LinkedIn's 4.77% Engineer Reply Rate Is Below Its Own 13% Floor

Every technical recruiter I talk to has the same quiet panic: the InMails aren't landing, the credits are burning, and finance wants to know why the per-hire number keeps climbing. The 2026 numbers are worse than the vibe. LinkedIn's software vertical is now replying at 4.77%, roughly a third of the 13% floor LinkedIn itself uses to threaten Recruiter seats with suspension.

The 4.77% number, and why it breaks LinkedIn's own policy

LinkedIn's software and SaaS vertical replies to InMails at 4.77%, while LinkedIn Recruiter's own policy requires 13% on any 100-InMail, 14-day window. The average tech sourcer is structurally in violation of the platform's quality rules every time they run a real engineering campaign.

The floor is documented in LinkedIn's Recruiter Help center: "Recruiters must keep their InMail response rate at or above 13% on 100 or more InMail messages sent within every 14-day assessment period." Fall below it and the punishment is graduated: a two-week warning, a seven-day grace period, then bulk InMail is disabled on the seat for 14 days and reassessed.

The 4.77% figure comes from Belkins' 2026 outreach study, which puts software and SaaS at the bottom of every vertical it measured. For context, Legal and Professional Services replies at 10.42%, more than double. Belkins attributes the SaaS gap to "message fatigue, particularly from automation-heavy approaches," which is a polite way of saying engineers have learned to swipe left on templated LinkedIn messages before the preview finishes loading.

4.77%

The mechanism is not mysterious. LinkedIn's 2024 talent blog reported that engineering and sales roles receive more InMails than any other function, and the top three in-demand titles on the platform are software engineer, JavaScript developer, and salesperson. Volume in, saturation out. The 13% floor was designed as a bad-recruiter tripwire back when senders competed for attention. It is now the ceiling nobody clears.

What an overage credit actually costs after October 2025

Additional InMail credits went from $3 to as much as $21 after LinkedIn's October 2025 pricing change, a 700% jump on the sticker price. Most enterprise buyers negotiate down to $8 to $12, with $10 as the modal number, but $21 now anchors every renewal conversation.

Run the arithmetic against a 4.77% reply rate and the picture gets grim quickly:

Even at negotiated $10, you are paying north of $200 for a single "yes, tell me more" from a senior engineer, before a recruiter has spent a minute on the phone. And the reply is not a callback, it is just a reply. Include "not interested," "I have a job," and "please remove me," and the cost per genuinely interested engineer is a multiple of that.

There is a second squeeze most buyers miss. LinkedIn also capped Open InMail sends (the free ones to profiles marked open to opportunities) to under 100 per month per account, down from roughly 800. That is an 87% cut to free outbound capacity, and every message it displaced now needs a paid credit. The pricing change was not a spam filter. It was a channel-mix trap that pushed volume from a free bucket to a $10-to-$21 bucket at the exact moment reply rates hit historic lows.

The refund policy penalizes the worst targeting

LinkedIn refunds a credit if the recipient replies within 90 days, including replies that say no. That sounds generous until you notice it inverts the incentive: the recruiter with the highest reply rate gets the most credits back, so effective cost per credit drops as targeting improves. At 4.77%, roughly 95 out of every 100 credits burn permanently. The buyers who most need relief get the least. The "credits refunded" line item in your contract is a rebate for people who did not need it.

Hiring Assistant made the shortlist smaller, not the pipeline bigger

LinkedIn's Hiring Assistant, generally available since September 2025, is marketed as reviewing 62% (and now, per LinkedIn's own product page, 81%) fewer profiles per role. That efficiency claim hides a systemic effect: every Hiring Assistant customer converges on the same "AI-approved" shortlist, so the same handful of engineers gets ten InMails in a week from ten different companies running the same tool.

LinkedIn cites Siemens for the 62% figure and Expedia Group for a 30-day cut in time-to-hire, alongside 66 to 69% higher InMail acceptance versus manual sourcing. The single-customer wins are real. The market-level second-order effect is that the review funnel narrows for everyone at once, which drives up outreach density on a smaller pool, which pushes reply rates down further. The tool is deflationary for the individual buyer in month one and inflationary for the market in month three.

LinkedIn raised the price of a message 7x while shrinking the audience willing to reply. That is not a quality initiative. It is a toll booth.

If you sell to LinkedIn's ecosystem, this is fine. If you buy from it, you are paying more per message to reach a set of engineers that is getting smaller and more fatigued by the week.

The rare-skill pool is now too small to spray

For specialist stacks, the volume-plus-InMail model has already collapsed mathematically. In Refolk's index, only 1,188 US-based Senior engineers list Rust as a skill, clustered at Cloudflare, SpaceX and xAI, Figure, and OpenAI. At a 4.77% reply rate, InMailing the entire population yields about 57 expected replies, and you would trip the 13% floor long before you finished.

The same index shows roughly 528,895 US professionals currently holding a Software, Senior Software, or Staff Software Engineer title, concentrated at Google, Microsoft, Figma, Datadog, Ashby, and Omada Health. The US-to-Germany ratio for the same titles is about 14.5 to 1 (528,895 versus 36,542), so US teams sourcing German engineers are working from a pool where a sub-5% reply rate is even more punishing per campaign.

The lesson from the Rust column is not "Rust is hard." It is that any pool under a few thousand people cannot survive a 4.77% response rate as a sourcing strategy. You need to find people who are not being InMailed at all, and reach them somewhere other than InMail. That is the exact gap Refolk closes: describe the person in plain English (say, "US-based senior engineers writing async Rust on GitHub in the last 12 months, not currently at Cloudflare") and get a ranked shortlist pulled from GitHub, LinkedIn, and the open web, not just the slice Hiring Assistant is willing to show you.

