Skip to content

Magento shop too slow? Hyvä makes it fast. More on Hyvä development

Abstract sculpture: two separate bodies with a wide dark gap between them, bridged by a single slender connection.

Project · Vitafy

Eleven Years on a Platform That Never Stood Still

Vitafy sells across several brands and markets. The shop ran on Magento for eleven years, and what ran around it changed more than once. I was the developer who kept the shop connected to the rest.

Where it started

In 2013 Vitafy was a shop with a growing range. The shop ran on Magento. What ran around it was not one system but several: order handling, shipping and stock sat in different places, and the data between them did not move continuously but in batches at fixed times.

Volume was never the problem. The shop was large from the start. What pushed that arrangement to its limit was the number of brands, countries and currencies that arrived.

What had to be connected

Shop and inventory

Orders leave the shop, stock levels and shipping status come back. That already ran automatically, but across hand-rolled one-off pieces. The gain was not automation, it was one connection instead of many.

Shop and CRM

Customer, order and campaign data had to reach the CRM without the CRM becoming a second source for order data. Every field has one side that owns it.

Several brands and markets

B2C and B2B sub-shops by market, country and currency on the same platform. One integration layer for all of them, rather than one connection per brand.

One platform over eleven years

A shop that grows for eleven years collects modules, special cases and legacy. Alongside the integrations came work on TTFB and platform performance.

The connector nobody was maintaining

In 2022 the warehousing moved onto a standard inventory platform. The shop was still running on Magento 1 at that point. The connector meant to join the two was no longer being developed by the company that made it.

There were two ways forward. Migrate the shop first and make the warehouse project wait, or take the connector over. We took it over. I rebuilt it around the business rules of the house: exports carry what those rules ask for, and the payloads of the API calls match what the shop actually sends rather than what the documentation said it sent.

Around that sat the work that eats most of the time on a project like this: data hand-offs between the systems, third-party extensions that had to be extended for the case in hand, and code of our own holding the two together.

The decision that made the difference

The obvious answer would have been to connect each brand separately. That is quick, and it comes back on you at the fourth one: four connections, four places to go wrong, four places where somebody forgets a change.

Instead there is one integration layer that every brand runs through, and exactly one system that owns each field of data. Product data, prices and stock have an owner. A change happens there.

My part in it

From 2013 to 2024 I was part of the internal development team. On the shop side I was the developer for the connections outward: to inventory, into the CRM, across the sub-shops by market and currency, along with work on TTFB and platform performance.

Alongside that came the work with the teams who handle the data every day. For business intelligence and for customer support I built what they needed to do their jobs. Someone with a report or a customer case in front of them notices things about an interface that appear in no specification.

That is the arrangement I still work in: inside an existing team, at the point where the depth is missing, owning a piece that runs without me afterwards.

What changed

Orders find their way without being reassembled on the journey. Stock is right on both sides, rather than reconciled afterwards. Brands and markets arrived without a separate connection being built for each. And moving the warehousing did not have to wait for a shop migration.

On figures. The integration improved the economics markedly against what came before, and revenue grew strongly over the years we worked together. There is deliberately no percentage here. The numbers I hold come from my own records, with no documented baseline and no independent source. What can be evidenced is above.

Is somebody sitting between two systems at your company?

Tell me which two systems they are and what happens by hand between them today. I will tell you whether it can be connected and where I would start.