How agencies set per-client LinkedIn connection caps that do not starve email sequences
LinkedIn invitation limits sit on the person who owns the seat, not on your agency brand. Email send caps sit on the client domain. When ops treats both as one shared daily budget, the LinkedIn seat burns out by noon and the email queue for that client sits idle until someone notices.
That is the failure mode this post fixes. Cap LinkedIn connection requests per client workspace and per seat. Keep email sequences on their own throttle. Orchestrate touch order so a LinkedIn invite pause never freezes the email lane.
What LinkedIn actually constrains
LinkedIn does not publish a clean daily invite number for general outreach. Its help language says invitations are limited to protect member experience. Community-observed operating ranges that agencies use in 2026 cluster around:
- About 15 to 25 connection requests per day on a warmed, established seat
- Roughly 100 invites per rolling week for many free and Premium seats (Sales Navigator often reported higher, still not a license to spike)
- Warm-up closer to 5 to 10 per day for new or dormant seats
- Free members limited to a small monthly count of personalized invite notes (LinkedIn publishes five per month for basic accounts)
Treat those as ceilings you stay under, not targets. Acceptance rate and sudden volume jumps matter as much as the raw count. Pending invites that pile up also eat headroom.
None of that is an agency-wide pool. If Client A and Client B both use the same LinkedIn seat, they share one fragile budget. Prefer one seat identity per client workspace when the commercial sender is the client brand.
Separate the two throttles
Write two explicit configs on every client workspace:
- LinkedIn connection cap. daily and weekly max invites for the seat(s) attached to that workspace, plus a warm-up stage.
- Email send cap. daily max on the client sending domain(s), already covered in your send-window work.
Do not let a shared "outbound actions today" counter decrement both. When LinkedIn hits its seat cap, email should keep dequeuing inside the domain limit. When email is paused for deliverability, LinkedIn can still send a small, paced invite batch if the seat is healthy.
Put the rule in the queue worker, not in a Slack reminder. Soft budgets get ignored on busy Fridays.
Per-client allocation when seats are scarce
Sometimes one AE LinkedIn seat must support more than one client for a short stretch. Then allocate invites deliberately:
- Split the daily seat budget by client priority (for example 12 / 8 on a 20-invite ceiling), and hard-stop each slice.
- Prefer warm or higher-intent LinkedIn targets first so acceptance stays healthy.
- Never dump leftover Client A invites into Client B at 4pm just to "use the budget." That trains spammy patterns and mixes ICP noise into the seat history.
- Log which workspace consumed which invite so weekly scorecards stay honest.
If you cannot explain the invite split on a buyer call, you do not have a real allocation.
Keep email sequences moving when LinkedIn pauses
Sequence design should assume LinkedIn may pause mid-day:
- Email steps should not require a prior LinkedIn accept as a hard gate unless the playbook truly depends on it.
- If step 2 is "LinkedIn connect" and the seat is capped, skip or delay that step and continue the email branch on schedule.
- Stop the whole sequence on any reply (email or LinkedIn message), same as always.
- Deduplicate so the same prospect is not connection-requested from two seats or emailed twice in one day across channels.
Orchestration without double-touch fatigue still applies. Caps are the volume layer. Touch order is the experience layer.
Ops board that catches starvation early
Run a daily board per client:
- LinkedIn invites sent / daily cap / weekly rolling total
- Pending invites and acceptance rate (pause if acceptance collapses)
- Email sends vs domain cap
- Sequences blocked waiting on LinkedIn (this number should stay near zero)
- Any seat restriction or challenge events
If "sequences blocked waiting on LinkedIn" is not zero, your caps are coupled wrong.
Practical defaults to start
For a warmed client seat supporting one ICP:
- Daily LinkedIn connects: start at 15, raise only if acceptance stays solid for two weeks
- Weekly LinkedIn connects: stay near 80 to 100 unless the seat's history clearly supports more
- Email: keep the existing per-domain daily cap and send windows
- Warm-up seats: 5 to 10 connects/day until the account looks normal in real use
Adjust per client. A founder-led seat with a strong network can behave differently than a brand-new client persona seat. Do not copy one client's numbers onto another workspace.
Why this belongs in the workspace, not in a spreadsheet
Agency outbound fails when LinkedIn and email share an informal budget. Client-isolated workspaces make the fix obvious: each client owns its LinkedIn identity budget and its email domain budget. GTM's motion is find → qualify → write → send/reply/book with both channels in that isolated workspace. Caps are how you keep the send/reply stage alive when one channel hits a platform ceiling.
Ship the two throttles. Keep email moving when LinkedIn is done for the day. Measure blocked sequences, not just invites sent.
If you want client-isolated outbound with email and LinkedIn under one ops model, join the GTM waitlist.
Sources: Understory LinkedIn outreach agency playbook, GTM-API LinkedIn connection limits 2026, Salesforge LinkedIn connection request limit, Expandi LinkedIn connections limit, GTM product knowledge.