Cost per Sold Unit and Cost per Lead: How Dealers Calculate Them by Source

Cost per sold unit is a lead source's marketing spend for a period divided by the vehicles sold to customers from that source in that period. Cost per lead is the same spend divided by the source's leads. Dealers should calculate both for each source, because cheap leads can still cost the most per sale.

At a glance

• Cost per lead equals source spend divided by leads from that source, and cost per sold unit equals source spend divided by units sold from that source.
• NADA Data 2025 puts average franchised-dealer advertising at $586,246 per dealership, but that figure blends every channel and every sale, so it cannot tell a dealer which lead source earns its budget.
• In the worked example below, the source with the cheapest leads has one of the highest costs per sold unit, which is why budget decisions should rest on cost per sale, not cost per lead.
• Sales that touched more than one source need one written credit rule (first touch, last touch or split credit) applied the same way every month.
• Small numbers mislead: a source with two or three sales in a month needs a rolling three-month view before anyone moves money.

The formulas for cost per lead, cost per sold unit and close rate

Cost per lead, close rate and cost per sold unit are three linked numbers, and each one is calculated per lead source for the same period, usually a calendar month. Write the definitions down once so the sales manager, the marketing manager and the accounting office all calculate them the same way.

• Cost per lead (CPL): source spend for the month divided by unique leads from that source in the month.
• Close rate: units sold to leads from that source divided by unique leads from that source.
• Cost per sold unit (also called cost per sale or advertising per unit): source spend for the month divided by units sold to customers credited to that source.
• The link between them: cost per sold unit also equals cost per lead divided by close rate. A source with leads at $40 and a 5% close rate costs $800 per sold unit.
• Optional middle steps: cost per appointment set and cost per shown appointment, which use the same spend divided by appointments set or appointments that showed.

The last formula explains most arguments about lead sources. Cutting cost per lead in half does nothing for cost per sale if the close rate on the cheaper leads also falls by half. David Boice, the CEO of marketing technology vendor Team Velocity, makes the same point in a May 2026 Digital Dealer column: cost per lead measures volume, not value.

What the industry average cost per sold unit does and does not tell you

The industry average cost per sold unit for franchised new-car dealers comes from NADA, and it is useful as context but not as a target for any single lead source. The NADA Data 2025 annual financial profile reports total dealership advertising of $9.96 billion and an average of $586,246 per dealership. Inside Radio's June 2026 summary of NADA Data 2025 reports average advertising of $739 per new vehicle sold.

The NADA figure has three limits a manager should keep in mind before comparing a store against it:

• The NADA figure is blended. NADA divides all advertising by all new units, including repeat buyers, referrals and drive-by customers that no paid source produced. A store's cost per sold unit for one paid source will usually be higher than its blended number, because the paid source gets credit for fewer sales.
• The NADA figure covers franchised new-car dealers. RV, marine and powersports dealers, and independent used-car lots, are not in that dataset, so the $739 average is not a benchmark for them.
• The NADA figure is per new unit. Many stores spend a large share of their budget selling used vehicles, so a store that counts used units in its own calculation is not comparing like with like.

Rob Anderson raised the blending problem in Auto Dealer Today in 2008, arguing that dealers should remove repeat, referral and drive-by sales before dividing. The article is old and its percentages are the author's assumptions, not data, but the method still holds.

There is no reliable published benchmark for dealer cost per lead by source. Vendor articles that quote cost-per-lead ranges, such as Autocorp's June 2026 article, do not say where their figures come from. Your own three-month history by source is a better yardstick than any outside range.

What to count as cost and what to count as a sale

Cost per sold unit is only comparable from month to month if the dealership decides in writing what goes into spend and what counts as a sale. Most disagreements about which source "works" turn out to be disagreements about these inputs.

What belongs in source spend?

• Include media spend (search, social, display, TV, radio, direct mail), listing-site subscriptions, per-lead fees paid to lead providers, agency or management fees tied to that source, and call-tracking or landing-page costs used only for that source.
• Decide on co-op. Choose whether to record spend before or after manufacturer co-op reimbursement, then apply the choice to every source. After reimbursement shows the store's true cost; before reimbursement makes month-to-month comparisons easier when co-op arrives late.
• Keep payroll separate. BDC and sales pay is a real cost of working leads, but it is shared across sources. Report it as its own line rather than spreading it by guesswork.

What counts as a sold unit?

• Count retail deliveries only. Leave out wholesale units and dealer trades.
• Decide whether a deal that unwinds is removed from the month it was sold or from the month it unwound, and stay consistent.
• Report new and used separately if the budget is split that way.
• Count a sale for a source only when the CRM record shows the buyer came from that source, not when a salesperson remembers it that way later.

