Source of Hire

How to track source of hire

The goal is one reliable source value on every hire's record in your applicant tracking system. That sounds easy and almost never is.

Use a short, fixed list of sources, grouped into categories. A practical list for most employers: job boards (with the specific board as a sub-source), programmatic job ads, career site, employee referral program, recruiter sourcing, agency, hiring events and job fairs, campus, social media, rehire, internal move and walk-in. Fifteen options is plenty; 60 means no one picks the right one.

Capture it automatically wherever possible. Tracking links with source parameters on every job ad, referral links tied to the referring employee and agency submissions logged through a portal are far more accurate than a 'How did you hear about us?' dropdown. Keep the dropdown as a backup and to catch channels that tracking cannot see, such as word of mouth or a radio ad.

Decide on an attribution rule and write it down. The most common is first touch: the channel that first brought the candidate into your pipeline. Some teams also record the last touch, for example a candidate who first saw a job board ad but applied after a recruiter text. Whatever you choose, apply it consistently.

Source effectiveness: going beyond the share of hires

Share of hires alone tells you where volume comes from, not which channel is best. A good source of hire report adds four columns per channel:

1. Applicants and yield. How many applicants each source produces and what percentage become hires.
2. Cost per hire by source. Channel spend divided by hires from that channel. Combine this with cost per applicant to see whether a channel is cheap at the top of the funnel but wasteful further down.
3. Speed. Average time to hire for candidates from each source.
4. Quality. 90-day retention, performance or ramp-up for hires from each source.

The last column usually changes decisions. Referrals and rehires often show higher retention than job boards even when they look more expensive per hire, and some high-volume job ad channels turn out to bring many applicants who never show up for an interview.

Worked example: reallocating a contact center budget

A BPO hires 1,000 agents a year for two US sites and spends $400,000 on sourcing. The talent acquisition manager, Chris, builds a source report for the last 12 months:

• Programmatic job ads: 520 hires, $260,000, 90-day retention 55%.
• Employee referrals: 180 hires, $90,000 in bonuses, 90-day retention 78%.
• Community hiring events: 150 hires, $30,000, 90-day retention 66%.
• Social media ads: 150 hires, $20,000, 90-day retention 52%.

Cost per hire looks cheapest for social and events. But cost per hire still employed at 90 days tells a different story: programmatic about $909, referrals about $641, events about $303, social about $256.

Chris keeps programmatic for volume, adds two more community events a quarter and raises the referral bonus from $500 to $750 paid in two halves (at start and at 90 days). Over the next two quarters, referral volume rises by about a third, and overall 90-day retention moves from 60% to 65%.

Why source data is often wrong

Most source reports carry errors large enough to change decisions. The common causes:

• Self-reported dropdowns. Candidates pick the first option, or choose 'Indeed' because they recognize the name, when they actually came from a referral.
• Job board aggregation. A job posted on your career site is scraped and shown on several boards; the candidate clicks through from one, but the ATS records 'career site'.
• Recruiters overwriting sources. A recruiter adds a candidate found in the database and labels them 'sourced', erasing the original channel.
• Referrals not logged. Employees tell a candidate to apply directly instead of using the referral link, and nobody gets credit.
• No source on hires imported from other systems, or on walk-ins entered later.

Audit a sample of 50 recent hires each quarter by asking them directly how they found the job. If more than one in five records is wrong, fix the capture process before acting on the report.

Using source of hire in high-volume recruiting

For volume hiring, review source data monthly, not yearly. Channel performance shifts quickly with seasons, local competition and ad pricing. A job board that worked in March can stop producing show-ups by July when a new distribution center opens nearby and posts higher pay.

Set a budget rule tied to outcomes, for example: shift spend toward channels whose cost per 90-day retained hire is below the site average, and cap or pause channels more than 50% above it for two months in a row. Keep a test budget, around 10% of spend, for new channels so you do not lock in a mix that worked last year.

Also look at fairness by source. If one channel produces a much less diverse applicant pool, leaning harder on it can narrow your pipeline without anyone intending it.

Common mistakes

• Reporting share of hires without cost, speed or retention.
• Mixing first-touch and last-touch attribution in the same report.
• One company-wide report for very different jobs.
• Cutting a channel based on one bad month.
• Treating 'Other' as a valid answer for more than a few percent of hires.