How Velven & Glow built its online home and won back its team's week

The problem

Velven & Glow was growing faster than the systems holding it up. The products were selling in two countries, but the brand had no website that carried its own identity, no clear read on which customers were actually generating the revenue, and a back office where orders, stock and customer email were all handled by hand, one message and one spreadsheet at a time.

Every hour spent reconciling data or answering the same email for the fifth time was an hour not spent on the brand itself. Growth was making that worse, not better.

The approach

I worked on four fronts. First, give the brand a digital home built from its own visual language instead of a template. Second, find out where the highest value customers actually were, so marketing spend could follow the money. Third, take the repetitive operational work off the team entirely by connecting their database to Shopify and letting it run itself. Fourth, hand customer email over to AI agents that could read, classify and answer it.

A website built from the brand, not a template

I designed and built the Velven & Glow website starting from the brand's own visual language rather than fitting it into something generic. The colour palette comes straight from the brand itself, the deep burgundy and soft pink that already lived on the packaging, and it carries through every page without a single off-brand shade.

The structure follows how customers actually shop rather than how a template wants them to, and the tone of every section matches the essence the founders had already defined for the brand. The result is a site that reads like Velven & Glow, not like a storefront that happens to sell it.

• ✓Colour, taken directly from the brand palette and applied consistently across the whole site.
• ✓Essence, with copy and imagery that match the voice the founders had already built.
• ✓Structure, laid out around the real purchase journey rather than a generic shop layout.

velvenandglow.com

Visit the live site →

Click the preview to open velvenandglow.com

A geolocation study of customers by revenue

A customer list tells you who bought. It does not tell you where the money is. I mapped every customer against the revenue they generated and plotted the result geographically across both markets.

That turned a flat spreadsheet into a map: which cities and regions were producing the highest value orders, which were producing volume without margin, and where the customers with real purchasing power were concentrated. With that in hand the team stopped spreading budget evenly across two countries and started aiming campaigns at the places that were actually paying for them.

03 Shopify

Database and Shopify, automated end to end

Orders, customers and stock lived in a database that somebody had to reconcile against the Shopify store by hand. It worked when the volume was small. It stopped working the moment it was not.

I automated the entire pipeline with AI. Data now moves between Shopify and the database on its own, records are matched and updated without anyone opening a spreadsheet, and the mismatches that used to surface days later now surface immediately. The team did not get a faster process, they got the process back.

Customer email, handled by AI agents

Email was the other quiet time sink, and the one that scaled worst with growth. I built AI agents that read every incoming message, classify it by what the customer actually wants, route it, and draft or send the answer on their own.

• ✓Customer follow up, with order status and post purchase sequences that no longer depend on anyone remembering.
• ✓Offers and promotions, matched to the right customer segment automatically.
• ✓Returns and refunds, identified on arrival and pushed straight into the right process.
• ✓General enquiries, filtered so the ones that genuinely need a human get to one quickly.