Here's the trap. You buy a plan with a credit allowance, you use the product properly, and at some point in month three you cross a line nobody drew. What happens next isn't in your contract, isn't on the pricing page, and isn't the same at any two vendors.
One HubSpot admin described it plainly on Reddit: the credit system "just lets you blow past what you've paid for", and they put the damage at roughly $11,000 for the year. Nobody told them there was no ceiling, because nobody tells anybody.
A credit meter is a taxi with the fare display switched off. The meter's running, the rate is fine, and you'll find out where you got to when you arrive.
There are three things a tool can do when you hit zero, and which one you get is the single most consequential thing about a credit plan. It's also the thing you cannot find out by reading the tier comparison.
Behaviour one: it switches you off
Kixie charges a $50 reserve when you first submit a card, separate from any subscription. It's a prepaid balance, and usage draws it down.
If your card fails and the reserve reaches zero, the account is disabled. Not throttled. Not queued. Off, on a product whose whole job is answering and placing phone calls.
That's arguably the most honest of the three, because the failure is loud and immediate. It's still worth knowing before your team discovers it mid-morning on a Tuesday.
Kixie is also a good example of why the pricing page isn't the whole story. Its number rotation and reputation management are automated rather than your job, and that automation costs $50 a month per user plus $1 a month per number. That figure appears on exactly one page across all Kixie properties, and it's a help-centre article. The vendor's own note there says nationwide teams typically need around 350 numbers.
Behaviour two: it lets you keep going
This is the $11,000 one, and it's the default assumption almost nobody checks.
The balance goes negative. The product keeps working. Everything feels fine, because from the inside a working tool looks identical whether you're inside your allowance or four times past it. The signal arrives on an invoice, and by then the spend is historic.
The reason this pattern exists isn't malice. A tool that cuts you off mid-campaign generates furious support tickets, so letting usage run is the kinder default in almost every case except the one where somebody automates a workflow and forgets about it.
Behaviour three: nobody says
Lead411 publishes no overage rate anywhere. Not on the pricing page, not in the terms.
Its pricing page shows $49 and $150 a month and describes the allowances as 1,000 exports and "1,000+". What happens at 1,001 is undefined in public. It's also the vendor whose page never uses the word seat while every figure on it is per seat, so the pattern is consistent.
An unpublished overage rate isn't necessarily a high one. It does mean the number gets decided after you're already over, by someone who knows you're already over.
So how fast are you actually burning them?
Knowing the cutoff behaviour only helps if you know how fast you're approaching it, and the word "credit" is doing different work at every vendor.
| Vendor | What one credit buys | Read |
|---|---|---|
| Lusha | One email. A phone number costs five | 12 Aug |
| Amplemarket | 600 phone vs 13,500 email per user per year | 11 Aug |
| BookYourData | One contact, email and mobile together | 12 Aug |
| Regie.ai | 120,000/yr, stated as 1,000 accounts | 5 Aug |
| Warmly | 10,000 a month. What consumes one: unpublished | 12 Aug |
| Lead411 | 1,000 exports, then unstated | 6 Aug |
Lusha charges five credits for a phone number and one for an email. If you run a calling motion, your allowance is a fifth of what the number on the page implies. And on an annual plan, unused credits are wiped at the end of the cycle, so the discounted plan is the one that confiscates.
Amplemarket makes the same split explicit and the ratio is starker: 600 phone credits per user per year against 13,500 email. That's not a plan with a calling feature. That's an email plan with twenty-two phone numbers a month attached.
Warmly is the one to watch for a different reason. It grants 10,000 credits a month and publishes nothing about what consumes one. You cannot forecast against that at all, which puts it in the same bracket as the pricing pages whose numbers aren't offers: technically published, practically unusable.
Why does nobody publish the overage rate?
Because publishing it turns a flexible conversation into a fixed one, and because the number is genuinely awkward to state.
Credit costs are not linear for the vendor. A phone number sourced from a live carrier lookup costs them real money each time; an email already in the database costs almost nothing to serve twice. A single overage rate has to average across both, so it is either too high to advertise or too low to sustain.
That is a real constraint rather than an excuse. It also means the rate you eventually get quoted is negotiable in a way a list price isn't, and that the right time to negotiate it is at signature rather than at the overage.
Ask for a capped rate, or a hard stop, or both. A hard stop at your allowance with an explicit opt-in to continue is the single cleanest term available here, and vendors grant it more often than buyers ask, because it costs them nothing and removes their worst support conversation.
What to ask before you sign
Four questions, and they take one email. Ask them together, because the answers only mean something as a set.
- What happens at zero: cut off, keep going, or something else? This is the question, and it has three possible answers rather than a range. Get it in writing, because it is the one that produced the $11,000.
- What is the overage rate, in the same units as the plan? If there isn't a published one, that is your answer and you should price the risk accordingly.
- What consumes a credit, itemised? Not "enrichment". A phone number, an email, a verified email, a company record, an export. The ratios between them decide your real allowance.
- Do unused credits roll over or expire, and does that differ on annual? On at least one vendor here the annual plan is the one that wipes them, which inverts the usual assumption that longer commitment means better terms.
Then set a calendar reminder at 70% consumption. Not 90%, because by 90% the decision is already made for you, and every one of these meters accelerates when a campaign scales.
So what happens when your credits run out?
One of three things, and the vendor picked which before you signed.
You get switched off, which is loud and recoverable. You get carried, which is quiet and expensive. Or you get an unpublished rate applied after the fact, which is the one where you have least leverage, because the negotiation starts with you already owing.
None of that is on a pricing page, and none of it will come up on a demo unless you raise it. It costs one email to ask and, on the evidence of one Reddit thread, about $11,000 not to.