Pricing pages in this category are unusually hard to read, and not always by accident. Tools bundle different parts of the pipeline, charge on different units, and quote a headline number that assumes you are already paying somebody else for the rest.
So this article does two things. It explains the four charging models and what each one is really optimising for, and it lists the costs that turn up later.
It deliberately does not quote prices. Published numbers in this category change often enough that any figure written today is misleading within months, and a stale price presented as current is worse than no price. What follows is the structure, which changes far more slowly.
The four pricing models
| Model | You pay for | Rewards | Punishes |
|---|---|---|---|
| Per seat | Named users | Small teams, high value per deal | Adding occasional users |
| Per contact / credit | Records touched or enriched | Precise, low-volume targeting | Experimenting with lists |
| Per mailbox / volume | Sending capacity | Predictable high volume | Bursty campaigns |
| Flat platform fee | Access, within limits | Heavy, consistent use | Light or seasonal use |
Per seat
The familiar SaaS model. You pay for each person with a login.
It suits teams where a defined group runs outbound and each person's activity is roughly similar. It becomes awkward the moment outbound is something several people touch occasionally, because you either buy seats that sit idle or share a login, and sharing a login breaks the attribution that makes reporting worth having.
Watch for what a seat includes. A seat that comes with one connected mailbox and a cap on sends is a very different unit from a seat with unlimited mailboxes.
Per contact or per credit
You pay for each record you enrich, verify, or contact. Credits are the same idea with an abstraction layer, and the abstraction is usually where the confusion enters: one credit may not be one contact, and different actions may cost different numbers of credits.
This model is genuinely good for low-volume, high-value outbound, where you contact few people and each is worth a lot. It has one structural problem: it charges you for learning. Building a target list is iterative. Your first definition of your ideal customer is wrong, you find that out by contacting some of them, and you refine. Under per-contact pricing, every iteration has a bill attached, which quietly discourages the exact experimentation that fixes the list, which is the thing most likely to be wrong.
If you choose this model, ask specifically what happens to credits for records that fail verification. Paying full price to discover an address is dead is a meaningfully worse deal than it first appears.
Per mailbox or sending volume
You pay for capacity: how many mailboxes are connected, or how many messages go out.
This aligns well with how outbound actually consumes resources, and it makes scaling costs predictable. It suits teams sending consistently. It suits campaign-shaped work less well, because you pay for capacity in the quiet months too, unless the plan allows genuine downgrades.
Flat platform fee
One number for access, usually with limits generous enough that most users never meet them.
Easiest to budget, easiest to compare, and the best value for anyone using the tool heavily. The thing to check is what sits outside the flat fee, because something usually does, and it is normally the data.
The bill nobody quotes
A working outbound stack needs all of the following. Some tools include some of them. No quote includes all of them by default, and this is the single biggest reason two apparently similar prices are not comparable.
1. Sending domains. Serious cold outreach does not go out on your primary domain. It goes out on a separate domain, redirected to your real site, so that a campaign that damages reputation damages something recoverable. That is a registration cost per domain per year, small but real, and it is a decision you can only make cheaply at the start. Why a separate sending domain matters.
2. Mailboxes. Volume has to be spread across several mailboxes to look like several people rather than one machine. Each mailbox is usually a paid seat with whatever provider hosts it, and this is frequently the largest recurring cost in the whole stack. It is also the one most often left off comparisons, because it is paid to a different company.
3. Email verification. Either bundled or bought per lookup. Not optional: bounces are charged to your sender reputation, and the cost of skipping verification arrives later disguised as a deliverability problem. The bounce rate guide covers the mechanics.
4. Data or enrichment. Where the leads come from. Some platforms include discovery; many start at "upload a CSV", which means the hardest stage of the pipeline is a separate subscription.
5. Warm-up. Mailboxes need a history of normal conversation before they carry campaigns. Sometimes included, sometimes a separate product.
6. AI usage. Some platforms include model costs, some ask you to bring your own API key. Bring-your-own is usually cheaper and always less predictable.
Write these six down as line items before comparing anything. The exercise takes ten minutes and regularly reorders the shortlist, because the cheapest platform fee is often attached to the longest list of things you now have to buy separately.
The only number that matters
Cost per send is a vanity metric. It always improves when you send more, and sending more is the easiest thing in the world to do badly.
The number that decides anything is fully loaded cost per positive reply:
(platform fee + data + verification + mailboxes + domains) ÷ positive replies
Positive replies, not replies. "Not interested" is a reply and costs you the same to produce.
You cannot know this figure before you start, which is inconvenient and unavoidable. You can know it after one properly run sequence to a properly built list, which is why the correct first budget is the smallest one that buys exactly that. A single real sequence produces a number specific to your market, your offer and your list. Every comparison table on the internet, including any we might write, is a worse input than that number.
Where the money actually goes wrong
Three patterns account for most of the wasted spend we see described.
Paying per contact for a list you have not validated. The cost of a bad list under per-contact pricing is paid in full, immediately, for the privilege of finding out it was bad. Validate the definition on a small batch first.
Buying sending capacity before fixing deliverability. More capacity into an unauthenticated domain produces more mail in more spam folders. The fix costs nothing and comes first. SPF, DKIM and DMARC in twenty minutes.
Generating meetings the business cannot take. Outbound software produces conversations. If nobody has time to have them, the constraint was never lead generation and the spend was aimed at the wrong problem.
A note on our own pricing
Leads Ranger is in an open pilot and every feature is currently free, which makes us a poor source of neutral pricing advice and an easy one to try. Discovery, enrichment, verification, warm-up, mailbox rotation and the AI agents are one product rather than six line items, which is the design decision behind it: the six-subscription version of this stack is where most of the confusion in this article comes from.
Run one sequence, work out your own cost per positive reply, and use that to judge everything else, including us.
Where to go next
- How does an AI SDR work? for what you are actually paying for, stage by stage.
- Best AI SDR software for evaluating capability rather than price.
- The benefits and challenges of an AI SDR for whether it will work for your team at all.