Skip to content
All journal entries

Field note · Sip Champagnes

The numbers, before the story.

By Daniel Blatchford4 min read

Most consultancy case studies start with a story and end with a vague claim. This one starts with the numbers, because the numbers are unusual enough to lead with.

Sip Champagnes is a UK wholesale distributor. It is also where ODEG’s working method was tested — by being applied, not described. We redesigned how the business operated and built the AI workflows, automations and internal tools to support the new shape. The figures below are taken from Sip’s own audited management accounts, comparing the nine months after switch-on to the nine months before.

+58%
Gross profit

On a slightly smaller top line. Margin doing the work, not volume.

+15pts
Gross margin

Year on year, against the same period a year earlier.

−25%
Cost of sales

Cleaner purchasing, fewer order-entry errors, tighter supplier terms.

−96%
Operating loss

From a sizeable loss to near break-even.

−24%
Admin expenses

Same team, more throughput. AI took the manual work, not the people.

−92%
Loss before tax

The whole P&L moved. Not a one-off, not a single line item.

What changed first.

Before any AI was switched on, the operating model was redesigned. Where work began, where it ended, who owned each handover, how exceptions were treated. Without that, AI just speeds up the wrong process.

The biggest unlocks were the most ordinary. Order processing moved from inboxes and re-typing to a single review surface. Stock and forecasting moved from a Friday spreadsheet to a continuously-updated picture. Invoice chasing moved from awkward phone calls to polite, automatic, on-time reminders. None of those are glamorous — they are simply where the friction sat.

We didn’t fire anyone. We just stopped asking them to do work a machine could do. They started closing deals and managing customers instead of fighting spreadsheets.

Daniel Blatchford

What held into the next quarter.

It is easy to move a P&L for a single quarter. The harder test is whether the trajectory holds. Through Q1 2026 it did:

  • Gross margin sustained above 35%, nine months after switch-on.
  • Operating run-rate at break-even, on track for post-tax profit this year.
  • Order-processing automation running above 95% straight-through on priced lines.

The point of this journal entry is not to claim the same numbers are possible everywhere. They will not be. The point is to be specific about what is possible when strategy and implementation sit in the same room — and to be specific in audited numbers, not vague language.

TransformationAI implementationWholesale
Reply to this