Multi-Source Fitment Intelligence
Reducing Lost-Sale Rate from 34% to 8% on Complex Fitment Resolutions
Industry
Automotive & Heavy Equipment
Scale
$320M Revenue
Duration
20 Weeks
Location
Memphis, Tennessee
Engagement
AI Consulting
Executive Summary
The VP of Sales at a 20-branch automotive and heavy equipment parts distributor in Memphis had a revenue problem disguised as a knowledge problem. When a fleet maintenance manager called needing a starter for a 2019 Peterbilt 579, the counter person's job was just beginning — that chassis could have a Paccar MX-13, a Cummins X15, or a legacy Caterpillar engine, each requiring a completely different starter from different manufacturers at different price points. The 4 senior counter specialists who could resolve complex fitment from memory closed the sale. The other 28 counter and inside sales staff put the customer on hold, walked to a specialist's desk, and hoped they weren't already helping someone else. 34% of complex fitment inquiries that took longer than 10 minutes resulted in a lost sale — the customer called the next distributor. We embedded ML-driven fitment intelligence into the distributor's counter and phone workflow on Karmak Fusion.
Business Impact
34%→8%
Lost-sale rate on complex fitment inquiries
$2.8M
Revenue retained on orders that would have walked
62 sec
Average complex fitment resolution time, down from 14 minutes
7.2%→2.1%
Return rate from wrong-part shipments
The Situation
The distributor served fleet operators, independent repair shops, dealership service departments, and field mechanics across Tennessee, Mississippi, Arkansas, and Northern Alabama. The catalog spanned 80,000+ active SKUs across light-duty automotive, Class 4-8 heavy-duty truck, and off-highway equipment applications from 200+ manufacturer lines — OE, aftermarket, and remanufactured.
The counter team fielded an average of 220 fitment inquiries per day across 20 branches. 40% were straightforward — light-duty, VIN decodes cleanly, ACES fitment data resolves it. The other 60% — heavy-duty, mixed-fleet, off-highway, and older equipment — required knowledge that lived in 4 people's heads and nowhere else.
- For Class 4-8 trucks, VIN decoded to the chassis manufacturer but not to the engine, transmission, or auxiliary configuration. A Kenworth T680 could have a Paccar MX-13 or a Cummins X15 — each requiring a different starter, alternator, turbocharger, and filter set. Determining the powertrain required asking the customer to read a tag off the engine block or looking up the serial number range in a manufacturer-specific system
- For off-highway equipment — Caterpillar loaders, John Deere excavators, Komatsu dozers — there was no standardized VIN decode at all. Fitment depended on serial number breaks that determined which engine, hydraulic system, and electrical configuration the machine left the factory with. A CAT 950 wheel loader from 2018 with a C7.1 engine had different filters, hoses, and injectors than the same model built in 2016 with a C6.6
- Once the application was identified, the counter person had to cross-reference across OE, aftermarket (Denso, Bosch, BBB Industries, WAI Global), and remanufactured options — each with different pricing, lead times, core charges, and warranty terms. The senior specialists could do this from memory across 40+ years of production. The other 28 staff toggled between manufacturer catalogs, Karmak lookups, and phone calls to manufacturer tech lines that averaged 20-40 minute hold times
- Wrong-part shipments ran at 7.2% across all branches — driven almost entirely by fitment errors on heavy-duty and off-highway applications where the customer provided incomplete information, the counter person made an assumption about the configuration, and the assumption was wrong. Each wrong-part shipment cost $85-$240 in return processing, restocking, and reshipping — plus the relationship damage of a mechanic standing next to a truck that's still down
- The 4 senior specialists had been with the distributor an average of 18 years. They were the most operationally critical people in the business. When one was on vacation, that branch's lost-sale rate on complex inquiries doubled
The distributor's value proposition was expertise — the fleet manager called them because they could identify the right part for any truck or machine in the yard. That expertise was real. It was also locked inside 4 people serving 20 branches.
