The three hidden revenue drains quietly bleeding placements out of your staffing firm, and how the best agencies are sealing them.

Every staffing leader knows the reflex: growth is down, so we go find more leads. More job-board spend. More resumes. More outreach. But in a recent Ringover panel, three operators who spend their days inside agency data made a sharper case, that the biggest gains right now aren’t at the top of the funnel at all. They’re in the leaks.

“It’s actually thinking about the leakage,” as the discussion opened. “Where’s the revenue draining through the business, through the agency?” The panel mapped three of them: a data leak, a network leak, and a communication leak. Plug all three and you grow without adding a single new source of leads.

The voices:

Here’s what they surfaced.

Leak 1: The data you’re already sitting on

Every agency has spent years pouring conversations, texts, emails, and resumes into its ATS. Very little of it is actually usable. Mark Hummel’s team runs free data audits across staffing firms, and the headline number is brutal: 78% of engaged candidates are missing a data field that the recruiting team actively wants to search on. You’ve already talked to these people. You just can’t find them again.

That’s the real moat hiding in plain sight. “Your competitive moat as a staffing agency is that you’ve had all these conversations and you uniquely know something about these candidates that your competition does not,” Hummel argued. The problem is that this knowledge gets buried in unstructured notes nobody mines.

And bad data doesn’t just slow you down; it actively costs you. The panel put hard numbers on the damage from spray-and-pray outreach:

  • Firms that allow mass texting see 3x the objection rate per response compared with those that don’t.
  • Generic, introduction-based SMS gets 5x the opt-out rate of personalized outreach at scale.

Every opt-out is a candidate you paid to acquire and just torched. The flip side is just as measurable. Using real pre-screen data points in a re-introduction message drove a 48% lift in meaningful responses, and agencies capturing call transcripts saw a 24% improvement in their engaged-to-submitted ratio.

That last point reframes a debate a lot of agencies are still having. The objection to logging full call transcripts has always been “that’s a lot of data just sitting in a note.” Not anymore. With tools that pull the highlights out automatically, Hugunin’s framing lands: every call is “data exhaust,” and captured properly, that exhaust is worth a fortune. “You should never have to go to the notes tab again.”

It also explains why so much AI in staffing underwhelms. When CEOs report “no meaningful ROI” on AI, the culprit is usually the foundation underneath it. Matching tools, Hummel noted, are often scoring against data that doesn’t exist for at least half the records, so they fall back on noise like “Microsoft Office” and “team player.” Fix the data hierarchy first and the same matching engine performs in a completely different league. It’s not who buys the best AI. It’s who has their data ready to feed it.

Leak 2: The network you already built

If the data leak is about what you know, the network leak is about who you already know, and keep paying to re-acquire.

David Folwell shared an audit from a $100M healthcare staffing agency: 15% of the placements they sourced from job boards were already sitting in their own database. They were paying full freight for the exact same talent twice. Pure, avoidable waste.

The fix everyone nods along to, referrals, is also the one everyone under-uses. Why? Folwell’s answer was refreshingly honest. It’s partly systemic (placements now happen over text and email, so recruiters never build the rapport that earns an ask), partly friction (a referral creates an extra step versus working a stack of inbound submissions), and partly culture. His one-question diagnostic: “You say you’re built on referrals, but do you bring up who got the most referrals on your weekly recruiter call?” Almost nobody does.

The economics make the neglect painful:

  • Agencies that build a referral program with cohesive automation (email, text, web) see a 20% to 40% jump in referral volume in year one, and it compounds year over year.
  • Staffing Referrals’ average customer gets 23% of starts from referrals; the top 5% get 41%; a few $100M agencies run over 50%.
  • Referred talent simply stays longer. Comparing candidate lifetime value, referrals showed a 78% increase in days worked versus a traditional job-board hire, because people referred in tend to work alongside, or near, a friend, and redeploy at far higher rates.

The order matters, though. “Automation is the fire on top of a good reputation and a good culture around referrals,” Folwell cautioned. Start with a dead-simple program every recruiter can explain, pay people out on time and without being chased, and build “referable moments” into the workflow. The instant someone’s placed, when a candidate becomes a promoter, any high point with the brand. Then layer automation, shareable links, dashboards, and even influencers on top.

And this isn’t just a candidate-side play. Hugunin pointed out the same physics govern sales. Since shifting Ringover’s US motion from a heavy inbound engine toward partner-sourced introductions, partner leads close at 50% to 60%, versus 15% to 20% for inbound. One vendor who’d been happy for three years referred them into two other professional baseball teams. “It’s all about who we know; use your networks. Don’t just be looking for net-new.”