A worked example: cost per lead and cost per sold unit by lead source

The worked example below uses invented round numbers for one month at a single store, chosen only to show the arithmetic; none of the figures are benchmarks. The store spent $23,500 on five paid sources and delivered 70 retail units in total, 29 of them credited to a paid source under a last-touch rule.

Source A: third-party listing site

• Spend $6,000, 150 leads, 9 units sold.
• Cost per lead $40, close rate 6%, cost per sold unit $667.

Source B: paid search

• Spend $8,000, 100 leads, 10 units sold.
• Cost per lead $80, close rate 10%, cost per sold unit $800.

Source C: social media lead ads

• Spend $3,000, 120 leads, 3 units sold.
• Cost per lead $25, close rate 2.5%, cost per sold unit $1,000.

Source D: manufacturer lead program

• Spend $2,500, 50 leads, 5 units sold.
• Cost per lead $50, close rate 10%, cost per sold unit $500.

Source E: direct mail

• Spend $4,000, 20 leads (calls and QR-code visits), 2 units sold.
• Cost per lead $200, close rate 10%, cost per sold unit $2,000.

What the worked example shows

• The cheapest leads were not the cheapest sales. Source C had the lowest cost per lead ($25) but cost $1,000 per sold unit, more than Sources A, B and D.
• The paid-source average is far above the blended figure. All paid sources together cost $810 per credited sale ($23,500 divided by 29). Dividing the same spend by all 70 units gives $336, which flatters the budget because 41 buyers came from referrals, repeat business, walk-ins or unknown sources.
• Two sales are not a verdict. Source E looks expensive, but one more sale would cut its cost per sold unit to $1,333. Hold judgment on any source with fewer than about five sales until three months of data agree. This threshold is a rule of thumb, not a statistical standard.

How to count sales that touched more than one lead source

Sales that touched more than one lead source need a credit rule the dealership chooses before the month starts, because the same deal can make two sources look good or neither. A buyer might submit a listing-site lead on day 1, click a paid-search ad and fill in a website form on day 12, and buy on day 20.

• First touch: the listing site gets the whole sale. First touch favors sources that introduce new shoppers.
• Last touch: paid search gets the whole sale. Last touch favors sources that catch shoppers who are ready to buy, and it is the simplest rule to apply from CRM data.
• Split credit: each source gets half a sale. Split credit is fairer but harder to explain, and it produces fractional unit counts.

Ad platforms use their own rules, and they will not match the CRM. Google's help page on attribution models in Google Ads lists two current models, last click and data-driven, and says first click, linear, time decay and position-based models are no longer supported. Each platform counts only the touches it can see, so the conversions each platform reports will usually add up to more sales than the store actually made.

Our recommendation for most stores is practical rather than perfect:

1. Use the CRM, not the ad platforms, as the record of which source a buyer came from.
2. Pick last touch or first touch as the primary rule for budget decisions and write it down.
3. Report a second column, "sales influenced," that counts every source the buyer touched, so a source that often starts a deal is not cut by mistake.
4. Ask every buyer how they heard about the dealership and record the answer, but use the answer to check the CRM, not to replace it. Showroom answers are often vague ("online", "Google") and should not decide a single campaign on their own.
5. Set a lookback window, for example a sale counts for a source if the lead arrived within 90 days of the sale, and apply it to every source.

A spreadsheet layout for tracking cost per sold unit each month

A monthly cost-per-sold-unit spreadsheet needs one row per lead source and one tab per month, with the same columns every time. Keep the source names identical to the lead source values in the CRM so the counts can be copied straight from a CRM report.

• Column A, Source: the exact lead source name used in the CRM.
• Column B, Spend: spend for the month under your written cost rules.
• Column C, Unique leads: leads after duplicates are removed.
• Column D, Appointments shown: optional, but useful for spotting a follow-up problem.
• Column E, Units sold (primary rule): retail units credited under first or last touch.
• Column F, Units influenced: retail units where the source appears anywhere in the buyer's history.
• Column G, Cost per lead: =B2/C2
• Column H, Close rate: =E2/C2
• Column I, Cost per sold unit: =IF(E2=0,"no sales",B2/E2)
• Column J, Three-month cost per sold unit: the sum of spend for the last three months divided by the sum of units sold for the same months.
• Column K, Notes: anything unusual, such as a campaign paused mid-month or a co-op credit received late.

Add a final row for sales with no paid source (referral, repeat, walk-in, unknown) with zero spend. A rising count of "unknown" sales is a data problem to fix in the CRM before the spreadsheet can be trusted. Gross profit per unit by source is a worthwhile extra column if the accounting office can supply it, because a source that sells low-gross units can cost more than its cost per sold unit suggests.

