After six weeks of vendor proposals the per-lead math never works

Ask both of them who writes the script and whether you can change it mid month without a change fee.

Different angle. Under B you are buying management work as well as dials, because three callers with no supervisor drift inside two weeks. Vendor B says they supervise. Do they? One QA person per how many seats? If the answer is vague you are paying $7,200 for three people on their own, and your cost per lead assumption is the optimistic one.

@beacon is right about the recordings but I'd push back on the conclusion. Per-lead pricing is not automatically the worse deal. It moves the contact-rate risk onto the vendor, which is exactly the number juniper says she can't pin down. If you genuinely don't know whether contact is 2 or 4 percent, buying the output instead of the input is a reasonable way to not fund somebody else's experiment. Just cap it. 40 minimum, and a quality clause where leads that fail an agreed test don't count toward the minimum. Vendors hate that clause, which tells you it has teeth.

On the list side: whoever scrubs against the national do-not-call registry and internal suppression, get it in writing which party is responsible and on what cadence. Rules on outbound calling and texting are federal and also state by state, and some states are stricter than others, so confirm with an attorney who does telemarketing compliance rather than with either vendor's sales rep. I've seen a service hand the client the scrubbing obligation in a schedule at the back of the agreement and the client never read it.

Something adjacent that came up when I was going through licensing coursework. Depending on the state and on how the person on the phone is compensated and what they say, prospecting for sellers on someone else's behalf can brush against activity a state requires a license for. It turns on the specific statute and how it defines the activity, so that's a question for a real estate attorney in your state, not a forum. Worth asking Vendor A whose name their callers use, because "we're calling for a local buyer" and "we're calling for juniper" are different conversations.

Actual numbers from my floor, six seats, last full quarter. 3,800 dials a day across the team on a three-line dialer. Contact rate 4.1 percent for skip traced numbers, and that drops to about 2.6 percent on records older than 90 days, which is the part your model is missing. Fresh data is doing most of the work in that 4 percent.

Conversion from live conversation to something I'd bill as a lead: 2.2 percent. From lead to signed contract for my clients: about 1 in 9, and that's on them not me. All in, my cost to produce one lead is $94 including data, dialer minutes, payroll and my own QA time. So $75 from Vendor A is under my cost. Either they have cheaper labor than I do, or their definition is doing the work. Both are possible and neither is a reason not to buy, you just need to know which.

Turnover is what will actually break your seat model, not contact rate, and five years of running a service is how I learned it. A caller who's been on your script four months produces roughly twice what a three week caller does on the same list. So when you compare $2,400 a seat, ask Vendor B the average tenure of their current callers. If it's under 90 days you're renting a training program. @cobalt's $94 is believable but it assumes a stable bench, and that's the assumption I'd interrogate before the 4 percent.