SWStrategies

Home › Case studies

Retail · Marketplace · 2019–2023

After $50M and two failed attempts, live in 10 months for under $1M

A B2B marketplace for a Fortune 5 retailer. Four years and more than $50M had gone into this platform before we arrived. An internal team had a proof of concept that would not scale; an external vendor spent a year and never reached one. We reached production in ten months.

Four years of failed attempts against ten months of delivery Above the line, four years and more than fifty million dollars: an internal team reached a proof of concept that would not scale, and an external vendor spent a year without reaching one at all. Below the line, SWStrategies reached a proof of concept in seven months for under five hundred thousand dollars, production three months after that, then added features for three years before handing the platform to the client's own team. Before us 4 years, $50M+, no working platform Internal IT team reached a POC that would not scale External vendor, one year never reached a POC at all Every attempt started from the same architecture. With us 10 months to production Month 7 — POC under $500K, scalable this time Month 10 — production live to 100,000+ merchants Three more years features, then handover in-house What changed was not the engineering budget. It was the architecture — microservices that scale without a rewrite — and an agile process that let the business see working software early enough to steer it. The platform now runs under the client’s own team, which was always the point.
A B2B marketplace for a Fortune 5 retailer, after two failed attempts. The second row is the same problem, a different plan.

What was wrong

  • The client had attempted multiple builds of a B2B marketplace over four years, spending more than $50M.
  • An internal IT team produced a proof of concept that could not scale.
  • An external vendor was brought in and after a year had still not reached a working proof of concept.

What we did

  • Identified the delivery-process and architecture problems that every previous attempt had inherited, before writing any code.
  • Built a microservice architecture that scales without a rewrite, rather than another monolith that would hit the same ceiling.
  • Designed and ran agile processes that got working software in front of the business early enough for them to steer it.
  • Added machine-learning recommendations against merchant profiles so buyers could find products across a catalog of millions.
  • When Covid shut down the trade shows the marketplace was meant to replace, the project went from strategic to urgent. Because the proof of concept had been built to scale, it became the production MVP rather than a throwaway.

What came out

  • Proof of concept delivered in seven months for under $500K.
  • The proof of concept went to production three months later, with no rewrite — ten months from a standing start.
  • Under $1M in total to production, against more than $50M already spent.
  • Live to 100,000+ merchants and millions of products, with headroom for ten times that.
  • Latin American merchants onboarded first, with the US and India next in line.
  • We kept adding features for three more years, then handed the platform to the client’s own team.

Built with

  • C# / .NET microservices
  • Angular
  • Python
  • Azure

Client names withheld by agreement. References available on request.

Is a platform build on its second or third attempt?

Every attempt after the first inherits the architecture that defeated the last one. An hour with your team, at no charge, and we will tell you what we would keep and what we would not.

Talk through your problem The other case studies