Application completion rate: the formula, how to track it and the fixes to test

Application completion rate is the share of people who start a job application and submit it: applications submitted ÷ applications started × 100, for the same jobs and period. It measures the step before "applied", which most funnel reports skip because they start counting at the submitted application. Track starts and submissions per job and per form step, fix the step where most people stop, and check that the candidates you gain still pass the first screen. As of October 2026 we found no reliable public benchmark from an original publisher, so compare against your own baseline.

What happens after the application arrives is covered in pipeline conversion rates by stage, and the wider candidate journey in how to improve candidate experience.

What counts as a start and a submit?

The rate is only as good as the two definitions, and tools count them differently. Write them down:

• Start: the first interaction with a field on the application form. Clicking "Apply" on a job ad is not a start; people who click and leave at an account-creation screen never reached the form, and that loss belongs in a separate ratio.
• Submit: reaching the confirmation page or a confirmed submission in the ATS, not pressing a button that may have failed validation.
• The unit: people or sessions. A candidate who starts on a phone and finishes on a laptop is two sessions and one person. Pick one and keep it.

The formula and the ratios around it

Apply-click rate        = apply clicks ÷ job views × 100
Form-start rate         = form starts ÷ apply clicks × 100
Completion rate         = submissions ÷ form starts × 100
Step completion (step n) = people reaching step n+1 ÷ people reaching step n × 100

Completion rate isolates the form itself. If job views rise and completion falls, the ad may be attracting the wrong people, or a form change may have broken something; the separate ratios tell you which.

Worked example: one job, one month

The biggest loss is at step 2: 130 of 430 people stopped at the resume upload. Splitting by device shows why:

Most people start on a phone, and a phone is where a resume file is least likely to be at hand. The first test is obvious: let mobile applicants upload from cloud storage or apply with contact details and a resume sent later.

How do you track it on a careers site?

If the form sits on a site that carries your Google Analytics 4 tag, the enhanced measurement option for form interactions records form_start the first time a user interacts with a form in a session and form_submit when they submit it. Google notes that the form parameters, such as form ID and name, can only be used in reports after you create custom dimensions for them, which you will need if the site has more than one form.

• Count submissions from the confirmation page where you can, so failed submissions are not counted.
• Track each step with a page view or a custom event per step for multi-page forms.
• Pass the job ID as a parameter so you can compare roles.
• Respect your cookie consent settings: visitors who decline analytics may not be counted, so treat the rate as a sample and keep the definition stable rather than chasing exact counts.

How do you track it in an ATS?

Many application forms are hosted on the applicant tracking system's own pages, where your analytics tag may not run. Check your ATS's reporting documentation for terms such as "incomplete", "started" or "draft" applications, and ask the vendor whether you can add your own analytics tag to its hosted pages. If neither is possible, measure what you can: apply clicks on your own job pages against submissions in the ATS, which gives a combined click-to-submit rate. It mixes the form with any redirect or account step, so it is a weaker diagnostic, but it still shows changes over time.

Why do candidates abandon applications?

• An account before applying. A password screen between the ad and the form loses people before the form starts.
• Re-typing the resume. Asking for work history that was just uploaded. The W3C's accessibility guideline WCAG 2.2 success criterion 3.3.7, Redundant Entry (level A) asks that information already entered in the same process be auto-populated or available to select, with limited exceptions.
• Long screening questionnaires. Every required question that would not change your decision is pure cost; see the pre-screening questionnaire template for knockout questions only.
• Assessments before submission. A 30-minute test inside the form loses people who would have taken it after a first conversation.
• Mobile failures. File uploads, date pickers and tiny fields on a phone.
• Errors that wipe entries, or sessions that time out without saving.
• Questions you should not ask at all, such as current salary where pay history questions are banned.

Which changes should you test first?

1. The step with the biggest loss in your step table, not the one with the most complaints.
2. Remove account creation before submission, or make it optional after.
3. Make re-entry optional when a resume is uploaded.
4. Cut required fields to those you would actually reject on.
5. Move assessments after submission, or after the first screen.
6. Add a progress indicator and save-and-return for long forms.

Change one thing at a time, compare at least four weeks before and after for the same jobs, and track the screen pass rate alongside completion. If completion rises and the share of applicants who pass the first screen falls sharply, you removed a filter that was doing useful work. Federal contractors should check which self-identification invitations they must still make before moving or removing those questions; see the Internet Applicant rule in 2026.

Is a low completion rate always bad?

No. Some drop-off is the form doing its job. A knockout question that tells a candidate the role needs a license they do not hold, or a stated schedule they cannot work, should end some applications, and those people are better off finding out in two minutes than after a phone screen. Separate the two kinds of exit in your tracking:

• Screened out: the candidate answered a knockout question in a way that ends the application, or chose to stop after reading a stated requirement. Count these separately and do not try to reduce them.
• Abandoned: the candidate stopped without a disqualifying answer, often at an upload, an account screen or a long questionnaire. This is the part to fix.

Compare like with like as well. Hourly roles with short forms, professional roles with longer ones, and candidates arriving from different job boards behave differently, so a single company-wide rate hides more than it shows. Read the rate by job family and source, and look at the trend within each.

What about candidates who cannot use the form?

Some people abandon because the form does not work with the assistive technology they use, or because a timed step or a required upload is a barrier for them. Put a short line on the form with another way to apply, such as an email address or phone number, and make sure someone answers it. The U.S. Equal Employment Opportunity Commission's guidance on job applicants and the ADA says an employer cannot refuse to consider an applicant because they need a reasonable accommodation to compete for a job. More on adjustments later in the process is in ADA accommodations in interviews. This is general information, not legal advice.

How should you report it?

Show counts beside every percentage, monthly, by job family and device. Rates on fewer than about 50 starts swing too much to act on. A row like this fits on the recruiter KPI dashboard.

Common mistakes

• Dividing submissions by job views and calling it completion; that is overall conversion.
• Mixing people and sessions between the numerator and the denominator.
• Counting duplicate submissions from the same candidate.
• Comparing tools with different definitions of a start.
• Maximizing completion while the quality of applicants falls.