“Buy call center leads” is two purchases wearing one search term. One of them is a lead: a person who raised a hand, sold to you per lead with the intent already priced in. The other is a record: a household chosen on a signal, sold per record, with the raising of the hand left to your setters. Vendors in both businesses call the product leads, quote it per unit, and offer a sample. The purchases have almost nothing else in common.
This guide is the buying sequence for either one, from the decision about which you are actually buying through the terms to get in writing and what a test batch has to prove. It does not re-derive the exclusive-versus-shared arithmetic or the post-delivery audit — both have their own pages, linked where they belong.
Disclosure: Scout Data is our product and we sell records, not leads — the second column below. The advice on the first column is what we would give a floor buying from someone else.
Two products behind one search
| Per lead | Per record | |
|---|---|---|
| What arrives | A named person who responded to something, recently or not | A household chosen on a signal, with a phone for the owner |
| Who did the work | The vendor generated the response; your floor closes it | Your floor generates the response; the vendor found the house |
| Price shape | Dollars per lead, tiered by freshness and exclusivity | Cents per record, tiered by volume and selects |
| What goes wrong | Resold, recycled, or not what the form said | Wrong-party phones, stale ownership, renters |
| Fits a floor that | Has closers and no setters, or needs volume this week | Has setter hours to fill and a pitch about the house |
The decision between the columns is about labour. A lead carries the setter’s work inside its price; a record does not. If the floor has idle setter hours, records fill them at a fraction of the unit price. If it has no setters, the lead price is buying a function you do not have. How to price a bought lead works backwards from your close rate to the most either one is worth.
Exclusive or shared
On the per-lead side this is the term that sets the price, and the argument about whether the shared discount is big enough has an answer: the discount has to be at least as deep as the drop in your own close rate on shared leads. The formula, a worked example and the cases where shared genuinely wins are in exclusive vs shared leads. Read it before the quote, because the vendor will not run the arithmetic for you. On the per-record side the equivalent question is whether the same rows are sold to your competitors in the same weeks — a compiled file usually is; a file built to your spec usually is not — and the honest way to ask it is “who else gets these rows, and when.”
Five terms to get in writing
1. Replacement
What counts as a bad unit — disconnected, wrong party, duplicate, already on your suppression file, a lead who never filled the form — and whether each is credited, replaced with a fresh unit, or simply yours. Then the window: bad units found in week one are arguable in a way that bad units found in month three are not, so the window has to be long enough to dial the file properly. A replacement policy that exists but is not written is a policy that exists until the first claim.
2. Refresh
How old the underlying data is on the day it ships, and how often it is re-verified. For a lead: when the form was filled and whether the vendor re-confirmed anything since. For a record: when ownership was last checked and when the phone was last verified live. “Verified” is a word with no date in it; ask for the date.
3. The resale window
For a lead: how many buyers, over what period, and whether the same lead is sold again as aged once the fresh window closes. For a record: whether the same rows are pulled for other buyers in the same geography in the same month. Get a number and a period, not an adjective.
4. Suppression before delivery
That the vendor will take your suppression file — opt-outs, complaints, customers, your last three orders — and apply it before the file is built, so you are not paying for rows you would discard and not re-dialing a household that asked you to stop. A vendor who applies it after invoicing is charging you for the dedupe.
5. Format
The columns, exactly; one number per row or a ranked set; line type on every number; the registry scrub date on the file; owner-occupied and the pitch-setting fields your script depends on. The full row is in homeowner phone lists for call centers, and what makes any file load cleanly is in dialer-ready lead lists.
What the test batch has to prove
Every vendor offers a sample, and most samples prove nothing, because they are dialed once by the best rep and compared to nothing. A test batch that means something is run against the list you dial today: same reps, same hours, same script, three passes over one to two weeks, with wrong-party logged as its own outcome. It has to prove exactly two things — a connect rate and a wrong-number rate, next to the incumbent’s — and it is not allowed to prove deals, because a thousand rows is enough to measure whether the phones ring the right people and far too small to measure sales. The method in full is in free lead samples: how to test a vendor.
Lines in a quote that mean trouble
- “Up to five numbers per record.” Padding. Five numbers with no ranking is the vendor declining to decide which one is the owner’s, and charging you for the indecision.
- “Verified” with no date. Verified when, against what. A phone verified in March is a phone in September.
- “TCPA compliant.” No purchased list is compliant on its own; a campaign is. What the phrase can and cannot promise is in TCPA-compliant homeowner leads.
- A match rate in the headline. Match rate counts rows with any number on them. It is the number that flatters address-matched files most.
- A monthly minimum before the test. The order of operations is sample, then commitment. A vendor who needs the commitment first is telling you what the sample would have shown.
- Per-lead pricing with no cap on buyers. A shared lead with an unwritten buyer count is a lead shared with everyone who paid.
The sequence
- Decide lead or record from your setter hours, not from the vendor list. Idle setters buy records; no setters buy leads.
- Write the spec — columns, signal, geography, volume — before the first call, so every quote is for the same thing.
- Get the five terms in writing on every shortlisted vendor. Drop the ones that will not put a number on replacement or resale.
- Run the test batch against the incumbent list, and decide on connect and wrong-number rates alone.
- Audit the first paid delivery on a sample you drew — match-back, duplicates, disconnects, wrong-party — using how to audit a lead vendor, and claim the credits the contract names.
- Stamp every row with its source at import and keep a ledger per source, so the vendor that wins the test can be checked every month it stays; how to measure lead source ROI is that ledger.
If the cheapest tier of the per-lead column is on the shortlist — the aged batch — its own economics are in aged solar leads: the discount is real, and so is the cadence it needs to pay off.
Frequently asked questions
What is the difference between buying leads and buying data for a call center?
A lead is a person who did something — filled a form, answered an ad — and is sold per lead with the intent priced in. A record is a household chosen on a property or demographic signal, sold per record, with the work of turning it into a conversation left to your floor. Leads cost more per unit and less in labour; records cost less per unit and need setters with hours to spend. Which one to buy is a question about your floor, not about the vendor.
What should a call center get in writing before buying leads?
Five terms: what counts as a bad record and whether it is credited or replaced; how old the underlying data is and how often it is re-verified; how many buyers see the same lead and for how long; whether the vendor accepts your suppression file before delivery; and the exact columns, line-type flag and scrub date the file will carry. None of those is on a rate card, and all of them decide what the rate actually buys.
What should a test batch of call center leads prove?
Two rates against the list you dial today: connect rate and wrong-number rate, measured on the same reps in the same hours over three passes. Not deals — a test batch is large enough to measure whether the phones ring the right people and far too small to measure sales, and a vendor who asks to be judged on appointments from a thousand rows is asking to be judged on noise.
Are exclusive call center leads worth the price?
Sometimes. The shared discount has to be at least as deep as the drop in your close rate on shared leads, and that is a number only your own tagged dispositions can supply. A floor that answers inside a minute can make shared leads pay; a floor where the closer is also the answerer usually cannot. The formula and a worked example are in our exclusive-versus-shared guide.