Why senior-engineering reply windows define the cadence

The reason a "good" tech-staffing cadence looks so different from a general-outbound cadence isn't the channel mix or the personalisation depth. It's when the copy arrives — and what the engineer is weighing when it lands. Senior engineers reply between comp refreshes, around quarterly all-hands, and in the half-hour after a Series-B announcement hits their work Slack. Designing around comp-refresh and equity-revaluation windows is what makes the cadence. The 3-7-12-21 spacing is downstream of that timing choice.

The 3-7-12-21 spacing — and why each gap matters

Touch 1 (day 0). Short, technical opener — under 110 words. Name the company, the role, one stack-specific detail, and the comp band. Engineers self-select on stack and comp before anything else.

Touch 2 (day 3). Two-sentence bump that adds a single new fact; repeating the opener changes nothing.

Touch 3 (day 7). A structured re-angle, not a re-pitch. We rotate through six angles — pain framing, results, case study, contrarian, referral redirect, equity-window — because picking one and repeating it produces diminishing returns. The strongest single angle at touch 3 is the equity-window re-anchor: "the equity grant on this role vests on a four-year schedule with a one-year cliff, and the Series-B extension is handled in the next 14 months." Naming the equity window directly converts at higher rates than the generic opener because it addresses the engineer's actual decision frame.

Touch 4 (day 12). A contrarian or results-brief re-angle that breaks the pattern. By day 12 the engineer has had two opportunities to engage on their own timing — a third restatement produces zero lift.

Touch 5 (day 21). Low-pressure permission reopen. A one-word yes/no ask converts higher than you'd expect because the social cost of saying yes is one word and the social cost of saying no is more effort than the engineer wants to spend.

Stack-specific openers and the credibility move

Once timing is right, the opener has to clear the engineer's first filter — which for senior engineers is stack specificity.

Backend / platform. "Senior Platform Engineer — Go / Postgres / Kubernetes, Series-B fintech, $220-280k base + 0.15-0.30% equity" outperforms the generic opener by roughly three times on first-reply rates.

Full-stack / product. "Staff Engineer — TypeScript / React / Node, marketplace company, $240-300k base + equity refresh" works because the equity-refresh mention tells the recipient the role lands inside their next comp cycle, not deferred to a year-three re-evaluation.

Infra / SRE. "AWS" alone reads as generic; "Terraform + AWS + observability stack" is the credibility move that says the recruiter has done the homework.

ML / data. "PyTorch" alone reads as a blast; "PyTorch + production inference" tells the engineer the role ships models rather than just trains them.

Both moves live in the templates we ship in the tech follow-up pack, alongside the six touch-3 re-angles and the permission-reopen breakup.

Bounce and re-engagement patterns for engineering lists

Engineering lists bounce more than general-outbound lists. Senior engineers route inbound to Gmail's "Promotions" tab aggressively — the most common failure mode isn't mail-server blocks, it's HTML filter. A pure-email sequence loses 6-10% of senior engineers to Gmail filter over a 14-day window on cold data, plus another 4-7% to personal-domain blocks.

The recovery pattern: switch channels rather than retry the same one. An email in Promotions becomes a LinkedIn DM referencing the role. A bounced personal-domain email becomes a re-send from a different sending domain with the role line moved to the subject. For cold engineering lists where the channel is unclear, we layer a LinkedIn touchpoint at day 3 — short, under 220 characters, referencing the role from the prior email without restating it. The redundancy earns the inbox-filter break without producing spam signals.

Reply-rate benchmarks, broken out by tech sub-vertical

Here's where the comp-windowed 3-7-12-21 cadence lands across roughly nine months of tech-staffing sequences at Outbound clients.

Backend / platform. First-reply 11-15%, comp-windowed timing pulls another 5-8%. End-of-sequence cumulative: 22-28%.

Full-stack / product. First-reply 13-18%, windowed timing adds 4-6%. End-of-sequence cumulative: 26-32%.

Infra / SRE. First-reply 9-13%, windowed timing adds 4-7%. End-of-sequence cumulative: 20-26%.

ML / data. First-reply 7-11%, windowed timing adds 5-7%. End-of-sequence cumulative: 17-23%.

The common thread: comp-windowed timing moves every range up by 18-25%, and most of the gain comes from the first touch landing in a comp-refresh window.

Where to go from here

If you're running a tech desk and want the whole cadence with timing, comp-window targeting, copy, and channel mix wired up, we build that end to end. The full run-through is on the staffing vertical page, with tiers on /pricing.

If you'd rather drop tech-specific follow-up templates into your own sequences, the tech follow-up pack ships 25 tech-tuned openers, single-fact bumps, the six touch-3 re-angles, the permission-reopen breakup, and bounce-recovery templates covering backend, full-stack, infra/SRE, and ML/data placements.