286 Minutes on the Phone. Tech CPH Doubled. Same Story.

The ASA/Prodoscore report dropped on June 4 with a number designed to make recruiting leaders feel good: 286 minutes on the phone per week, the highest on record, and recruiters now running an average of 1.36 AI tools. A week later, Appcast's 2026 Recruitment Marketing Benchmark made everyone feel less good: job-board cost-per-hire jumped across the board, and for technology roles it more than doubled in a single year. These are not two stories. They are the same story, told from two ends of the funnel.

What the 286-minute number actually means

Recruiter call time 2026 is up roughly 2x from Q1 2024. The ASA framing, echoed by Prodoscore CEO Sam Naficy, is that automation has freed recruiters to do the high-judgment human work: more conversations, deeper relationships, better placements. That sounds nice. It is also a misread of the data.

Look at what else moved over the same window. Ashby reported in May that applications per open role have tripled since 2021, now above 300. Monster found 77% of US candidates worry their resumes will be filtered out before a human reads them, so they pad and re-apply. And in late 2025, LinkedIn quietly capped Open InMail at under 100 sends per month, down from a practical limit closer to 800. That is an 87% drop in outbound capacity, overnight, on the channel most tech recruiters built their week around.

87%

So: inbound volume tripled, the dominant outbound channel got throttled by an order of magnitude, and the phone is now the only lever that still moves. Of course call time hit a record. It is not a productivity win. It is the sound of recruiters running out of options.

Engineering and sales were already the worst-responding cohorts

LinkedIn's own data shows software engineering and sales receive more InMails than any other functions, which means response rates sit below average even before the cap. Signal-driven, personalized outreach can still hit 15 to 25% InMail response and 30%+ connection acceptance. Generic templated outreach comes in under 5%. After the cap, the math is unforgiving: you have fewer shots, on the hardest cohort, and volume tactics are now strictly worse than they were last year.

That is why the phone went up. It is the only channel where the marginal dial still pays.

What the doubled tech CPH actually means

Appcast's 10th annual benchmark, drawn from 302M+ clicks, 27M applications, and nearly 1,200 employers, was explicit about the cause: "shifts in job board pricing and programmatic media models." Hiring slowed in 2025. Apply rates stayed high. CPH rose anyway, because programmatic auctions don't care about your headcount plan.

SHRM's 2025 Benchmarking Report pegs non-executive cost-per-hire at $5,475. ScoutLogic's 2025 analysis puts technical and engineering roles at $6,200 to $8,000, driven by candidate scarcity. The Appcast number behind the headline is that the median job-board CPH reached $1,340 across all occupations in 2025, and tech CPH more than doubled in a year. Stack those together and the picture for an engineering org is brutal: total tech CPH is climbing through both the paid-media line and the recruiter-time line at the same time.

The phone went up because nothing else converts. That is not a productivity story. It is a triage story.

The macro narrative calls this a "low-hire, low-fire" market. For tech that framing is wrong. Tech unemployment is 2.8% against a 4% national rate. Robert Half says 87% of US tech leaders struggle to find skilled talent. BLS still projects 1.2M unfilled US developer roles by 2026. The candidate-supply story and the CPH story are not in tension. They are two markets sitting inside one benchmark report.

The buried takeaway from the ASA report

The celebratory headline about 1.36 AI tools per recruiter hides the part of the ASA findings that actually matters. The StaffingHub coverage notes that firms only see revenue gains when they strategically redirect the time AI saves. Without a plan for the freed capacity, the efficiency gains vanish into more dials, more calendar bloat, and more time on platforms that no longer convert.

The same June StaffingHub brief on the ASA 2026 State of Staffing Benchmarking Report flagged something even more direct: 50% of growth agencies relied entirely on owned sourcing channels, versus just 27% of slower-growth firms. The fast firms are not buying more apply volume. They are building pipeline they own.

This is the part where the "sourcing tool ROI" question stops being abstract. If your stack does not either cut dials-per-hire or replace job-board spend with better-targeted owned outreach, it is a P&L liability this quarter. Everything else is decoration.

The new sourcing math

Here is the equation recruiting leaders should be running for Q2 and Q3.

Dials-per-hire, not minutes-on-phone

Stop benchmarking recruiters on call time. Benchmark on dials-per-hire and InMails-per-hire. A recruiter who makes 60 well-targeted calls to land a senior staff engineer is beating a recruiter who makes 240 calls to land the same hire, even if the second one logs more "productivity" in Prodoscore. The 286-minute number rewards the wrong behavior.

Signal plus identity beats volume

With InMail throttled and job boards inflating, the moat is intent signal combined with cross-platform identity resolution. That means: funding events, hiring posts, GitHub activity, conference talks, and recent commits, tied to a single resolvable person across GitHub, LinkedIn, and the open web. This is what makes a personalized first message land in the 15 to 25% response band instead of the sub-5% generic band.

