Add a client to an agency that runs acquisition for other businesses and the revenue side gains one line. The cost side gains about eight. A CRM seat so somebody can actually work the account. A lead database seat, or a draw against a credit block you committed to in January for volume you were guessing at. A separate sending workspace, because mailing for a new client from another client's domain is not a thing you want to explain later. An answering service if the phone matters, billed by the minute. A scheduler seat. A builder subscription for the campaign page. An automation plan underneath it charging per task. Your retainer is priced against what that one client gets out of the arrangement. Everything beneath it is priced against a number the client never sees, which is how many other clients you have.
Some vendors will let you park client accounts under an agency umbrella. You still administer them one login at a time, and you still pay per attachment. The umbrella is a billing convenience, not a change in how you are charged.
Work out what one client costs you before anybody touches a keyboard
Mission Control is $149 a month. Three AI workers, unlimited apps, 500 credits. Spread across ten clients, the acquisition stack costs $14.90 per client per month. Across four, $37.25. Those are whole numbers you can put in front of a partner without a footnote about what happens at renewal. If you are two or three clients in rather than ten, Frontier at $79 gives you two workers and 250 credits, with meetings booked straight onto a calendar.
Every meter in the standard stack runs on something you cannot control
HubSpot charges per seat, and the tier jumps tend to land exactly when you need one more capability. Apollo, Lusha and Clay meter credits and seats, with the good rate usually attached to an annual commitment. Calendly is per seat. Zapier bills the plumbing by the task, so the better the plumbing works, the more you pay for it. Answering services such as Ruby and Smith.ai charge by the minute or the call, which puts your cost of goods on the same curve as your client's success. A busy month for them is an expensive month for you. You either eat it or you send an invoice with a line on it nobody discussed, and the second option costs more than the money.
Three workers means three live outbound operations, and you choose which
A worker finds buyers, writes to them, reads what comes back and sorts it. Three of them is a real ceiling and I am not going to pretend otherwise. It is a ceiling on the thing agencies already ration, though, which is attention. Agencies rarely come apart because somebody lacked a fourth login. They come apart because six accounts were nominally live and two of them had not been touched in a month.
The inbound line is what clients notice first. Their number answered around the clock, recordings and transcripts on every call, transfer to a person mid-call, after-hours handling that does not just take a name. You stop having the conversation where a lead went cold because nobody picked up on a Saturday, and if that conversation happens anyway, you have the recording.
Unlimited apps moves a cost line onto the invoice you send
Every engagement wants an asset. A landing page for the campaign. A booking flow. A quote form that writes into the pipeline rather than into somebody's inbox. Some small internal thing the client's office manager opens forty times a day. Lovable, Base44 and Bolt meter generation and price per project, so an agency's library of client assets becomes a recurring liability that grows every quarter, and the twelfth asset costs about what the first one did.
Unlimited apps with custom domains flips that. The asset lives on the client's domain, it takes Stripe payments if it needs to, and the marginal cost of the next one is nothing. Price it into the retainer at whatever number you think it is worth.
Size the market during the discovery call and pay nothing for it
Searching, counting a market and previewing 25 people by name are free. That is a sales tool before it is a data tool. Rather than promising a list in the proposal and finding out later, you count it while the prospect is still on the call. If their real universe turns out to be a few hundred companies rather than the tens of thousands everyone assumed, you both learn it before signing, and the proposal you write is a different document.
Then the arithmetic on the data itself. Adding a person costs one credit. A verified phone and email cost five, and they only bill when the data is actually found, so you are not paying for misses. Mission Control's 500 monthly credits go a long way if you are mostly adding people and a much shorter way if you enrich everyone. An agency running heavy volume for a single client will hit that. Budget credits per account the way you already budget hours.
Forty handoffs instead of four
One client on five tools gives you four seams, and you paper over them yourself between other jobs. Ten clients on five tools gives you forty, spread across a team, each with its own way of failing on a Friday evening. A reply lands in the sending tool and the CRM never hears about it. The answering service takes a call from someone who is three emails into a sequence and treats them as a stranger. The scheduler books over a slot. Nobody catches any of it until the client asks a question you cannot answer.
What an agency actually sells is judgment about who to talk to and what to say. Time spent moving records between systems is not that, and it grows with every client you sign.
Bounce suppression, sending-domain checks, reply classification, pipeline stages, custom fields, tasks, quotes, CSV import and export, seats for your own people. In one place, the eleventh client is an afternoon of setup instead of another copy of a fragile assembly nobody documented.
What is left when a client leaves
Reselling access to somebody else's software has an expiry date on it. Sooner or later the client finds the vendor's pricing page and does the subtraction, and that meeting goes badly no matter how good the work was. Selling the operation holds up better, because the operation is a list you built, messages that get replies, a phone that gets answered, meetings sitting on a calendar and a pipeline you can export and hand over the day they go. Tag contacts by client with custom fields. Set pipeline stages that match how that particular business actually closes, not how the last one did.
All of it starts the day the contract is signed. Published price, a card, live in minutes, cancel from the dashboard. No procurement call and no annual commitment made before you know whether the account renews in month four.