The Challenge
The VP of Sales tracked a single week in September 2024 to quantify the problem. Across all 20 branches, the counter teams received 1,087 fitment inquiries. Of those, 643 were complex — heavy-duty, off-highway, or multi-source cross-reference required. Resolution data showed a clear split:
When a senior specialist handled the inquiry directly, average resolution was 3-4 minutes and the close rate was 91%. When a non-specialist handled it — looking up catalogs, calling manufacturer tech lines, walking to a specialist's desk — average resolution stretched to 14 minutes and the close rate dropped to 66%. The 34% lost-sale rate on extended-resolution inquiries represented the customers who said "let me call you back" and never did, or who called a competitor while on hold.
- The 4 specialists couldn't scale. Each handled their branch's complex inquiries plus phone requests from 2-3 adjacent branches. During peak hours, the queue backed up and non-specialists defaulted to "let me research this and call you back" — a response that lost the sale 78% of the time on urgent jobsite requests
- Heavy-duty fitment was the highest-value and most complex category. A turbocharger for a fleet truck was a $1,200-$3,500 sale. A wrong turbo — wrong turbine housing, wrong compressor wheel, wrong wastegate configuration — was a $300 return cost plus a mechanic who lost a day of labor waiting for the right part. The margin on getting it right was enormous. The cost of getting it wrong was equally large
- Off-highway was the fastest-growing segment and the hardest to serve. Construction and mining equipment parts carried 30-35% gross margins — the highest in the catalog — but the fitment complexity was extreme. Serial number breaks, field modifications, and proprietary manufacturer numbering systems meant that even the senior specialists sometimes had to call the OE dealer to confirm fitment. The distributor was winning off-highway business on expertise but couldn't scale that expertise beyond the branches where the specialists sat
- The VP had tried creating a "fitment guide" — a binder of cross-reference charts organized by equipment type. It was 400 pages within 3 months, already outdated on 15% of its content, and nobody used it because flipping through a binder while a customer waited on hold was slower than walking to the specialist's desk
The Solution
We spent 5 weeks in discovery across 4 branches — the Memphis headquarters, the Nashville branch, the Little Rock branch, and the Jackson, Mississippi branch — observing how counter staff handled complex fitment inquiries and documenting the decision logic the senior specialists applied.
The critical finding: the specialists' fitment resolution followed a consistent diagnostic sequence — identify the chassis, determine the powertrain configuration, narrow to the specific component application, then cross-reference across sourcing options. What made them fast wasn't memorization of 80,000 part numbers — it was the ability to move through the diagnostic sequence in the right order, asking the right qualifying questions, and knowing which configuration details actually mattered for the specific component being requested. A starter fitment depended on engine and flywheel housing. A turbocharger depended on engine, horsepower rating, and emissions tier. A hydraulic pump depended on machine serial break and aux hydraulic configuration. Each component had a different diagnostic path — and the specialists had internalized all of them.
The system analyzed signals including:
- VIN and serial number decode logic across every chassis and equipment manufacturer the distributor served — resolving not just to the base vehicle but to the specific powertrain, auxiliary, and emissions configuration using serial number range tables, production year breaks, and OE build data
- Complete specification and fitment data across 200+ manufacturer lines — OE, aftermarket, and remanufactured — mapped into a unified cross-reference layer that connected every available option for a given application with pricing, availability, core charges, and warranty terms from Karmak
- The senior specialists' diagnostic sequences documented and encoded as component-specific decision trees — which qualifying questions to ask for which component type, and which configuration details determined fitment versus which were irrelevant for the specific part being requested
- 3 years of return data analyzed for fitment failure patterns — revealing the 14 most common wrong-part scenarios (wrong turbo housing on pre-2017 ISX engines, wrong alternator bracket on Paccar MX vs. MX-13, wrong hydraulic filter on CAT machines across the C6.6/C7.1 serial break) that accounted for 68% of all fitment-related returns
- NLP processing of customer communications — parsing text messages containing serial number photos, emails with equipment lists, and phone call notes where fleet managers described machines using colloquial terms ("the big Cat loader," "the white Pete with the Cummins") that needed to be resolved to specific model and configuration data
68% of fitment-related returns traced to 14 recurring wrong-part scenarios. The senior specialists knew to watch for these. The system flagged them automatically — before the part shipped.
