GoodDom Case Study: Why a 3-Million-Ruble Accounting System Lost to a Claude SetupNote · digest for November 7
Entrepreneur's desk displaying custom warehouse system code in Cursor alongside a furniture hardware mockup.
Note2 sections≈ 3 min

GoodDom Case Study: Why a 3-Million-Ruble Accounting System Lost to a Claude Setup

Once your marketplace store passes 10–20 million rubles in monthly turnover, you inevitably hit a glass ceiling in accounting. Standard Excel spreadsheets start bursting at the seams, FBS stock stops matching the warehouse, returns vanish inside sorting centers, and managers drown in endless manual reconciliations.

Every sensible entrepreneur's first thought is: "Time to grow up. Enough makeshift tools — let's buy a professional WMS (Warehouse Management System) from a reputable integrator."

That is exactly what the leadership of furniture brand GoodDom, members of Dmitry Kovpak's business community, did. The company paid 3,000,000 rubles for a turnkey "off-the-shelf" warehouse accounting system. The expectation was simple: get a polished powerhouse connecting bin-location storage, barcode picking, and their Wildberries and Ozon dashboards.

Reality hit like a cold shower.

The Heavy-Software Trap: Why Multimillion-Ruble WMS Platforms Slow Sellers Down

The core flaw of classic WMS platforms is that they were architected for sluggish wholesale distributors circa 2010. In modern e-commerce, marketplace rules shift every two weeks: logistics coefficients update, new label formats roll out, and КГТ receiving regulations change.

When the GoodDom team needed to configure custom logic for checking hardware-kit completeness and tie it into a daily operational metrics dashboard (РНП — "finger on the pulse"), the integrator threw up its hands:

  • "That's a non-standard workflow; we need to draft a spec for custom development."
  • Every tweak was quoted at an extra 150,000–500,000 rubles.
  • Worst of all, the release timeline for any change was 3 to 6 months!

While the integrator wrote reports, warehouse operations stalled, and customizations lagged a full season behind the business's actual needs.

The Custom-Code Revolution: How a Practitioner Builds a Tailored System in One Evening

Realizing the project had turned into a bottomless money pit, GoodDom's owner made a radical move: canceled the contractor agreement and decided to build the software in-house.

Sounds crazy? A factory owner with no computer science degree. Yet he holds the one thing no hired developer has: an intimate grasp of his warehouse floor, pickers' routes, and unit-economics bottlenecks.

The owner sat down with the Claude language model inside the Cursor IDE and wrote a custom inventory tool in a single evening! The system did precisely what the furniture factory needed: tracked every part's path, verified spare-hardware stock, and streamed key metrics to a smartphone screen in real time.

And this case isn't an outlier. Among sellers in Kovpak's Accelerator, it has become a steady trend:

  • entrepreneur Maxim used Cursor and an LLM to build a unified, secure manager workspace on his own server, controlling prices and stock across WB and Ozon without handing master passwords to third-party services;
  • Miyumi founder Ruslan deployed an architecture of 35 specialized AI agents monitoring logistics, ads, and unit economics 24/7;
  • other sellers build end-to-end influencer-marketing analytics in 1–2 days, spotting bloggers with inflated reach by tracking actual gross SKU buyouts.
Limits of Custom-Built Systems

Don't get carried away thinking AI completely replaces traditional engineering. Homemade microservices carry two vulnerabilities: once transactions scale to hundreds of thousands of rows a day, the database needs proper index design to avoid locks. There is also the "bus factor": if only one person holds the tool's logic in their head, the business must enforce architecture documentation.

In our piece on optimizing business automation costs, we noted that the winners aren't those who buy the priciest software on the market, but those who adapt digital tools to their sales funnel fastest.

Want a deep diveVotes help us choose what to break down in detail. This material has 0 votes.
Sources · 2