A new sending profile carries no history the platform trusts, belongs to a real person whose network you can damage in a week, and keeps whatever reputation it earns in its first fortnight for months afterwards. That fortnight can be written down in advance as a schedule, day by day, and this is ours.
We run outbound as an operator. Between 27 July and 11 September 2026 we opened a ramp on sixteen LinkedIn profiles: our own, rented ones, and personal profiles belonging to people at the companies we work for. This article is the schedule we actually run, the numbers it produced, and the two places where the numbers say something different from what the schedule promises.
Day zero happens before the first invitation
The first thing a stranger sees is the headline on the profile, and they see it before they decide whether to accept. So the ramp starts with a gate: until the headline describes the angle the campaign is about, the queue stays shut and the daily ceiling stays on its lowest rung.
We learned the cost of skipping this on a rented profile whose headline announced a selling role in plain words. Twenty eight matured invitations went out from it and zero came back accepted, against 13.1% across every other profile in the same period, writing to the same kind of people. At a 13% base rate, zero out of 28 lands somewhere near a 2% chance. The headline was doing the work.
Two rules came out of that. A sending profile is dressed as an operator inside its own company or as a promise of usefulness, and the headline names what the person does for others. And the gate stays live after the first send: in September one of our rented profiles moved from one campaign to another while its headline still described the old one, and a single invitation went out under it before the check caught up. The gate now pauses invitations every morning the headline and the campaign disagree, and the pause expires by itself on the first morning they agree again.
Across our sixteen profiles the median gap between connecting a profile and its first invitation was two days. One ran to nineteen, because the person who owned it wanted to approve the recipient list first, which is their call and worth the wait.
Measure what the profile arrives with
A rented profile arrives with somebody else's history, and the platform counts that history as the profile's own. The measurable part is the pile of connection requests already sent and still unanswered.
We pulled that number for seven profiles and put it next to their acceptance rate. Five profiles carrying between 4 and 80 pending requests accepted at 14% to 25%. One carrying 889 sat at 9.3%. One carrying 546 sat at zero. On that last profile, 512 of the 546 had been sent before we ever touched it, 457 of them with notes attached, by a previous renter who had been mailing hard for two months.
A second rented profile shows the other half of the same story. Our own sending from it totalled 35 invitations, while the profile carried 845 withdrawn requests from its earlier life. Two weeks in, the platform began refusing new invitations on it with a temporary limit, and the refusals kept coming for a fortnight after that.
Withdrawing the old requests carries its own price, because the platform blocks a repeat invitation to that person for about three weeks afterwards while the reputation stays where it was. So the number gets measured before the rental agreement, in one call to the sent invitations list, and a profile carrying more than 200 pending requests goes back.
The ladder, and how long it really takes
Every profile gets a written ladder: a list of daily invitation ceilings, one rung per working day, applied by the system each morning. The shapes vary with the profile. A brand new personal profile starts at one invitation on day one and adds one per working day to a ceiling of ten. Personal profiles belonging to client staff, which have real networks and years of history, start at two and reach ten in eight working days. A rented profile that has been sending ten a day right up to handover keeps that ceiling, because it already holds the sending history a ladder exists to build, and resetting it to one would cost the campaign its first week.
Day one opens on an event, the first batch entering the queue, rather than on a calendar date. A ladder tied to the calendar ticks away while the profile sits waiting for a headline or an approval, and the profile then arrives at day eight holding the ceiling of day eight and the sending history of day one.
Here is what the ladder produced in practice, counted as invitations per active profile per day of its own life:
- day one: 2.5
- day six: 5.2
- day eight: 6.3
- day twelve: 8.9
- day fourteen: 9.2
The other number worth knowing is how long a ladder written in working days takes in calendar days. Across the sixteen profiles, the climb from first rung to last took a median of 11.5 calendar days, with a spread from 8 to 25. Weekends, platform pauses and days with an empty queue all stretch it. So a plan that says "we reach full volume in ten days" is honestly stated as two to five calendar weeks, and it is worth saying that out loud to whoever is paying, on the day you start.
What the ramp buys
Now the uncomfortable part, and the reason this article has numbers in it.
Sorted by the age of the sending profile, acceptance climbs. On invitations at least a week old, sent from profiles other than our one long warmed account: days one to five accepted at 10.0% on 150 invitations, days six to ten at 18.9% on 259, days eleven to twenty at 17.4% on 476, and day twenty one onward at 22.4% on 664.
That looks like a clean case for ramping. It is also exactly the shape you would get from survivorship, because the profiles that reach day 21 are the ones that were working well enough to keep. So we ran it again inside each profile, comparing its own first ten days against everything after. Fourteen profiles have both windows: early acceptance across them is 16.1% on 496 invitations, later acceptance is 20.7% on 1,285. Narrow it to the five profiles that cleared thirty invitations in both windows and three rose while two fell, one of them from 21.6% to 10.4%.
Read together, those three cuts say something specific about what a ladder is for. It buys a profile the right to still be sending in month two. The acceptance rate it gets there with is set by the audience, the headline and the message, and those are the three things to work on when the number needs to move.
The thing a ramp definitively buys is quiet. Our platform refusals fell from 6.1% of attempts to 0.4% after we spread volume evenly across the sending window, and the profiles that ran a ladder from day one have logged their refusals in short clusters rather than in the account restrictions that end a campaign.
Four ceilings to set before day one
Take these from the platform's own documentation and the published guidance of the tools in this market, then measure your own.
Count your own rolling week. Around 100 invitations per rolling seven days is the ceiling to track in your own numbers, so you meet it as a decision instead of as an error message. The highest weekly total we have recorded on a single profile is 101, and it was deliberate.
Price a refusal at one attempt. A refused invitation pauses that profile for half an hour, and it then goes to a different person. The day closes after three refusals in a row with nothing sent.
Spread the allowance across the day. The daily quota is divided over the whole working window, recalculated every cycle, with at most one cold touch per minute. A cluster in the first hour reads as machinery to the platform and as spam to the recipient.
Set working hours by the recipient's clock. Our strongest profile by acceptance, at 25.0%, was the one sending between 9 and 11am in the recipient's local time. Several others were firing at 3am local, because their working hours had been set in the sender's timezone while their audience lived five to nine hours away. That one setting was worth more than any rung on any ladder.
The plan you can run tomorrow
Write the headline first and read it as a stranger would. Measure the pending invitation debt on any profile you rent or inherit, and walk away above 200. Set working hours by your audience's timezone. Start at one or two invitations a day, add one per working day, and stop at ten until you have a month of clean acceptance. Spread the day's allowance across the day. Count your own rolling week. Then tell whoever is waiting for pipeline that full volume arrives in two to five calendar weeks, because that is what the calendar has actually done sixteen times in a row.
If you want the same ramp running under someone else's supervision, the service levels page shows where that sits, and the playbooks page has the rest of what we run.