The honest unit of cost in cold outreach is not a monthly subscription. It is one contact: what you paid to find them, what you paid to reach them, and what it took out of your week when they wrote back. Priced that way, the number on the pricing page turns out to be the small end of the bill, and founders who abandon outbound in month three usually abandon it over the parts nobody itemized.

So let's do the arithmetic properly.

Data is sold by the row, and the dead rows cost the same as the good ones

Every contact database in this category prices in credits. Apollo, Lusha, Clay: you buy a bucket, then spend from it to reveal an email or a direct dial. Clay will push a single row through provider after provider until something comes back, which means one contact can consume four or five credits before you have written a word to them. That is how the meter is built, and you are the one feeding it.

The part that gets skipped in the sales conversation is that you pay for the misses too. A credit spent on an address that stopped working in March is gone. You find out three days later when the bounce arrives, and the bounce takes a piece of your sending reputation with it, which is a cost you feel a month later and cannot put a figure on at the time. Take your own bounce rate from the last list you mailed. Whatever it is, that is the multiplier sitting on top of your per-contact data cost, because you bought the entire list in order to reach the working part of it.

You do not buy sending once, you keep it running

Send cold email at volume from your company domain and you will eventually damage the domain your customers use to reach you. The standard fix is secondary domains. Each one needs mailboxes, DNS records set correctly, and a warmup period measured in weeks before it carries real volume. Each mailbox carries a per-seat fee from whoever hosts your mail. Then a sending platform on top, priced by how many mailboxes you connect and how much leaves them, so the bill climbs at exactly the point the campaign starts working.

For a modest setup that lands in the low hundreds a month before a single stranger has heard from you. Still not the expensive part.

The line item nobody writes down is their own time

Take a month at five hundred contacts. Small by any agency's standard, plenty for a founder fitting it between customer calls. Put your own numbers in if mine are wrong, because the shape of it does not change.

  • Building and filtering the list, then filtering it again when the titles turn out wrong: 3 hours
  • Cleaning the enrichment output, deduping against people already in the CRM, deleting the rows that came back as garbage: 2 hours
  • Writing the sequence, then the variants, then rewriting the opening line because it reads like a template: 4 hours
  • Domain and mailbox upkeep, warmup babysitting, working out why one inbox went quiet: 2 hours
  • Reply handling, which is the whole ballgame: see below

Assume five in a hundred write back, because you have to assume something. Twenty five replies. Twelve minutes apiece is an optimistic average once you count reading it, working out who the person is, checking whether a colleague already spoke to them, writing something that does not sound canned, and going back on Thursday when they go quiet. Five hours.

Sixteen hours. Two working days out of twenty, taken from the person who decides what the company builds, and delivered in ninety-second interruptions spread across three weeks, which is the worst shape time can arrive in. Divide those hours across five hundred contacts at whatever you think an hour of your attention is worth, then set the result next to the credits and the mailbox fees. It is not close.

Almost none of that time was spent thinking about the customer

Look at where the sixteen hours go. Export, import, dedupe, open a tab, copy, paste, check, remember. The list tool has never heard of your pipeline. The sending tool does not know your phone was answered on Tuesday by somebody from the same company. The calendar has no idea a deal exists. Every boundary between two products is a place where the work stops and waits for a human, and in a young company there is one human.

Five subscriptions do not integrate themselves. The founder does that job, on top of the actual job, for free.

Outbound rarely gets dropped because the messages failed. It gets dropped because nobody can keep being the connective tissue between five products and still run a business.

What that same month costs on Rocketship

Launch is $24.99 a month. It finds the buyers, writes to them, reads what comes back and answers it, and your number is answered 24 hours a day by an AI receptionist that transfers to a human mid-call, records and transcribes, and handles after hours. Searching the data, counting how big the segment actually is, and previewing 25 people by name are free, so you can size a market before you spend anything at all. Adding a person to your list is one credit. A verified phone and email is five, charged only when they are found. Launch includes 100 credits a month. Frontier is $79 with 250 credits, a second AI worker, and a scheduler that books meetings onto your calendar. Mission Control is $149 with 500 credits and three workers.

Replies get classified. Bounces get suppressed. Your sending domain gets checked before you send from it. All of it lands on one contact record inside one pipeline, and that is precisely where the sixteen hours were disappearing.

A hundred credits is not unlimited outbound, and the limit points the right way

Be clear about what the entry tier buys. One credit to add someone, five for a verified phone and email, so a fully enriched contact runs six credits and 100 credits is around sixteen of them a month. Frontier and Mission Control buy more. Search and market counts stay free at every tier, so nothing stops you mapping ten thousand companies and then spending credits on the forty that matter.

I would rather have that constraint than not have it. When contacts feel free and your own hours are the scarce input, the rational move is volume, and volume is what burns the domain and kills the program by spring. When each contact costs something you can see, and the replies are handled without you sitting there, writing to forty people properly beats writing to five hundred badly. That is the version of outbound a company with one founder and no sales hire is still running in month six.