Measuring a Disability Employment Program: What to Track

Most organisations measure one thing: what percentage of the workforce has self-identified as disabled. It is the number that gets reported, and it is close to the least informative one available.

For a separate commercial perspective on workforce analytics, see workforce analytics software; it should be evaluated independently against accessibility, privacy, and employment-law requirements.

Employers can compare inclusive-employment practices with resources from the Employer Assistance and Resource Network.

It tells you nothing about whether people can apply, whether they stay, or whether they advance — and it moves for reasons that have nothing to do with anything you did, because self-identification rates depend mostly on whether people feel safe answering.

Here is what to measure instead.

Process metrics

These are the leading indicators, and they are entirely within your control.

Time from accommodation request to resolution. The single most predictive number in this area. A figure in days indicates a functioning process. A figure in weeks predicts your retention number a year later.

Accommodation request volume. Counter-intuitively, low volume is usually bad news. It means people are not asking — either because they do not know how, or because it is not worth the cost of asking.

Proportion of requests granted, and of those refused, how many came with an alternative. A refusal with no alternative offered is a process failure regardless of the merits.

Application completion rate. What proportion of people who start your application finish it, and where they abandon. A drop-off concentrated at one step is a barrier, not disinterest.

Accessibility conformance in procurement. What share of systems bought this year had accessibility assessed before signing.

Outcome metrics

Retention at 12 and 24 months, compared with your overall rate. If people are being hired and leaving faster than average, the hiring programme is converting recruitment spend into turnover. See retention.

Promotion rate, compared. Hiring without advancement is a specific and common failure — visible only if someone looks.

Pay progression, compared.

Time to full productivity for new hires with accommodations, against the general figure. A gap usually points at onboarding rather than at people. See onboarding with accommodations.

Referral source effectiveness. Which channels produce candidates who are hired and stay. See where to find candidates.

Experience metrics

Numbers do not capture this and it is where the useful information is.

An anonymous survey question set, asked consistently:

• Do you know how to request an adjustment to how you work?
• If you have one in place, is it working?
• Did your manager support it?
• Has anything changed that means it should be reviewed?

Exit reasons, collected the same way every time and read for patterns rather than individually.

Candidate feedback on the hiring process. One optional question at the end: was anything about this process difficult? It will tell you more about your systems than any audit.

Asking creates an obligation. Asking and doing nothing is worse than not asking.

Collecting data without causing a problem

Self-identification must be voluntary, confidential, and separated from employment decisions. Aggregate reporting only, never accessible to a hiring manager or a supervisor.

Explain what it is for and who sees it. Response rates depend almost entirely on trust, which is why the number is a measure of your culture at least as much as of your workforce.

Do not chase individuals.

Keep accommodation records separate from personnel files, as medical information requires.

Federal contractors have specific obligations around self-identification and reporting, and those are currently subject to proposed regulatory change. See Section 503 in 2026.

Where automated tools are used in hiring or assessment, several jurisdictions impose specific obligations — bias audits, advance notice, disclosure. Check before deploying anything of that kind.

What the numbers usually reveal

Three patterns come up repeatedly.

Low self-identification with high accommodation volume. People are getting what they need but do not trust the reporting system. Your culture is better than your number.

High self-identification with low accommodation volume. People are comfortable saying so and not comfortable asking for anything. The gap is worth investigating.

Good hiring, poor retention. The most expensive pattern, and the most common. Effort is going into the front of the funnel while the back of it leaks.

Where to start

If you measure nothing today, start with three:

1. Time from accommodation request to resolution
2. Application completion rate, by step
3. Retention at 12 months, compared with your overall rate

The first two are within your control and improvable this quarter. The third tells you whether any of it is working.

Headcount can stay on the report. It should not be the number anyone is managing to.