Separation, not volume, is the hard part
An in-house team runs one message into one market. An agency juggles ten positionings, ten tones of voice and ten sets of personas, and has to stop them bleeding into each other. Most LinkedIn tools were designed for the first case and then adapted, badly, to the second.
Three costs show up that nobody warns you about when you sign your first retainer.
Account separation is a hard constraint, not a preference
Each client needs their own sending identity. Not a shared inbox, not a rotating account, but their own. Per-seat pricing makes that expensive quickly, which is why agency plans exist. HeyReach bundles 25 senders at $999 a month; Skylead gives 50 seats at the same price. Compare on cost per client account rather than on headline price.
Sales Navigator is where it gets genuinely messy. Whether a seat can be shared across accounts is one of the most searched and worst-answered questions in this market, and the answer has real billing consequences. Get it wrong and you either overpay or put client accounts at risk.
Setup time is your actual margin
Your unit economics live in the hours between signing a client and sending the first message, not in the tool subscription: understanding their offer, defining personas, writing sequences, building lists, then maintaining all of it as campaigns decay.
Do the arithmetic on your own numbers. Take the hours it takes to onboard one client, multiply by clients per month, add the weekly maintenance across every live account, and set the total against your retainer. For most agencies the tool licence is a rounding error and the setup time is the whole business.
This is why the campaign model hurts agencies more than anyone else. Every new client multiplies it.
Reporting is what gets you renewed
Clients do not renew because your acceptance rate is good. They renew because they can see what they paid for. Yet almost no LinkedIn tool ships client-ready reporting, so you export CSVs and rebuild a deck every month.
Whatever you use, decide early what the monthly report should look like and make the tooling serve it, rather than reporting whatever the tool happens to export.
White label: when it matters, and when it does not
Presenting the tool under your own brand matters if you sell an outcome and want the method to stay yours. It matters much less if your clients know they are buying execution on a named platform, which is increasingly common.
Ask a vendor what exactly gets rebranded, rather than whether white label exists at all: the client-facing dashboard, the reports, the notification emails, the login domain. The answers differ enormously, and "white label" on a pricing page tells you nothing.
Spruce offers white label with volume-based per-seat pricing, so you can calculate your margin per client and how many accounts one person can realistically handle. See how it works.
What to test before you commit
Run one real client through a trial rather than a demo dataset. Watch how long the setup actually takes, whether the tool keeps client contexts genuinely separate, and what the monthly report looks like with no manual work behind it. Those answers predict your margin better than any feature grid.