Caldwell Vineyard

Snapshot

What they do

Caldwell Vineyard crafts a diverse portfolio of estate wines from 28 grape varieties in Napa, California. The business reaches its customers directly through several channels at once: wine clubs led by the Society of Smugglers, the tasting room, JFC on the Road private events, a barrel program, and online sales.

Delivery

Co-delivered by Value-First Team and USS.

Shape of the work

Coaching, not implementation. The Caldwell team built the system themselves, session by session, with strategic guidance alongside.

The relationship

Sep 2025

Two scoping sessions with the full team — owner, wine club, tasting room, marketing, technical — mapping every revenue stream and every place customer data was living.

Oct 2025

Coach-only engagement begins. The team chose to build internal capability rather than hand the work to an outside implementer, and the plan was structured in three milestones that each had to prove their value before the next one started.

Nov – Dec 2025

Order tagging, product taxonomy and channel attribution built in Shopify and carried into HubSpot. In December, the first channel revenue number the team could actually trust.

Jan – Feb 2026

Membership sync rebuilt, the manual club-run process mapped end to end, and the first automations replaced work that had been done by hand from personal inboxes.

Mar – Apr 2026

Attention turned to replacing the wine-club platform underneath everything. The team documented its own non-negotiable requirements and worked through a side-by-side comparison of four alternatives.

Where they started

Napa was in the middle of a genuine downturn. Tasting room visits had fallen by about half across the valley, and the email channel that had carried new-release announcements for years was returning a small fraction of what it once did. The team was adapting — taking the wine to customers through JFC on the Road rather than waiting for customers to come to the valley — but they could not measure whether it was working.

The reason was that nothing agreed with anything else. Customer information sat in the commerce platform, in HubSpot, in the wine-club system, in a long list of spreadsheets, and in the heads of the people who had always handled it. Contact counts differed between systems and nobody could say which was right. The connector between commerce and CRM no longer carried the tags and notes that held the context, so the richest information stayed where it was created and was invisible everywhere else.

The practical cost was that the team could not answer their own most important question — which parts of this business actually generate revenue. JFC on the Road was believed to matter. It could not be proven. And the club run, the core of the business, was executed almost entirely by hand: a member emails asking to swap a wine, and someone works through the club system, the commerce platform and their own inbox to make it happen, carrying the rules in their head because they were written down nowhere.

What we delivered

A full current-state analysis, a future data model and an implementation roadmap, produced from the scoping sessions with the whole team in the room.
A working frame for what the platform is actually for — one view of the customer, one view of revenue, one view of the business, and a foundation the team and AI can both work from — used to decide what to build and what to skip.
Channel revenue attribution: product taxonomy rebuilt by variety, vintage and club, orders tagged at source, and reporting that splits revenue by the channel that produced it.
Transactions moved onto HubSpot Orders as the standard for everything originating in the storefront, with negotiated arrangements kept separate.
The membership sync diagnosed and rebuilt against customer records rather than the club system's own objects, producing membership lists that finally matched one to one.
Automated communications the team had been sending by hand: tasting confirmations to the guest and to whoever referred them, follow-ups for visitors who joined the club and for those who did not, and a welcome email that now fires on membership rather than a stale tag.
Reporting consolidated into one dashboard, so the numbers stopped living in one person's exports.
Contact-data technique taught directly — associating records and confirming the history before merging them, rather than merging and hoping.
Requirements and a four-way comparison for replacing the wine-club platform, written from the team's own stated non-negotiables rather than from vendor feature lists.

What changed

In December the team put a number on JFC on the Road and it landed within about three percent of what the owner had believed it to be. That sounds small. It was the whole point: the largest thing they had been doing on instinct was now measured, and the measurement agreed with the instinct. From that moment the reports were trusted, and a question that had been unanswerable for years — which channels drive this business — had an answer they could act on.

Work that had been manual stopped being manual. Confirmation emails that went out from a personal inbox now go out from the system. Membership lists that had to be pulled and checked by hand are now trusted enough to send club communications against. The spreadsheets that ran the tasting room and the events were no longer the only record.

The change that outlasts the engagement is the capability. The team stopped waiting to be shown. Their operations lead built a business dashboard himself with AI tools after describing the technology as less frightening than he expected; the tasting-room lead moved from watching to running her own workflows; the club manager started asking why a process worked the way it did rather than repeating it. And when the underlying wine-club platform finally had to be confronted, the decision to replace it did not come from us. It came from the people doing the work, unanimously, with their own requirements documented and their own comparison in hand.