Why a lead source's cost per sold unit changes when follow-up changes

A lead source's cost per sold unit depends on the dealership's own follow-up as much as on the source, because close rate sits in the formula. A source can look expensive simply because its leads arrive after hours and wait until morning for a reply. Before cutting a source, check how fast its leads were answered and how many got a second and third attempt; our guide to measuring dealership lead response time shows how to pull that number by source.

The same caution applies to phone leads. Calls that ring out never become CRM leads unless someone records them, so sources that drive phone calls can look worse than they are. The missed call text-back guide explains how to count unique missed callers.

If you want an outside view of where leads are slipping between the ad and the sale, the Revenue Leak Audit reviews lead response, follow-up and conversion with your team at no cost.

Where Shiftix Cloud fits in cost-per-lead and cost-per-sale tracking

Shiftix Cloud can supply the lead-side inputs for these formulas, while sold units still come from the dealership's deal records and DMS. Shiftix Cloud is not a DMS; it works alongside the DMS and accounting systems a store already uses.

• Marketing Monitor: the marketing page describes reviewing available impressions, clicks, leads, spend and cost-per-lead data across connected campaigns, with campaign and ad-level detail where connected, and carrying campaign and source details into the CRM opportunity where available. The page also says the tool complements advertising and listing platforms rather than replacing them.
• CRM: the CRM page describes keeping lead source, vehicle interest and recent activity together on the opportunity record, and routing leads by conditions that include source.
• Connections: the integrations page lists Google Ads and Meta / Facebook under marketing, and lead providers such as Cars.com, AutoTrader, CarGurus, RV Trader and Boat Trader as lead sources. The page describes these as examples of potential connection targets, with availability confirmed during implementation.

Next steps

The next step depends on what your current numbers look like:

• If you only track cost per lead today, add units sold by source for the last three months, using last touch from the CRM, and calculate cost per sold unit before the next budget meeting.
• If many sales show "unknown" or "walk-in" as the source, fix lead source capture in the CRM first. The automotive CRM guide covers what a CRM should record.
• If one source has cheap leads and few sales, check its response times and follow-up before cutting it; the BDC manager's weekly scorecard shows which handling numbers to review.
• If the BDC works paid leads, compare cost per shown appointment as well; the guide to what a car dealership BDC does explains how BDC costs fit into cost per sale.

Frequently asked questions

What is a good cost per lead for car dealerships?

A good cost per lead for a car dealership is one that produces an acceptable cost per sold unit at the store's own close rate, so there is no single right number. A $100 lead that closes at 10% costs $1,000 per sale, while a $30 lead that closes at 2% costs $1,500. Published cost-per-lead ranges from vendors rarely state their data source, so compare each source against its own three-month history instead.

What is the difference between cost per lead and cost per sale?

Cost per lead divides a source's spend by the leads it produced, while cost per sale divides the same spend by the vehicles sold to those leads. Cost per lead shows how much attention a source buys; cost per sale shows what each delivered vehicle actually cost in marketing. The two are linked: cost per sale equals cost per lead divided by close rate.

Should BDC and sales payroll be included in cost per sold unit?

BDC and sales payroll are usually kept out of cost per sold unit by source and reported as a separate line. Payroll is shared across every lead source, so splitting it by source requires estimates that can distort the comparison. Some stores add a fully loaded cost per sale for the whole store, which includes payroll, as a second measure for the dealer principal.

Every dollar spent on a lead is only recovered if the lead is answered, followed up and worked to a sale. The Revenue Leak Audit looks at lead response, follow-up, conversion, customer retention, accountability and workflow. It is complimentary, carries no obligation and needs no software purchase, and Shiftix Cloud walks through the findings with you.

When your numbers show which sources earn their budget, more shoppers may be worth reaching. Searchshop, a vehicle marketplace run by the same team as Shiftix Cloud, is where buyers search dealer inventory, and its dealer listing packages put a dealership and its inventory in front of those shoppers.

Sources

• NADA, NADA Data 2025: Annual Financial Profile of America's Franchised New-Car Dealerships
• Inside Radio, Turning Data Into Dollars: Key Insights From 2025 NADA Data (June 2026)
• Digital Dealer, When "Cost Per Lead" Lies: Smarter Metrics for Dealer Marketing Performance (May 2026)
• Auto Dealer Today, What Should My Cost Per Sale Be? (March 2008)
• Autocorp, article on what car dealers pay for leads (June 2026)
• Google Ads Help, About attribution models
• Shiftix Cloud, Marketing Monitor
• Shiftix Cloud, CRM
• Shiftix Cloud, Integrations