Leak 3: The communication you can’t see

The third leak is where the rubber meets the road, and where good data and a warm network either convert or quietly die.

The single biggest mistake, per Hugunin: forcing everyone to communicate the same way. “We get stuck in ‘you make 100 calls.'” His alternative is what he calls “all bound,” let reps and candidates connect on the channel that actually works for them, then capture every touch back into the system. A nurse mid-shift won’t answer a call; a developer won’t either. Healthcare skews to text, IT to LinkedIn, and the right time of day differs for every cohort. Mandating volume over fit doesn’t just underperform. It pushes reps onto personal phones, and that data vanishes forever.

He described the perfect cautionary tale: a rep handed a rigid “100 calls a day” KPI became a power-dialer, firing off 100 one-second calls. It hit the number, hurt the brand, and got the firm’s line flagged as spam by the carriers. The metric was met; the business was harmed.

What works instead is measurable. A Bullhorn impact study the panel cited found that simply capturing communication data cut time-to-placement by 12% and lifted submissions/placements by 18%, nearly a 30% combined swing from visibility alone. Another finding: teams using a connected stack sourced 7% more leads from their own internal database instead of defaulting to external sources for talent they already had.

The future Folwell sketched is signal-based outreach, “fishing with a bobber instead of throwing the line in and just reeling it in.” Reach out because something happened, with a message personalized from what you actually know. As AI makes personalized outreach at scale cheap for everyone, the winners will be the firms reaching out on the right channel, at the right moment, with the right context, not the ones blasting the most messages.

The thread tying it all together: AI, used with discipline

AI ran underneath every leak, and the panel was bullish and clear-eyed.

The upside is real and measured. Agencies running 3+ AI-enabled recruiting processes are growing meaningfully faster than those with around two. New-hire cohorts that lean on AI throughout the day showed a 10x increase in 90-day productivity, because instead of learning by osmosis next to a senior recruiter, they ask their own data the questions they’d have asked that mentor. And the leverage extends to management: with AI surfacing insights, one manager can oversee 20 recruiters instead of 10.  A direct hit to the bottom line.

But the guardrails were just as emphatic:

  • Data foundation first. AI on broken data produces confident nonsense.
  • Stop over-engineering it. The era of 370 if/then prompts is over; state your goal in a sentence and let the model work.
  • Fact-check everything. One CTO rebuilt an entire sales process on an AI-generated model expecting +10%, and triggered $3M in churn, losing more than 10% instead. (A reminder that a large share of what models train on is unvetted internet opinion.)
  • Demand honesty. Push past the sycophancy. “Be brutally honest with me. What did I miss? What was the subtext?”

The most exciting use the group landed on: run prompts across your top 5% of recruiters’ call transcripts to find what they actually do differently, when and how they ask for referrals, the micro-moments that never make it into a training doc, and build the DNA of a high-performing recruiter you can coach the rest of the team against.

The 90-day playbook

Asked what they’d fix first if handed a 40-person agency tomorrow, the panel was unanimous on the shape of it: measure, then move.

  • Find what your best people actually do (Hummel). Before buying anything, research your own ATS. Where does your top performers’ real process diverge from the one you’ve documented? Turn that into data points and make those your assumptions.
  • Start with measurement by channel (Folwell). Look at candidate lifetime value and profitability per channel, then allocate toward what’s actually profitable, usually more referrals, more direct relationships. And stop paying twice for talent you already have.
  • Get a holistic view of your communication infrastructure (Hugunin). “You can’t optimize what you can’t see.” Capture everything, review weekly rather than annually, and operate like a startup making fast, progressive changes. “Perfect is the enemy of good enough.”

The bottom line

The takeaway from the hour was a quiet challenge to the industry’s default setting. As the session closed: it’s not always about finding new leads. It’s better use of your data, better activation of your network, and better execution in your communications.

The leads you’re chasing are expensive. The placements you’re leaking are already paid for. Seal the three leaks first.

This article distills a live panel hosted by Ringover (moderated by Jason Lawson) with Mark Hummel of Toro, David Folwell of Staffing Referrals, and Erich Hugunin of Ringover. 

Ringover helps staffing firms capture every conversation across calls, SMS, and channels — and turn that communication data into placements. Learn more →

RELATED ARTICLESMORE FROM AUTHOR