Writing
Writing from the operator seat
What actually happens inside a running outbound engine: the signals worth chasing, the numbers we measure on our own pipeline, and the parts that break. Three a week, no email gate.
Articles
24
Latest
Oct 26
Months
3
October 2026
Oct 22, 2026
Connect LinkedIn to ChatGPT or Claude for prospectingWhat it takes to let your own model work a LinkedIn pipeline, measured on our door: 80 keys, 924 tool calls, 186 ms for a data read, and one step that runs on a two-day clock.Oct 19, 2026
Cost per meeting from LinkedIn outreach: our numbersWhat a cold LinkedIn meeting costs on our own pipeline: 4 110 invitations, 1 044 accepts, 267 first replies, 19 qualified conversations, and the arithmetic between them.Oct 15, 2026
AI SDR alternatives for founder-led sales: four routesFour ways to run outbound before your first sales hire, priced and measured on our own pipeline: 1 632 invitations, 27.3% accepts, 192 self-run keys, one month.Oct 12, 2026
Fractional CRO with an AI outbound engine: what you actually getAn inventory of thirty days of operated outbound: 1 563 invitations, 59 026 text checks, 282 logged decisions, and the 19 of them a person made.Oct 8, 2026
LinkedIn outreach from Claude via MCP: setup in 10 minutesWhat the first ten minutes with a pipeline key actually look like. 31 keys, 882 tool calls, a median of 24 seconds to the first call, and the one step that takes a day.Oct 5, 2026
Managed AI SDR pricing in 2026: what the fee coversFive pricing structures you will meet this year, and the numbers from 30 days of our own pipeline to put against each one: volume per identity, data, email and the human load.Oct 1, 2026
AI SDR, agency or a hire: the cost per meetingCost per meeting is a fraction, and the denominator decides it. Numbers from a live pipeline on what one sending identity produces, priced against all three routes.September 2026
Sep 28, 2026
MCP for revenue teams: read access to your own pipelineWhat happens when you point your own model at a live outbound pipeline. 777 tool calls over five weeks, the read and write split, the caps, and what the logs say.Sep 25, 2026
A/B testing outbound copy when the sample is smallFour weeks, three copy variants, 871 first messages. What a small outbound sample settles, what it leaves open, and the assignment key that decides both.Sep 21, 2026
Reply handling: classes, SLA, and what a human still ownsHow we triage 720 inbound messages, which reply classes survive counting, the clock that replaced approvals, and the decisions a person keeps for themselves.Sep 18, 2026
Sales Navigator, Apollo and job boards: which answers whatFour outbound data sources measured on 6,320 sourced people and 2,628 invitations, with the one question each of them answers well and the price of asking the wrong one.Sep 16, 2026
Outbound exclusion lists: what goes on them, and whyFour sources for a do-not-contact list, measured across 543 listed companies and 707 rejected candidates, plus the one empty field that quietly defeats the whole check.Sep 14, 2026
Founder-led outbound: what only your own profile can doNumbers from 2,338 cold invitations sent across founder, employee and rented LinkedIn profiles, showing which parts of founder-led outbound actually pay.Sep 11, 2026
Warming up a new LinkedIn sending profile, day by dayThe ramp we run on every new sending profile, with acceptance, refusal and pacing numbers measured across sixteen profiles between July and September 2026.Sep 9, 2026
Personalization that survives contact: one checkable factWe sourced 4,887 people in ten weeks and measured how old the personalized fact was when it landed. Median four days. Here is what ages well and what turns into an apology.Sep 7, 2026
Hiring signals: reading a company's own job boardWe scanned a few hundred company domains over six weeks. Board coverage ran from 54% to zero depending on the slice. What an open role is worth as a sales signal, and how to check it.Sep 4, 2026
How to read an outbound weekly report without fooling yourselfSix weeks of our own numbers. One week reads 8.7% or 1.8% depending on the denominator you pick. Seven checks to run before you trust a Friday outbound report.Sep 2, 2026
The five steps of outbound, and which one an agent closesOutbound splits into five steps. Across 3,706 sourced people and 1,651 invitations from our own profiles, a model closed one of them end to end. Here is the split, with numbers.August 2026
Aug 31, 2026
What to send after the first reply to a cold messageAcross 139 cold conversations in six weeks, 28% produced a second reply. Here is what a first reply contains, what we sent next, and the follow up that finally earned a number.Aug 28, 2026
Why LinkedIn limits your account, and the pacing that stops itPlatform refusals across 1,824 invitations from our own sending profiles in 33 days: why 14 of 18 profiles took zero, and the four pacing rules that decide which group a profile lands in.Aug 21, 2026
Buying signals that still work when everyone has the listTrigger events are sold as an edge. Here is what a hiring filter, a role change and a plain profile filter actually returned on live invitations, with every denominator written next to the rate.Aug 19, 2026
What an SDR actually costs in 2026, fully loadedA line by line cost model for one SDR seat in 2026: compensation, employer load, tools, ramp and management time, with funnel numbers from a live pipeline to turn it into cost per meeting.Aug 18, 2026
Cold LinkedIn benchmarks in 2026, and what actually moves themAccept rate, reply rate and meeting rate from a running outbound engine, with the sample size next to every number and the four things that changed ours.RUN IT ON YOUR PIPELINE
Re:Vault runs this for you
Operated LinkedIn outreach: we find the buyers, write in your voice, handle replies and book the meetings. $2,000 a month, month to month. Your own Claude can watch the whole thing from $199 a month.