# Analytics benchmarks that matter: from days-in-inventory to churn
A dealer group in Ohio bragged about "record inventory turns" until someone pulled the raw feed: their 42 days' supply looked lean, but the industry sat near 60 days at the time. They were not efficient. They were understocked, losing sales they never recorded. That is the danger of a dashboard without a baseline. A number means nothing until you know what "normal" looks like for the automotive sector.
This lesson calibrates the measurement benchmarks automotive executives actually watch, so your dashboards compare against realistic sector reality, not gut feel.
Every metric below is trivial to calculate. The hard part is the data feeding it. Before you trust a benchmark, ask three questions:
Mismatched definitions across rooftops (individual dealership locations) are the single biggest reason group-level dashboards mislead.
Days' supply answers: at the current sales rate, how many days until we sell out?
Formula:
Days' supply = (Units in inventory / Units sold in period) x days in periodWorked example. A dealer holds 300 new units and sold 150 last month (30 days):
Days' supply = (300 / 150) x 30 = 60 daysDays-in-inventory is different: how long a *specific* car has physically sat on the lot. Days' supply is a forward-looking ratio; days-in-inventory is a backward-looking age. Confusing the two is a classic dashboard error.
Benchmarks (US, as of late 2025, industry estimates): new-vehicle days' supply commonly runs in the 50 to 70 day range for a healthy market, though it swings hard by brand and segment. During the 2021 to 2022 chip shortage it collapsed below 30. European figures are harder to standardize because the market skews toward build-to-order, especially in Germany, so days' supply is a less dominant metric there than in the US stock-heavy model.
Cox Automotive publishes regular US data-drivendata-drivenAn approach where decisions are systematically informed by data analysis rather than intuition alone.Voir la définition complète → commentary worth bookmarking: Cox Automotive market insights.
Rule of thumb: below 30 days signals shortage and lost sales; above 90 signals overstock, aging units, and floorplan carrying cost pressure.
Fixed-first-visit (FFV), also called first-time-fix, measures the share of repair jobs resolved on the first visit without a comeback for the same issue.
FFV rate = (Repair orders fixed on first visit / Total repair orders) x 100Why it matters as data: it is a leading indicator of both customer satisfactioncustomer satisfactionCustomer Satisfaction Score, a direct measure of satisfaction captured right after a specific interaction or experience, usually on a short rating scale.Voir la définition complète → and warranty cost. A low FFV means repeat visits, wasted technician hours, and inflated warranty claims that OEMs scrutinize.
The measurement trap: what counts as a "comeback"? If a customer returns in 7 days it is clearly a repeat. At 30 days it may be a new fault. Fix the window in your definition and apply it identically across every rooftop, or your comparison is noise. There is no single public sector benchmark for FFV; targets are usually set internally by OEMs, often in the high 80s to low 90s percent range as an aspirational floor. Treat any external FFV number with suspicion unless the window and scope are stated.
Service retention measures how many vehicle owners keep coming back for maintenance rather than defecting to independent shops or chains.
A common definition:
Service retention = (Customers with a service visit in period / Eligible customers in period) x 100The words "eligible" and "period" carry all the weight. Do you count a 6-year-old car whose owner uses a local independent shop? Rolling 12-month windows are standard. Longer windows flatter the number.
Retention is where the money is. Service and parts generate a large and stable share of dealership gross profit, and a retained service customer is far likelier to buy their next vehicle from you. That is why OEM dashboards track it obsessively.
Data sources to reconcile:
The gap between "due for service" and "actually visited" is your defection signal.
Net Promoter Score (NPS) asks one question: "How likely are you to recommend us, 0 to 10?" Promoters score 9 to 10, passives 7 to 8, detractors 0 to 6.
NPS = % Promoters - % DetractorsWorked example. Of 200 survey responses: 120 promoters (60%), 40 passives (20%), 40 detractors (20%):
NPS = 60 - 20 = +40NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters.Voir la définition complète → ranges from -100 to +100. Automotive NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters.Voir la définition complète → varies widely by brand; premium and EV-native brands often post notably higher scores than mass-market brands, but published cross-brand figures are inconsistent and methodology-dependent, so treat any headline NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters. as an estimate tied to a specific survey.
Churn is the mirror image, and in automotive it comes in two flavors:
1. Service churn: owner stops using your service bay (visible in DMS gaps).
2. Brand churn (defection): at trade-in or lease-end, the customer switches brands. This is captured in conquest and defection data, often sourced from registration datasets (for example S&P Global Mobility registration data in the US and Europe).
The mistake is watching each metric alone. They interact:
A useful discipline: for every benchmark on your executive dashboard, store three fields alongside the value: source system, definition version, and as-of date. Without those, a green number and a red number are not comparable.
Simple governance snippet for a metric definition registry:
metric: service_retention
definition_version: 2.1
formula: serviced_customers / eligible_customers
window: rolling_12_months
eligible_rule: vehicles_under_7yr_with_active_owner
source_systems: [DMS, CRM, OEM_warranty]
as_of: 2026-01-15This tiny file is what stops two rooftops from reporting "retention" that means two different things.
Vérification des acquis
1. The Ohio dealer group celebrated a low days' supply as a sign of efficiency, but analysis revealed the opposite problem. What core lesson does this illustrate?
2. A dashboard shows a car's 'days-in-inventory' as 45 days and the store's 'days' supply' as 60 days. What is the correct interpretation of the difference between these two metrics?
3. Why does the lesson describe benchmarks as 'a data-quality problem, not a math problem'?
4. Select ALL correct answers. Before trusting a benchmark, the lesson recommends asking which of the following questions about the underlying data?
Sélectionnez toutes les réponses correctes.
5. Select ALL correct answers. Why do mismatched metric definitions across rooftops cause group-level dashboards to mislead?
Sélectionnez toutes les réponses correctes.
Three cautions before you present any of these numbers upward.
Segment before you compare. A pickup truck franchise and a luxury EV showroom have structurally different days' supply and service patterns. Blending them hides the story.
US and Europe differ structurally. The US stock model produces meaningful days' supply figures; much of Europe's build-to-order flow makes that metric less central and less comparable. Do not port a US benchmark onto a German dashboard without adjustment.
Beware survivorship in retention and NPS. Customers who already left do not answer your survey. A rising NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters.Voir la définition complète → can simply mean your detractors churned out of the sample. Cross-check NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters. against churn data, never in isolation.