What should my cost per item on leads be?

 Cost per lead is the wrong metric to optimize. Here are the four numbers that actually tell you whether your lead spend is working. 

When an agency owner asks, "what should my cost per lead be?" the question sounds reasonable. But the honest answer is that the cost per lead number by itself is almost meaningless.

Here's why. A $10 lead that closes at 3% is more profitable than an $8 lead that closes at 0.5%. The cheaper lead costs more to run a business on. And if you're making sourcing decisions based on the $8 number, you're actively destroying ROI while thinking you're being smart about budget.

The question to ask isn't what a lead should cost. It's what your cost per item, cost per sale, and ROI look like, and whether those numbers are healthy enough to justify spending more, not less.

What actually determines lead cost

Cost per lead in the third-party internet lead market is driven by supply, demand, and the filters you select. That's it.

Wide filters mean a large pool of available leads. A large pool means competition among buyers is lower, and prices stay reasonable. Tight filters,  specific zip codes, confirmed homeowners, particular credit tiers, and certain vehicle profiles shrink the supply pool. When supply is limited, and demand stays constant, the price goes up. You pay to play for the targeting.

There's also a meaningful difference in lead type that affects what you should expect to pay and what you should expect to convert.

The first-party and organic rows in that table are worth dwelling on for a moment. Most of the industry conversation around lead costs focuses entirely on third-party internet leads. But agencies that invest in their own content, their own ad funnels, or their own SEO are generating leads that convert at significantly higher rates because the prospect came to them specifically. The cost math is harder to calculate on a per-lead basis, but the ROI tends to be superior. It's a longer game and requires different skills, but it changes the cost-per-item picture entirely for the agencies that build it.

The four metrics that actually matter

Stop anchoring decisions on cost per lead. These are the four numbers that tell you whether your lead operation is healthy.

Tracking cost per item, cost per sale, and ROI across multiple vendors and lead types manually is unsustainable. NCC's Data Dashboard surfaces these numbers in real time.

Cost per item is the metric most agencies should be running their business on, day to day. It answers the question every insurance agency owner actually cares about: how much am I spending to bind one policy? That number accounts for lead cost, contact rate, quote rate, and close rate all at once. It's the efficiency metric.

ROI is the override. You might look at your cost per item and feel like it's running high. Without ROI in the picture, that's enough to make you nervous. But if your ROI is 50% or better, the operation is profitable. The business is working. High cost per item in a high-ROI operation isn't a problem. It's just an optimization opportunity.

The vendor comparison mistake that costs real money

Here's the most common and most costly mistake we see when agencies evaluate lead vendors: comparing cost per lead without comparing close rates.

Vendor A charges $8 per lead. Vendor B charges $10. The agency cuts Vendor B.

But Vendor A's leads close at 0.5% and Vendor B's leads close at 3%. At those close rates, a 100-lead sample from Vendor A produces half a closed deal. The same 100 leads from Vendor B produce three. The $10 lead was cheaper to operate on by every measure that actually matters.

Once you know the ROI for each vendor, the conversation changes entirely. You stop asking "is this lead too expensive?" and start asking "how much more can I spend here before returns start to compress?" That is a fundamentally different question, and the agencies that get there make better sourcing decisions with the same budget. When you're comparing close rates across vendors, lead quality affects the inputs. What goes into an NCC lead quality check? explains what separates a lead that's likely to close from one that isn't before you ever dial. 

Adding an SDR: how the math changes

When you layer an SDR operation on top of lead costs, the fully loaded cost per acquisition goes up on paper but should go down in practice because a well-run SDR team improves contact rates, qualification, and warm transfer quality enough to move the cost per item in the right direction.

The key question is whether the SDR operation is performing.  An SDR team that isn't moving your contact rate doesn't justify the added cost. You're paying for SDR labor, your lead cost hasn't changed, and your contact rate and transfer quality haven't improved. Every dollar of SDR cost sitting on top of a flat contact rate is pure margin erosion.

The math only works when SDR performance is producing meaningful contact rate improvement and delivering genuinely warmer prospects to your closing agents. When it does, the fully loaded cost per acquisition often comes down even though the gross spend went up, because the conversion rate at every stage of the funnel improves.

At what volume does adding an SDR make sense? The honest answer is that it depends on whether your current team has a contact rate and speed-to-lead problem that SDRs can solve. If you're already hitting 35% contact rate with a well-structured internal team, an SDR layer may not move the needle enough to justify the cost. If you're stuck at 15% because internal reps are splitting time between dialing and quoting, an SDR team that handles the front-end calling exclusively can change the economics significantly.

What your data needs to look like before these numbers mean anything

Before you can trust any of these metrics, you need two things: enough leads and enough time.

The minimum sample size before cost per item and ROI figures are meaningful is 400 to 500 leads. Below that threshold, you're concluding noise. One strong week or one slow closer skews the entire dataset in ways that won't reflect the true underlying performance.

The other trap is pulling leads into your analysis too early. Never calculate cost per item using leads from the last five to seven days. Those leads are still working through your call cadence and sales cycle. Include them, and they'll drag your close rates down artificially, making a healthy system look broken.

In practice: if you're buying 100 leads per week, you need roughly a month of data before your numbers are trustworthy. Track everything from day one, but hold your conclusions loosely until you hit that threshold. You're still collecting, not deciding.

How to reverse engineer what you can afford to spend

Once you have clean data, the calculation runs backward from where you want to end up.

• Start with your target profit per bound policy. What do you need to make on each item to justify the lead spend?
• From there, calculate the maximum cost per item your margins can support. This is your ceiling.
• Now apply your close rate. If you're closing 2% of leads into bound policies, and your maximum cost per item is X, then your maximum cost per lead is X multiplied by 0.02.
• That's the number you can afford to spend. Not the industry average, not what the vendor charges by default. The number that comes from your own margin requirements and your own close rate.

Most agencies never run this calculation. They either accept whatever the vendor charges or cut based on sticker price. Neither approach is a business decision. The reverse engineer is.

If you want to run these numbers against your current lead spend and see where cost per item and ROI actually stand, reach out to our team. We do this analysis regularly with new and existing NCC partners, and the picture that emerges is almost always more actionable than whatever the agency thought they knew going in.

 This blog is a collaborative piece by: