messages. Write the acknowledgement, decline and invitation emails now, so that nobody improvises under pressure.
Stage 1: Application questions (only for real eligibility)
Short questions at application can save time, but they are a screen, and they can exclude people unfairly. Use them only for facts that decide legal or practical eligibility, such as the right to work in the country, a licence required by law, or availability for shift work if the role needs it. Phrase them neutrally, and allow a way to explain or ask for an adjustment.
| Good use | Risky use |
|---|---|
| “Do you hold a valid forklift licence?” for a role that legally requires one | “How many years have you been working?” which may indirectly screen out older or younger candidates |
| “Are you able to work Saturday shifts?” if every shift pattern includes them | “Do you have reliable transport?” which may exclude people for reasons unrelated to the job |
| “Do you have the right to work in [country]?” with a note on how to show it | “Are you a native speaker?” which points to national origin; state the language level needed instead |
Stage 2: The 45-second review
Reviewers read each CV for the must-haves only. They mark yes, no or unclear for each. They do not rank or compare. Use a sheet with one row per application and one column per must-have. This stage removes clearly ineligible applications and sorts the rest.
| Outcome of stage 2 | Rule | Next step |
|---|---|---|
| All must-haves yes | Moves to stage 3 | Record the CV line for each yes |
| At least one unclear, no no | Send one short question or check the application form again | If the answer shows the must-have, move to stage 3 |
| At least one clear no | Decline with a standard message | Record the must-have that was not met |
Treat unclear marks with care. CVs are short summaries. A candidate with strong skills may write badly about them. If many applications are marked unclear for the same must-have, the ad or the application form probably does not make the requirement clear. Fix it and tell reviewers.
Stage 3: Evidence and the short list
For applications that pass, spend two to three minutes finding the exact evidence for each yes. Copy the CV line into the sheet. This is a person-led step. Tools can help you find passages, but a person decides what counts as evidence. FastHire is being built to show each job requirement next to the exact CV passage that supports it, so a person can check the evidence quickly. FastHire is coming soon. It does not rank, score or reject candidates. People make every decision.
When you have more qualified people than you can interview, do not rank by a hidden formula. Use a neutral, written rule: random selection among all who meet every must-have, or one extra published criterion. Record the rule and apply it to everyone.
Calibrate your reviewers
- Before the screen starts, all reviewers read the same 10 CVs and mark them alone.
- Compare marks. For each difference, discuss the must-have, agree what counts and write an example.
- Repeat with 10 new CVs. Aim for the reviewers to agree on nearly all marks. If they still differ often, the must-have is too vague. Rewrite it.
- Midway, repeat with five CVs to check for drift.
Quality checks that catch errors
| Check | How | Why |
|---|---|---|
| Sample of no decisions | A second reviewer re-reads 10 percent of the “no” decisions, chosen at random | Finds errors and shows whether the criteria are applied consistently |
| Sample of yes decisions | Re-read five to ten yes decisions per reviewer | Finds reviewers who are too lenient |
| Near-miss review | Look at applicants who failed only one must-have | Shows whether that must-have is really essential |
| Time check | Compare how long each reviewer spends per CV | Very quick reviews may mean low accuracy |
| Pattern check | Look at who passes and fails by source channel and by must-have | Helps spot channels or criteria that need review |
If you track outcomes by group, such as for equal opportunity monitoring, store that data separately from decision records and follow the rules of your country. We cover this in more detail in our upcoming guides to diversity reporting and fairness monitoring.
Candidate communication at volume
- Send an automatic acknowledgement with the closing date and the date by which people will hear the outcome.
- Send a clear decline to everyone who is not moving forward. Silence damages your employer reputation and wastes candidates’ time. See our guide to recruiter outreach messages for tone.
- Offer a way to request an adjustment at any stage.
- Tell people how long you will keep their data. The UK Information Commissioner’s Office explains the duties of employers and recruiters in its guidance on recruitment and selection.
A worked example with 500 applications
This example is fictional. It shows the flow, not the results of any real company. A retailer opens 12 customer service roles and receives 500 applications in two weeks. Three reviewers calibrate on 10 CVs in 45 minutes and rewrite one must-have that was unclear. Each reviewer reads about 167 applications in stage 2 over two sessions. 310 applications have all yes marks, 60 are unclear and 130 have a clear no.
