Worked example

Roost & Co. — a Cape Town coffee roaster

Three systems, one dashboard, four insights that changed how they made decisions. Here's the full breakdown.

Background

Roost & Co. is a specialty coffee roaster based in Woodstock, Cape Town. Founded in 2019, they roast small batches for a growing list of wholesale cafés across the Western Cape, sell single-origin beans direct to consumers through their online store, and run a flagship roastery café where most weekend revenue comes from card payments.

Annual revenue sits around R14 million, split roughly across wholesale, direct online sales, and the roastery café. The team is small: a head roaster, two baristas, a part-time bookkeeper, and the founder who handles everything else — including pulling numbers together at month-end.

Their stack

Three systems, none of them talking

Xero

Accounting, invoicing, supplier costs

Shopify

Online store and subscriptions

Yoco

In-store card payments at the café

The problem

What the founder couldn't see

No way to view total revenue across all three channels in one place — month-end took a full Saturday of spreadsheet work.
No visibility into the café's trading pattern — whether weekends were driving the majority of revenue, and what that meant for staffing and stock ordering.
Product margin was invisible — some top-selling beans had very different cost structures that eroded their contribution, but this was impossible to see across separate systems.
Online repeat-purchase behaviour lived in Shopify; the founder had no view of how it compared to the café's repeat customers.
Inventory decisions on raw beans were made by gut feel — there was no link between roasting volumes, supplier costs, and downstream sales velocity.

What we built

One Power BI dashboard, six pages, refreshed daily

We connected Xero, Shopify, and Yoco into a private Azure environment, modelled the data so wholesale, online, and café revenue could be viewed together, and built a Power BI dashboard with pages for sales overview, customer analysis, product performance, debtors, channel comparison, and data freshness. The whole build took three weeks from kickoff. Roost now logs in each morning to one place — no more month-end spreadsheet Saturday.

What the data revealed

Four insights from the first month

Revenue concentrated in a handful of customers

A handful of wholesale cafés were quietly carrying the business — invisible across three separate systems.

In-store revenue is heavily concentrated on weekends

The roastery café's takings skew strongly toward Fridays and weekends — directly shaping how Roost plans staffing, stock orders, and weekly cash flow.

Revenue rank and margin rank don't agree

The gap between the highest and lowest margin products is significant. Some top sellers by volume are far less profitable than they appear — only visible once costs and revenue are modelled together.

A strong online repeat-purchase rate — higher than expected

Higher than industry average — gave them confidence to launch a subscription product.

Live demo dashboard

Browse the report yourself

The actual Power BI report we built for Roost & Co. — fully interactive. Use the page tabs at the bottom to explore each view.

Want to explore it without the page constraints?

Open the report full screen in a new tab.

Open full screen →

Roost & Co. is a worked example we built ourselves to show exactly how we approach a build. The systems, problems and insights reflect the situations we see in real South African small businesses.