The Challenge
The VP of Sales tracked a single week in September 2024 to quantify the problem. Across all 20 branches, the counter teams received 1,087 fitment inquiries. Of those, 643 were complex — heavy-duty, off-highway, or multi-source cross-reference required. Resolution data showed a clear split:
When a senior specialist handled the inquiry directly, average resolution was 3-4 minutes and the close rate was 91%. When a non-specialist handled it — looking up catalogs, calling manufacturer tech lines, walking to a specialist's desk — average resolution stretched to 14 minutes and the close rate dropped to 66%. The 34% lost-sale rate on extended-resolution inquiries represented the customers who said "let me call you back" and never did, or who called a competitor while on hold.
- The 4 specialists couldn't scale. Each handled their branch's complex inquiries plus phone requests from 2-3 adjacent branches. During peak hours, the queue backed up and non-specialists defaulted to "let me research this and call you back" — a response that lost the sale 78% of the time on urgent jobsite requests
- Heavy-duty fitment was the highest-value and most complex category. A turbocharger for a fleet truck was a $1,200-$3,500 sale. A wrong turbo — wrong turbine housing, wrong compressor wheel, wrong wastegate configuration — was a $300 return cost plus a mechanic who lost a day of labor waiting for the right part. The margin on getting it right was enormous. The cost of getting it wrong was equally large
- Off-highway was the fastest-growing segment and the hardest to serve. Construction and mining equipment parts carried 30-35% gross margins — the highest in the catalog — but the fitment complexity was extreme. Serial number breaks, field modifications, and proprietary manufacturer numbering systems meant that even the senior specialists sometimes had to call the OE dealer to confirm fitment. The distributor was winning off-highway business on expertise but couldn't scale that expertise beyond the branches where the specialists sat
- The VP had tried creating a "fitment guide" — a binder of cross-reference charts organized by equipment type. It was 400 pages within 3 months, already outdated on 15% of its content, and nobody used it because flipping through a binder while a customer waited on hold was slower than walking to the specialist's desk
The Solution
We spent 5 weeks in discovery across 4 branches — the Memphis headquarters, the Nashville branch, the Little Rock branch, and the Jackson, Mississippi branch — observing how counter staff handled complex fitment inquiries and documenting the decision logic the senior specialists applied.
The critical finding: the specialists' fitment resolution followed a consistent diagnostic sequence — identify the chassis, determine the powertrain configuration, narrow to the specific component application, then cross-reference across sourcing options. What made them fast wasn't memorization of 80,000 part numbers — it was the ability to move through the diagnostic sequence in the right order, asking the right qualifying questions, and knowing which configuration details actually mattered for the specific component being requested. A starter fitment depended on engine and flywheel housing. A turbocharger depended on engine, horsepower rating, and emissions tier. A hydraulic pump depended on machine serial break and aux hydraulic configuration. Each component had a different diagnostic path — and the specialists had internalized all of them.
The system analyzed signals including:
- VIN and serial number decode logic across every chassis and equipment manufacturer the distributor served — resolving not just to the base vehicle but to the specific powertrain, auxiliary, and emissions configuration using serial number range tables, production year breaks, and OE build data
- Complete specification and fitment data across 200+ manufacturer lines — OE, aftermarket, and remanufactured — mapped into a unified cross-reference layer that connected every available option for a given application with pricing, availability, core charges, and warranty terms from Karmak
- The senior specialists' diagnostic sequences documented and encoded as component-specific decision trees — which qualifying questions to ask for which component type, and which configuration details determined fitment versus which were irrelevant for the specific part being requested
- 3 years of return data analyzed for fitment failure patterns — revealing the 14 most common wrong-part scenarios (wrong turbo housing on pre-2017 ISX engines, wrong alternator bracket on Paccar MX vs. MX-13, wrong hydraulic filter on CAT machines across the C6.6/C7.1 serial break) that accounted for 68% of all fitment-related returns
- NLP processing of customer communications — parsing text messages containing serial number photos, emails with equipment lists, and phone call notes where fleet managers described machines using colloquial terms ("the big Cat loader," "the white Pete with the Cummins") that needed to be resolved to specific model and configuration data
68% of fitment-related returns traced to 14 recurring wrong-part scenarios. The senior specialists knew to watch for these. The system flagged them automatically — before the part shipped.
