Impact study // Barnes Crossing Auto Group // 16 rooftops

The group that stopped renting its own reality.

How Barnes Crossing Auto Group - 16 rooftops across the Southeast - rebuilt its data layer from the inside - one owned warehouse, fifteen production systems on top of it, and a reporting stack that runs itself every morning.

Barnes Crossing is the group Vespoke's founder runs as Director of Operations - the proving ground for everything on this site. Want store-level detail? Ask for the live walkthrough.

IMPACT SO FAR

  • 26,135 vehicles delivered in the trailing 12 months - every deal, RO, and lead lands in a warehouse the group owns.
  • Fifteen production systems run on that one foundation - reporting, audiences, opportunity mining, recalls, widgets, automations.
  • The group sales number lands in every GM's inbox at 6:00am, 365 days a year - nobody keys anything.
  • 100 DMS-verified, consent-scrubbed leads handed to a store's AI outreach vendor in one morning - the first activation batch of many.
  • Tens of thousands of competitor listings crawled nightly into the market-intel base.
  • Rented website widgets replaced with owned ones on the first stores - leads arrive structured, and the data stays home.

01 / CHALLENGE

Sixteen rooftops, and no single version of the truth.

Barnes Crossing had what every dealer group has: a DMS full of transactions, a CRM full of leads, a website vendor holding the analytics, and a shelf of tools each reporting on its own performance. The numbers didn't agree, arrived late, and lived in systems the group didn't control.

Corporate learned group sales at month-end. A simple question - how many 84-month deals have we written this year? - took days, because it routed through whoever knew the report writer. Vendor reports graded the vendors' own homework. And each new tool meant another integration fee to access the group's own data.

< 1 in 3

dealers are satisfied with the customer data they get - 83% receive it regularly, 7 in 10 say it lags too much to be useful, and half see sources that conflict. (Cox Automotive, 2026)

~$42,000

a year is what an average dealer pays in third-party integration fees - roughly $3,500 a month, per store, to move its own data between vendors. (Dealertrack integration-fee analysis)

02 / SOLUTION

Build the warehouse first. Then let it compound.

Instead of buying another tool, the group built the layer underneath the tools: one warehouse, in its own name, fed by every source it already paid for.

The foundation went in first. DMS transactions through the DMS's own API. A full replica of the CRM across 14 rooftops. First-party website activity. Inventory feeds. Public APIs - federal recall data, book values - that vendors resell at a markup. One bridged customer identity across all of it.

DMSdeals · service · GL CRMleads · notes · history SERVICE CRMappointments · check-ins WEBSITEclicks · forms · first-party MANUFACTUREROEM portals · allocation MARKET + APIScomps · recalls · book values /00 THE WAREHOUSE one customer. one record. yours. TEXTS + EMAILoutreach, receipted DASHBOARDSone live picture ASKplain-English answers AUTOMATIONSwork that runs itself AI AGENTSthe data, safely AI PLATFORMfed real context

Everything the group already pays for feeds the warehouse - and the warehouse feeds everything back out, in the group's own definitions.

Then the systems, in order. Intelligence first: self-running morning reports, a live P&L, a plain-English answer engine, a CRM funnel cross-checked against deliveries. Then activation: audiences in plain English, an opportunity engine computing payment-relief and equity plays per customer, recall outreach off the free federal feed, one governed comms inbox. Then the customer-facing layer: owned trade-in widgets, campaign pages, service-lane tools. Then the autonomous layer - the pieces that run every night without a human.

Built from the inside. No 18-month data project, no consultants' roadmap. Each system shipped in weeks, survived contact with real staff and real month-end, and earned the next one. The full map - all fifteen, with what each replaced - is on the Work page.

THE SHIPPING LOG - EIGHT LAUNCHES IN SIXTY DAYS, JUN–JUL 2026

JUN 1 JUL 1 JUL 31 Comms queueJUN 8 Campaign pagesJUN 26 Competitor intelJUL 1 Recall SMSJUL 4 Opportunity engineJUL 12 Conversations inboxJUL 21 Disposition trackingJUL 25 100-lead vendor batchJUL 28

Real launch dates from the build log - this is what "ship every month" actually looks like. Vendors quote quarters; the warehouse makes it weeks.

03 / IMPACT

Same stores. Different operating picture.

VEHICLES DELIVERED BY MONTH - TRAILING 12, FROM THE WAREHOUSE ITSELF

0 1,000 2,000 1,567Aug 2,157Sep 2,151Oct 1,823Nov 2,312Dec 2,104Jan 2,280Feb 2,568Mar 2,404Apr 2,337May 2,377Jun 1,923*Jul*

Retail + lease deliveries by month, queried straight from the group's warehouse. *July 2026 is month-to-date. Hover any bar for its value.

View as table
MonthUnits
Aug 20251,567
Sep 20252,157
Oct 20252,151
Nov 20251,823
Dec 20252,312
Jan 20262,104
Feb 20262,280
Mar 20262,568
Apr 20262,404
May 20262,337
Jun 20262,377
Jul 2026 (MTD)1,923
26,135

vehicles delivered in the trailing 12 months, flowing through the owned warehouse with full deal, service, and lead history.

15

production systems live on one foundation - each one either replacing a rented tool or doing work no vendor sells.

06:00

the group number in every GM's inbox, every morning, 365 days a year - with tripwires that catch bad data before a human repeats it.

days

from "we should act on our equity" to verified lists in an AI vendor's hands - because the data was already home.

1 inbox

for customer conversations across rooftops, with frequency caps enforced by the system instead of memory.

$0

in new licenses for recall outreach - the federal recall feed is free; matching it to owners is just warehouse work.

The quieter change is cultural: arguments about whose number is right ended, because there's one set of definitions and everyone reports from it. When the principal asks a question in a meeting now, the answer arrives in the meeting.

Your group has the same raw material.

A 45-minute working session: we'll map your sources, your vendor stack, and what owning the layer underneath would change. No deck.

Book a working session