Where engineers actually reply in 2026

Engineers reply where outreach is specific, technical, and off LinkedIn: GitHub, warm intros through former colleagues, conference talks, and open-source maintainer networks. The ranking is not about channel novelty; it is about outreach density. Any place where the average engineer receives fewer than five recruiter messages a month still works. Any place they receive fifty does not.

Concretely, this is where senior sourcers have moved budget:

• GitHub-first outreach. Pull commit history, filter by language and recency, message the email on the commit. Reply rates in the 20 to 40% range are common because engineers know you actually read their code.
• Warm intros through prior teammates. Two or three ex-colleagues at the target company beat any InMail template. A one-line intro from a former Datadog or Figma coworker lands where a cold message does not.
• Conference speakers and OSS maintainers. Public signal, low outreach density, high signal-to-noise. Someone who gave a KubeCon talk on distributed tracing is a better outbound target than 400 engineers who list "distributed systems" on a profile.
• Alumni networks of concentrated employers. Google, Microsoft, Figma, Datadog, Ashby, and Omada Health are the top current employers in Refolk's US software engineer index. Ex-employees of these companies are the pool most tech recruiters underweight because LinkedIn's title search does not reward "left 4 months ago."

For teams that want to keep LinkedIn Recruiter for its ATS integrations but stop burning credits on cold InMails, the practical move is to use it as a verification layer, not a discovery one.

What a 2026 tech sourcing stack actually looks like

A working stack in 2026 replaces LinkedIn as the discovery engine and keeps it, if at all, as a directory. The bill of materials:

1. Discovery layer. GitHub, open web, professional index. This is where you find the person. Refolk sits here: ask in plain English, get people back across GitHub, LinkedIn, and the open web.
2. Enrichment layer. Email finder, current-employer verification, GitHub-to-LinkedIn stitching.
3. Outreach layer. Email plus, where relevant, GitHub messages or Twitter/X DMs. InMail becomes the last resort, not the first.
4. CRM or ATS. Whatever you already have. Ashby, Greenhouse, Gem, Lever. The sourcing tool feeds it, does not replace it.

The point is not to fire LinkedIn Recruiter tomorrow. It is to stop treating InMail credits as the primary channel when the platform's own numbers say the channel does not work for your role family. Recruiters targeting sales, legal, or professional services can defend the spend at 8 to 10% reply rates. Recruiters targeting software engineers, at 4.77%, cannot.

The 90-day move

If you manage a tech sourcing budget, three things are worth doing before your next Recruiter renewal:

• Audit reply rate by role family, not overall. Your executive search reply rate is masking your engineering reply rate. Pull the last 90 days by function and look at each independently.
• Reprice per reply, not per credit. Take your fully loaded credit cost and divide by function-level reply rate. Present that number to finance. It is almost always 5 to 20x what leadership assumes.
• Move 30% of engineering outbound off InMail. GitHub-first outreach, warm intros, and open-web discovery. If reply rate on that 30% comes back at 15%+, the case for a bigger shift makes itself.

The 4.77% number is not going to recover. LinkedIn has raised the toll, shrunk the free lane, and pointed every buyer at the same AI-narrowed shortlist. The recruiters winning engineering roles in 2026 are the ones who noticed and stopped paying $210 a reply to compete on the same 1,188 names.

FAQ

Is LinkedIn actually suspending Recruiter seats for missing the 13% floor?

Yes, though enforcement is graduated rather than immediate. LinkedIn's Recruiter Help center describes a two-week warning, a seven-day grace period, then a 14-day suspension of bulk InMail sending followed by reassessment. Most enterprise seats never get all the way through because sourcers stop sending as soon as the warning fires, which is precisely the behavior LinkedIn wants. The practical effect is not mass suspensions; it is that engineering sourcers self-throttle and end up with fewer, more cautious messages, which does not fix the underlying reply-rate problem.

Why has the SaaS reply rate collapsed to 4.77% specifically?

Because software engineers get more InMails than any other function on the platform, per LinkedIn's own 2024 talent blog. Add the surge of AI-drafted messages from tools like Hiring Assistant, which converge on the same shortlist, and outreach density on any given engineer has multiplied while message quality has flattened. Belkins' study names "message fatigue, particularly from automation-heavy approaches" as the primary driver, and the math tracks: the top three in-demand titles on LinkedIn (software engineer, JavaScript developer, salesperson) are the same three sitting at the bottom of reply-rate leaderboards.

What are the realistic LinkedIn Recruiter alternatives for sourcing software engineers?

Discovery tools that work off GitHub, the open web, and other public professional signal rather than LinkedIn's title index. Refolk fits here: describe the engineer you want in plain English and get a ranked shortlist across GitHub, LinkedIn, and the open web, so you can reach out on the channel where the person actually replies. Other adjacent tools include HeroHunt, Leonar, and Pin. The common thread is that they surface people who are either not on LinkedIn or are on it but not actively responding, which is where the 4.77% problem gets solved.

Should I just cancel LinkedIn Recruiter?

Probably not, but you should reprice it. Recruiter still has value as a verification layer, a directory, and, for non-engineering functions, a viable outreach channel. What has broken is treating it as the primary discovery and outreach tool for software engineering. Cut engineering-role InMail volume, shift discovery to GitHub and the open web, and keep the seats for verification and for the functions where reply rates still clear the 13% floor. Your per-hire number will move within one quarter.