Implementation
Deployment occurred over a 01 – 05 period.
Multi-Source Fitment Resolution Engine
ML-driven identification of chassis, powertrain, and auxiliary configuration from VIN, serial number, or partial customer description — resolving to the specific application across OE, aftermarket, and reman options simultaneously.
Component-Specific Diagnostic Workflows
The senior specialists' decision logic encoded as guided sequences — prompting the counter person with the right qualifying questions for the specific component type being requested.
Cross-Reference Intelligence
Unified cross-reference across 200+ manufacturer lines showing every available option for the resolved application — ranked by stock availability, margin, lead time, and warranty terms.
Fitment Failure Prevention
ML-trained flags on the 14 most common wrong-part scenarios, alerting counter staff before the order is confirmed when the selected part matches a known failure pattern.
Karmak Fusion Integration
Fitment intelligence surfaced within the existing counter and order entry workflow — the counter person sees resolved options on the same screen where they process the order.
Multi-Source Fitment Resolution Engine
ML-driven identification of chassis, powertrain, and auxiliary configuration from VIN, serial number, or partial customer description — resolving to the specific application across OE, aftermarket, and reman options simultaneously.
Component-Specific Diagnostic Workflows
The senior specialists' decision logic encoded as guided sequences — prompting the counter person with the right qualifying questions for the specific component type being requested.
Cross-Reference Intelligence
Unified cross-reference across 200+ manufacturer lines showing every available option for the resolved application — ranked by stock availability, margin, lead time, and warranty terms.
Fitment Failure Prevention
ML-trained flags on the 14 most common wrong-part scenarios, alerting counter staff before the order is confirmed when the selected part matches a known failure pattern.
Karmak Fusion Integration
Fitment intelligence surfaced within the existing counter and order entry workflow — the counter person sees resolved options on the same screen where they process the order.
Strategic Impact
Revenue Retention at the Counter
The lost-sale rate on complex fitment inquiries dropped from 34% to 8% within 22 weeks. The $2.8M in retained revenue came from the fleet managers and shop owners who previously hung up after 10 minutes on hold and called the next distributor. The VP of Sales: "We stopped losing $1,200 turbocharger sales because the counter couldn't identify the right housing configuration fast enough. That was always our sale to lose — and we were losing it on speed, not price."
Return Rate Reduction
Wrong-part returns dropped from 7.2% to 2.1% — driven primarily by the fitment failure flags that caught the 14 most common error patterns before the part shipped. At an average return processing cost of $160, the return reduction alone saved $380K annually. The operations team noted the downstream impact: fewer return shipments meant fewer receiving dock disruptions, fewer restocking errors, and fewer credit processing delays.
Expertise Scaled Across 20 Branches
The 28 non-specialist counter and inside sales staff went from resolving 52% of complex fitment inquiries independently to resolving 89% without specialist intervention. The 4 senior specialists' daily interruption volume dropped from 25-35 inquiries to 4-6 — the genuinely novel applications where machine-level expertise was irreplaceable. The Jackson branch — which had no senior specialist on site and had historically routed every complex inquiry to Memphis by phone — saw its complex inquiry close rate rise from 54% to 87%.
Key Takeaway
In automotive and heavy equipment parts distribution, the counter is where revenue is won or lost — and the counter's effectiveness on complex fitment is determined by whether the person answering the phone can resolve a 2019 Peterbilt 579 starter to the right Paccar MX-13 application in the time it takes the fleet manager to consider calling someone else. This distributor's expertise was real — 4 people who could identify the right part for any truck or machine across 40 years of production, 200+ manufacturers, and 80,000 SKUs. The problem was that 4 people can't serve 20 branches simultaneously. The system didn't replace their expertise. It encoded the diagnostic logic and cross-reference knowledge they applied hundreds of times a day and made it available to every counter person in every branch — turning a 14-minute research exercise into a 62-second resolution.