This is also exactly the problem we built Refolk to solve. You describe the person in plain English ("ex-Stripe payments engineer, now at a Series B, has shipped a Rust crate in the last year") and get a ranked shortlist with the context you need to write a message that actually earns a reply. The point is not "more candidates." The point is fewer, better first dials.

Owned channels over rented ones

The Appcast finding (CPH rose because of auction dynamics, not because candidates disappeared) means every dollar moved out of programmatic job-board spend and into owned sourcing is a dollar that does not get re-bid against you next quarter. The ASA growth-agency split (50% owned vs. 27% at slower firms) is the same insight from the staffing side.

The cleanest way to read this: outbound sourcing reply rates are now the single most important leading indicator in your funnel, because they determine whether you need the phone at all. If reply rates are at 4%, the phone is the only thing keeping the pipeline alive, and call time will keep climbing. If reply rates are at 18%, the phone becomes a closing tool again, not a triage tool.

A concrete Q2 reallocation

If you run technical recruiting and you are staring at these benchmarks, here is the move that maps to the data.

1. Cut programmatic job-board spend on tech reqs by 30 to 50%. Appcast's own data says you are paying auction premiums on cohorts that are still applying anyway.
2. Reallocate that budget to sourcing tools and seats that produce identity-resolved, signal-scored shortlists. Measure them on dials-per-hire and reply rate, not list size.
3. Rebuild your InMail playbook around the sub-100/month reality. Every send needs a specific reason. Generic sequences are now strictly negative ROI.
4. Set a dials-per-hire target per req level. For senior tech roles, anything above 80 dials per offer accepted means the front of the funnel is misaligned, not that recruiters need to dial harder.
5. Audit your AI tools. 1.36 per recruiter is meaningless if none of them reduce the time between "open req" and "first qualified conversation." That is the only metric that matters.

A specific note on step 2. The reason Refolk works for the kind of req where LinkedIn Recruiter has stopped converting (staff and principal engineers, infra leads, applied ML) is that the search starts from a description, not a Boolean string, and the result includes the GitHub and open-web signal you need to write a message worth one of your remaining 100 InMails. That is the whole game now.

What the benchmarks will say in Q3

If nothing changes in the sourcing stack, the Q3 readings are predictable. Call time will climb past 300 minutes. Tech CPH will rise again, because the auction does not stop. The 1.36 AI-tools-per-recruiter number will tick to 1.5 or 1.6, and revenue per recruiter will not move, because none of those tools will have addressed the actual constraint.

The firms that bend the curve will not be the ones with the most AI tools. They will be the ones who treated the 286-minute number as an alarm and rebuilt sourcing around owned signal, identity resolution, and reply-rate as the north-star metric. That is the version of technical recruiting benchmarks Q1 2026 that will look obvious in hindsight.

FAQ

Is the 286 minutes per week recruiter call time number sustainable?

No, and it is not supposed to be. It is a coping mechanism for a broken outbound stack. The same ASA report shows recruiters running more AI tools than ever, which means the time savings exist but are being absorbed by the only channel still converting. Once reply rates on owned sourcing recover, call time should fall back toward 2024 levels, not because recruiters are doing less, but because each dial is landing on a warmer contact.

Why did tech cost-per-hire double when the macro market is "low-hire, low-fire"?

Because the macro and the tech market are decoupled. Appcast attributes the broader CPH rise to job-board pricing and programmatic auction dynamics, not to candidate scarcity. For tech specifically, you also have 2.8% unemployment, 87% of leaders struggling to find skilled talent, and the InMail cap squeezing the dominant outbound channel. Auction pressure plus a tight tech labor market plus a throttled outbound channel is exactly the recipe for CPH to double in a year.

What is a realistic InMail reply rate for engineering roles in 2026?

Generic templated sends are at sub-5% and falling. Signal-driven personalized outreach (referencing a recent commit, a talk, a funding event, or a specific project) lands in the 15 to 25% range with connection acceptance above 30%. Given the post-cap reality of under 100 sends per recruiter per month, the only sane strategy is to optimize every send for the upper band. Volume tactics are now strictly worse than they were 12 months ago.

How should I measure sourcing tool ROI given these benchmarks?

Two metrics. First, dials-per-hire and InMails-per-hire by req level, tracked monthly. If a tool does not move these numbers down within a quarter, it is not earning its seat. Second, allocated cost-per-hire net of the job-board spend you can retire. If a sourcing tool lets you cut programmatic spend by 30% on tech reqs while holding time-to-fill flat, the tool has paid for itself regardless of what its dashboard claims.