What 300,000 cold calls a month taught us about list quality (real numbers inside)

Question

I've been managing outbound calling for wholesaling operations for a while now — currently sitting around 300,000 dials and 40,000 connects a month, mostly Texas and Tennessee.

Every cold calling benchmark in this industry traces back to somebody's course sales page, so we finally went through our own numbers properly. Four things stood out.

Connect rate is a data problem before it's a caller problem. We hold about 13% across everything, but that average hides the real story. Fresh skip trace vs stale list, same agents, same script — the gap is enormous. Most teams see a bad connect rate and start rewriting scripts or replacing callers. Usually the list is the variable.

Cost per record is a trap. A dead record costs you the record plus the agent minutes spent burning through it. When we started measuring cost per connect instead, our vendor ranking flipped completely. The cheapest source stopped being the cheapest.

Connects are not right party contacts. Someone answering isn't the owner answering. If you manage callers on connects but pay for leads, you're measuring one thing and buying another — and you'll eventually lose a good caller over a number that was never theirs.

The one I didn't enjoy: we spent a stretch responding to weak results by adding dial volume. It did close to nothing. Teams with a disciplined callback queue out-produce teams with more dials and no cadence, every time. More volume on a leaky follow-up process just leaks faster.

Two things I'm still working on and would take input on: the right cadence between call, text and voicemail drop without wrecking your numbers on A2P compliance, and whether 13% is actually normal at this volume or whether we're an outlier.

What connect rate are you holding, and off what data source? I'd rather this thread produce real numbers than the ones on the sales pages.

Michael Eskenasy · Answer

One thing I’d add to your numbers that I think gets missed a lot:

13% connect rate may not actually be the metric to optimize. The more useful number is the decay curve underneath it.

At your volume, I’d want to know the probability of a verified right-party conversation by:

• data source
• days since skip trace
• attempt number
• call vs. callback
• market
• time/day
• eventual downstream outcome

Because two lists can both show a 13% connect rate and be completely different assets.

One might produce most of its owner conversations in attempts 1–2 and then die. Another might look mediocre initially but keep producing through callbacks and later attempts. If you only rank them on first-pass connect rate or cost per record, you can kill the better list by accident.

That’s basically why we do this the way we do it.

We keep each record as a living history instead of treating every dial as an isolated event: source, trace date, attempts, channel, dispositions, right-party contact, callbacks, qualification, offers, and eventual outcome. Then we can query the population later and see where the actual drift is.

So instead of asking, “What’s our connect rate?” we can ask:

“At what age and attempt count does this source stop producing economically useful owner conversations?”

That tells you when to recycle, retrace, switch cadence, or stop burning agent time.

And I think your point about callback discipline is probably connected to this more than it looks. More volume on a bad queue leaks faster, but a good queue can actually extend the economic life of the list.

If you want to connect, I’d be interested in comparing how you’re storing the attempt/callback history against the way we structure it. At 300,000 dials a month, you probably have enough volume to see patterns most people literally can’t see.

James Wise · Answer

300,000 cold calls? You should give us your office location so we can drive there and slap you for spamming us.

Igor Ganapolsky · Answer

Mohamed, your first insight — that connect rate is a data problem before it's a caller problem — is exactly what led me to tax deed lists. The logic follows your framework perfectly:

If the list is the variable, then the best list is one where the government has already identified, legally notified, and processed the motivated sellers for 1-3+ years. County tax deed lists are published on a schedule (auction dates set, parcels listed publicly). The data is as fresh as it gets because the county just published it. No skip tracing needed — owner names and mailing addresses are public record on the county assessor/appraiser site.

Your cost-per-record trap becomes irrelevant when the record costs $0. Broward County FL tax collector website publishes hundreds of parcels for Auction #113 (October 26, 2026) — free. Marion County IN auditor publishes the annual tax delinquent list — free. Tarrant County TX, Cuyahoga County OH (Cleveland), Lucas County OH (Toledo) — all free public records.

The model I'm working with: pull the tax deed list (free) → cross-reference county assessor values (free) → calculate spreads between tax debt and market value → connect cash buyers who want to bid at auction. No cold calling because the sellers are already identified by the county legal process. The "connect" happens when a cash buyer reviews the spread and decides to bid.

Your 13% connect rate benchmark is impressive for cold calling, but tax deed auctions flip the question — instead of asking "can I reach the owner?" you're asking "does the spread work for a cash buyer who can close same day or within 24-48 hours?" That's a different conversation with a different type of buyer.

Curious whether your callers have tested tax deed auction lists as an alternative to traditional wholesaling lists — same research discipline (measure cost per connect, track fresh vs stale), but the data source is county government instead of third-party skip tracers.

Jack T. · Answer

Your flooding the threads with the same Ai slop about tax deeds with different variations is rather frustrating

Mohamed Hanafy · Answer

Lydia, thank you so much for taking the time to reply to my post.
I use mainly Deal Machine as our callers are assigned 15-20k homeowners on a monthly basis. I have tested Batchleads by Propstream & Data Sift but have seen far better results with Deal Machine specially while dealing with such volume at sosbpo.com. Separation is mainly measured by connection rate, accuracy & conversion to lead. What I mainly do is hire cold callers for wholesalers, investors and real estate agents. Each cold caller does 5000 dials on a daily basis generating 30+ leads monthly. I agree PPL is more expensive and gives less volume but more valuable. Can I know how much where you investing on cold calling? Were you doing it yourself or with an off-shore team? What was the bottleneck you faced with cold calling?

Lydia R. · Answer

Thanks for sharing this information, its very interesting! I've mostly pivoted away from cold calling to focus entirely on PPL. Its more expensive but it delivers more warm leads in a much shorter amount of time. I would love to understand more about your cold calling operation. Who are you using for skiptracing to get the most accurate data? Do you skip the same list with multiple tracers and compare the accuracy? How do you separate a bad lead list from bad phone numbers? How do you narrow down which is the problem so you can solve it?

What 300,000 cold calls a month taught us about list quality (real numbers inside)

What 300,000 cold calls a month taught us about list quality (real numbers inside)

Most Popular Reply

9 Replies