Candidate Experience Software: Fix What Applicants Actually Feel

The numbers behind this

What does candidate experience software actually do?

It instruments the hiring process from the candidate's side and gives you something to fix. Three functions matter. Measurement: application completion rates, time to first response, time between stages, and survey results from people who were hired and people who were not. Automation of the promises humans forget, such as status updates, interview confirmations and rejection notes. And visibility for the candidate, usually a portal or a status page that answers the question nobody wants to email about. Most of this lives inside an applicant tracking system rather than in a separate product, and buying a standalone tool to compensate for a system that will not send a status update is rarely the right order of operations. Start by finding out where people actually leave. The fix is usually a workflow change and a template, not new software. When it is software, it is the tracking system underneath.

Where do candidates actually drop out?

At the application form, long before an interview. SHRM, citing Appcast data, reports that 92% of applicants drop off before completing an application. SHRM also reports that 51% of candidates abandon long applications, and an InFlight audit cited by SHRM found the average online application takes 51 clicks to complete. Those three numbers describe the same failure from different angles: the form asks for more than a stranger will give at first contact. Shortening it is a careers page and application-flow job. The second drop-off point is silence after applying. The third is the gap between an interview and any news about it. None of these are candidate-quality problems, though they get reported that way. Look at your own funnel before accepting anyone's benchmark. Take application starts, finishes, and the number contacted within five working days. The worst of those ratios is where to start.

How fast does a response have to be?

Faster than the candidate expects, which in practice means an acknowledgement immediately and a human decision within a working week. The acknowledgement is trivial to automate and still gets missed. What matters more is the second message, the one that says yes or no. Talent Board's CandE research puts the share of applicants who never hear back after applying at 36%, a bar a single workflow rule clears. Set a service level per stage and let the system enforce it: applications reviewed within a fixed number of days, feedback after an interview within a fixed number of days, and an alert to the recruiter when a candidate has been sitting untouched. Publish those targets internally so hiring managers know what they signed up for. A stage-level clock is more useful than an overall time-to-hire figure, because it names the person who can fix the delay.

What should a candidate be able to see about their own application?

Where they are, what happens next, and roughly when. That is the whole requirement, and most processes fail it. A candidate who can see that their application is under review, that two interviews follow, and that the team responds within a week does not email to ask, and does not assume the worst when a fortnight passes. Give them a status page or portal tied to the application, with plain-language stage names rather than internal codes. Send an update when the stage changes, automatically, so it does not depend on a recruiter's memory on a busy Friday. Any modern hiring system can do this out of the box. Where a process genuinely stalls, say so: a short note explaining that a decision is delayed by a week costs nothing and prevents the silence candidates read as rejection. Honesty about a slow process beats a fast process nobody described.

How do interview logistics shape the experience?

More than the interview questions do. Candidates judge a company on whether the meeting started on time, whether the interviewer had read anything, whether the link worked, and whether they knew who they were meeting. Scheduling is where most of the damage happens, because it is the step that involves three calendars and a stranger's working day. Self-scheduling links, timezone-correct invitations and a clear agenda remove most of it, and dedicated scheduling tooling covers the mechanics in detail. Two smaller things matter more than they look. Send the interviewer names and roles in advance so candidates can prepare. And tell people what format to expect, including whether there is a task, how long it takes and whether it is paid. Surprise exercises sprung on a candidate at 6pm are a common complaint, and they are entirely avoidable.

How do you reject someone without damaging your brand?

Quickly, specifically enough to be useful, and from a person where the stage warrants it. Rejection is the experience most candidates get, so it deserves more design attention than the offer letter. Greenhouse's 2025 Workforce & Hiring Report found that 72% of job seekers say a bad experience changed their view of a company, and rejection is where most candidates form that view. Set the rule by stage. Applications can be declined by template, provided it goes out in days rather than never. Anyone who took a call deserves a note that says where the process stopped. Anyone who reached a final stage deserves a conversation, and an honest answer about whether to apply again. Avoid feedback that is really a legal hedge dressed as encouragement. And ask the ones you liked whether they want to stay on file, because today's rejection is next quarter's shortlist.

How do you survey candidates without fooling yourself?

Ask the people who were rejected, not only the people you hired. A survey sent to new starters measures how pleased someone is to have won, which is not the same as how your process felt. Send a short survey at two points: shortly after the application decision, and after the final outcome. Keep it to a handful of questions, one of them open. Make responses anonymous by default and say so, because a rejected candidate who thinks a bad rating hurts a future application will tell you it was all lovely. Segment results by stage reached, by role family and by source, since an average across everyone hides the one broken step. Expect low response rates and treat the free-text comments as the useful part. Then close the loop internally: publish results to hiring managers with the drop-off numbers beside them, and pick one fix per quarter.

Which candidate experience numbers are worth reporting?

Five, and they should fit on one slide. Application completion rate, meaning finishes divided by starts, which is the cheapest number to improve. Time to first human response after applying. Time from final interview to a decision communicated. Percentage of candidates who received any outcome at all, which should be 100 and rarely is. And candidate survey score split by rejected against hired, because a single blended figure flatters everyone. Add offer acceptance rate if you want a commercial argument, since declined offers are the expensive end of a poor process. Report these next to the operational figures in your recruiting metrics set so nobody treats experience as a separate soft topic. The point of the reporting is not the dashboard. It is being able to say which single step lost the most people last quarter, and what changed as a result.

Candidate experience by stage: what breaks and what to measure