The 60 unclear applications receive a short question about availability. Forty reply, and 25 of them show they meet the must-have. That gives 335 applications for stage 3. The team can interview 60 people. Since 335 people meet every must-have, they use a published rule: random selection among all who meet the must-haves, repeated if people decline. A second reviewer checks 50 “no” decisions and 20 “yes” decisions at random, finds three errors and corrects them. Every applicant receives an outcome within the promised date. The whole screen takes about 20 reviewer hours, spread over two weeks.
Common mistakes at volume
- Changing the criteria after you see the pile. If you must change them, change them for all applications and record why.
- Screening on proxies. Gaps in employment, school names, postal codes and graduation years are poor proxies for skill and can create unfair patterns.
- Relying on a tool to decide. If you use software, a person must decide and understand the basis. Ask the vendor how the tool works. See our Boolean search guide for a search method that you can explain and check.
- Skipping the sample checks. Errors at volume are common, and a short check finds many of them.
- Forgetting the bin. Delete application data when you no longer need it, according to the rules of your country.
A screening sheet for volume
| Column | Content | Notes |
|---|---|---|
| Application ID | A number, not a name, for the first pass if your system allows it | Supports anonymous review. Link IDs to names in a separate, protected sheet. |
| Date received | Date | Helps you meet the time you promised candidates |
| Must-have 1 to 5 | Yes, no or unclear | One column per must-have |
| Evidence (stage 3) | The CV line, copied or paraphrased | Only for people who pass stage 2 |
| Reviewer and time | Initials and date | Shows who decided and when |
| Second check | Yes or no, with a note | Only for sampled records |
| Outcome and message sent | Invited, declined, held, with the date | Closes the loop with the candidate |
| Delete by | Retention date | Supports data protection rules |
A reviewer time budget
Use arithmetic to plan before you start. This table shows how to calculate the time for 500 applications. The numbers are planning assumptions, not benchmarks. Replace them with your own.
| Stage | Time per application | Applications | Total hours |
|---|---|---|---|
| Stage 2: first review | 45 seconds | 500 | 6.25 |
| Unclear questions and replies | 2 minutes | 60 | 2.0 |
| Stage 3: evidence notes | 3 minutes | 335 | 16.75 |
| Second checks on a sample | 1 minute | 70 | 1.2 |
| Messages to everyone | Templates; mostly automated | 500 | 2.0 (checks and edits) |
The total is about 28 hours, which three people can share over two weeks. If the total looks too high, you can cut stage 3 for the group that you will not interview, but only after you apply the written selection rule. Do not skip evidence notes for people you invite.
If your must-haves knock out too many people
Sometimes the pass rate is very low, and you wonder if the criteria are too strict. Check the near misses: look at 20 applicants who failed one must-have. If many of them clearly could do the job, the must-have may be a preference. Move it to nice-to-have with the hiring manager’s agreement and write the reason down. Then re-apply the changed rule to every application, not only to the ones you saw. If a must-have is unavoidable, such as a legal licence, explain this in the ad so that fewer unsuitable people apply.
Frequently asked questions
How long should I spend on each CV in a first pass?
About 30 to 45 seconds, against written must-haves only. The aim is to sort, not to judge the whole person.
Is it fair to reject applications that miss one must-have?
Yes, if the must-have is job-related, written down before screening and applied to everyone. Review near misses to check that the must-have is really essential.
Can I use software to screen at volume?
Yes, but the decision should stay with a person, and rules on automated decision making may apply. Ask the vendor how the tool works and keep your own records.
How many reviewers do I need?
Enough to finish in a few short sessions. Three reviewers who calibrate together are better than one tired reviewer.
What do I do if I have too many qualified candidates?
Use a neutral, written rule such as random selection among all who meet every must-have, or one extra published criterion. Do not use a hidden score.
Related guides
- Hiring guides for recruiters and HR teams
- How to source candidates: ten channels and when to use each
- How to screen applications quickly: a 20-minute method for small teams
- How to define a role before you hire: a one-page role brief
- All hiring guides for employers
- Structured screening: how to compare every applicant on the same evidence
- How to reduce bias in resume screening: 15 practical steps
- How to write a screening call script
- How to reduce time-to-hire without lowering quality
- Blind CV screening: does it work and how to do it
- How to build a hiring dashboard in a spreadsheet
- How to choose recruiting software: 25 questions to ask every vendor
- How to screen resumes against must-haves with evidence
- How to run a hiring day or open day
- How to hire in volume for seasonal peaks
- How to manage hiring across the US, UK and EU
- How to write fair knock-out criteria
This is general information, not legal advice.