Marketplace IT Landscape: When Does Custom Code Pay Off — or Just Burn You?
9 June 2025
In just 10 months, an experienced team launched the fastest-growing marketplace in Kazakhstan – but not without hitting serious forks in the road. This isn’t just another startup story. It’s a real-world breakdown of when building your own software stack gives you an edge – and when it nearly wrecks your timeline, budget, or team.
Oleg Aksenov, CIO at Teez, presents his view on decisions such as:
- Why they built their own Warehouse Management System to cut 20 clicks per order down to 5
- How they chose technical stack for performance and delayed microservices until absolutely necessary
- How they approached hiring process with limited resources
- How custom logistics from China to doorstep in CIS region became their killer feature
- How bureaucracy and over-planning can nearly freeze feature delivery for a month
If you’re facing the build vs. buy decision in logistics, operations, or platform design, this story shows exactly where custom code can win – and where it costs more than it’s worth.
The backbone of the product – its’ team
The choice of the right founders can lead to vast improvements in productivity and, therefore, results. There are combinations of archetypes that are 12 times more effective than the others.
Oleg and his team decided upon the strategy to build a project in such a short time. The most important task was to identify what was going to be the main feature, the competitive advantage, and other, less important parts of the system could be outsourced or purchased as a white-label.
A rather innovative approach was taken for hiring of new personnel: not to try to get the best people from the market but to build a balanced structure that has all the cogs needed to run the business without overpaying to several rockstar-developers. Every hiring was digitized into one of nine sectors.
As the company scaled, their release cycle slowed dramatically. Initially, they launched core systems in just 3 months with weekly releases. But after splitting into multiple teams and adding rigid rules, deployments dropped to one feature release per month. Bureaucracy and over-planning clogged the pipeline. The turning point came when they separated core systems (where precision matters) from customer-facing features (where experimentation is fine), halving delays and restoring a faster, leaner release process.