Methodology

How to Evaluate an Enterprise Lead Partner

By LeadSquad8 min read
A dark grid of glass panels receding into shadow, with a single column lit in red — evoking the selection of one vendor from many.

Most lead vendor evaluations are run as a procurement exercise. Three vendors, a spreadsheet, a column for cost per lead, and a decision that lands on whoever quoted lowest with a tolerable reference call.

That process works for commodities. Demand is not a commodity, and the spreadsheet is measuring the one variable that predicts the least.

Here is the problem with a CPL-anchored evaluation. Cost per lead is an input price. What you are actually buying is a rate of revenue production per seat on your sales floor, and the distance between those two numbers is enormous. A $45 lead that your reps contact 22% of the time and close 6% of the time is far more expensive than a $180 lead they contact 71% of the time and close 19% of the time — but only the second number shows up in the CFO's model six months later, after the vendor decision has already been made.

The evaluation below is built to surface that distance before you sign. It has six dimensions. None of them is price.

1. Source ownership and traceability

The single most predictive question you can ask a demand partner: where does this come from, and can you show me?

You are trying to establish whether the partner generates demand or resells it. Both are legitimate businesses. They carry completely different risk profiles, and only one of them can improve.

Ask for:

  • The percentage of delivered volume from owned-and-operated properties versus partner or affiliate traffic
  • Source attribution at the record level — not "paid search," but the specific property, campaign, and ZIP
  • Whether you can see, in your own reporting, which source produced any given delivered lead
  • What happens to attribution when they use a partner source: is the sub-source disclosed, or does it collapse into "network"

What a strong answer looks like: they can name their properties, they show you source data down to the individual record, and they disclose affiliate traffic as affiliate traffic.

What a weak answer looks like: "proprietary methodology," "our network of publishers," or source data that arrives as a single undifferentiated channel label.

The reason this matters more than it sounds: when quality degrades — and it will, in some market, in some month — a partner with record-level source visibility can isolate and cut the bad source in a week. A reseller can only apologize and issue credits. You are not evaluating today's quality. You are evaluating how fast the floor can be found when quality moves.

2. What is actually verified, and by whom

"Verified" is the most abused word in this category. It can mean a database append, a form validation, a machine-scored intent model, or a live human conversation. Those are not comparable, and vendors will let you assume the strongest interpretation.

Get specific. For every delivered record, ask what was confirmed and how:

Attribute Confirmed how?
Property ownership Data append, or verified against a record of ownership?
Occupancy Is the contact the owner, or a resident?
Decision-maker present Confirmed live, or inferred?
Project scope Self-reported on a form, or confirmed in conversation?
Timeline Stated range, or confirmed against a specific window?
Budget authority Asked, or assumed?

Then ask the question underneath all of them: who performs the verification, and where do they sit? A live qualification desk and an automated scoring model produce different set rates and very different sit rates. Both may be described as "verification." Only one of them will hold up when your reps start reporting that the homeowner doesn't remember filling anything out.

Ask to listen to recorded qualification calls. A partner running a genuine desk will play you five. A partner running a model will explain why that isn't possible.

3. Exclusivity — read the contract, not the deck

Every vendor says exclusive. The word is doing different work in different contracts.

The three definitions you will encounter:

Genuinely exclusive. The record is sold to you and to no one else, ever, in any form.

Category-exclusive. Sold to one roofing company — and also to a windows company, a solar company, and an HVAC company. The homeowner receives four calls in an hour. Your rep is the third.

Exclusive-for-a-period. Yours for 30 or 90 days, then resold into a lower tier or an aged-lead market. You do not find out about this from the contract; you find out when the homeowner tells your rep they already signed with someone.

Ask for the clause in writing, and ask these four questions specifically:

  1. Is the record sold to any other buyer in any vertical?
  2. Is it resold after any period of time, including as an aged or remarketed record?
  3. If the homeowner does not convert, does the record re-enter your partner's inventory?
  4. What is the remedy if exclusivity is breached — a credit, or a termination right?

If the answer to (4) is "a credit," exclusivity is not a term of the agreement. It is a service level with a small penalty attached, and it will be violated whenever violating it is profitable.

4. Delivery fit to your actual sales motion

An in-home consultative sale, a call-center close, and a dispatch-based service model each consume demand differently. A partner delivering the wrong format will show acceptable lead-level metrics and unacceptable revenue.

Map it explicitly:

  • If you sell in-home: you need booked appointments with confirmed decision-maker availability and a confirmation cadence. Sit rate is your governing metric, and it is the one most vendors will not discuss.
  • If you close on the phone: you need live transfers with a defined handoff standard — how long the prospect is held, what they were told, whether they know who is calling.
  • If you dispatch: you need volume calibrated to technician capacity by territory and day, not weekly volume dumped into a queue.

Then ask the operational question: can delivery volume be paced to capacity by market, by day, and by hour? Most vendors sell you a monthly volume commitment. A monthly commitment delivered unevenly produces idle reps in week one and abandoned leads in week four, and both are pure margin loss.

5. Compliance posture and where the liability actually sits

This is a diligence item, not a checkbox, and it is the dimension where the gap between vendors is widest.

TCPA class action filings hit record volume through 2025 and 2026, with statutory damages at $500 per violation and up to $1,500 for willful violations. Roughly 80% of TCPA suits are filed as class actions. A class of 100,000 improperly contacted consumers is a $50M to $150M exposure. For a platform carrying acquisition debt, that is not a legal expense. It is a solvency event.

The consent was obtained by your vendor. The call was placed by your company. Ask:

  • Is consent captured and retained at the record level, with the exact disclosure language, timestamp, IP, and the URL where it was given?
  • Can they produce that consent record for a specific homeowner within 24 hours of a demand letter?
  • Are DNC and litigator scrubs run per contact, or per list, or per upload?
  • Are state-level rules applied where they exceed the federal standard?
  • What does the indemnification clause actually cover — and is it capped at fees paid?

That last one decides everything. An indemnity capped at trailing fees is not indemnification against a class action. If a vendor generated your exposure, their willingness to stand behind it is the clearest available signal of how confident they are in their own consent chain.

6. Capacity and geographic elasticity

The last dimension is the one that only matters if you are growing, which is why it gets skipped and then becomes the reason you switch partners in 18 months.

  • Can they produce volume in a market before you have brand presence there?
  • What is the ramp curve in a new metro — first meaningful volume in weeks, or months?
  • Can they hold quality standards while scaling volume 3x in a single territory?
  • If you acquire a brand in an adjacent vertical, does the same program extend to it?

Ask for a market they entered cold and the month-by-month volume and quality curve that followed. A partner who has done it will have the chart. A partner who has not will describe the process.

Scoring it

Weight these to your situation, but a defensible default for an operator with an expansion plan:

Dimension Weight
Source ownership and traceability 25%
Verification depth 20%
Compliance posture and indemnity 20%
Delivery fit 15%
Exclusivity terms 10%
Capacity and elasticity 10%

Price enters after scoring, as a tiebreak between partners who clear the bar — not as a dimension inside it. If you weight price inside the model, it will dominate the model, because it is the only variable expressed in dollars while every other variable is expressed in language.

Three answers that should end the conversation

"We can't disclose our sources — that's proprietary." Source composition is the primary determinant of quality volatility. A partner unwilling to disclose it is asking you to accept unmeasurable risk on an unbounded liability.

"Exclusivity is guaranteed" — with no clause and no remedy. If it is real, it is in the contract with a termination right attached. If it is in the deck only, it is marketing.

"Indemnification is capped at fees paid." This tells you the vendor has priced their own compliance risk and decided not to carry it. You would be carrying it.

The one-sentence version

You are not buying leads. You are buying a rate of revenue production per seat, and a liability position on every phone call your company makes. Evaluate for those two things, and price will resolve itself.

TCPA figures current as of August 2026.

Put this framework to work.

This framework is the same one we use internally when scoping a program. If you want a working version of the scorecard, or a read on how your current sources would score against it, talk to our team.