What it takes to move a platform forward inside a Fortune 50
A blueprint for what building actually looks like inside a large, legacy-heavy, sales-driven organization. Two years as Platform Lead on a Fortune 50 e-commerce platform. 315,000+ employees. $92B revenue. Elixir backend (yes, really).
If you're building something smaller and simpler - good. Learn from the pain so you don't recreate it.
What you walk into
A platform that technically works but isn't really finished. Migrated from Node.js to Elixir, running in production, but raw - rough UX, missing features, reliability gaps.
One more layer of complexity: the platform team was an acquired startup, still semi-independent inside the larger group. Which meant building with enterprise constraints but without enterprise resources. The worst of both worlds, if you're not careful.
20+ engineers across multiple squads with no clear boundaries. People float to whoever has the most organizational pull that week. Priorities shift based on politics, not strategy.
On the B2B side: hundreds of customers placing orders through emailed spreadsheets. Stale product catalogs in Excel. Manual entry, manual stock checks, errors everywhere. A $92B company running a core revenue channel like it's 2004.
What actually has to happen
The first job isn't technical. It's organizational.
Multiple divisions pulling the platform in different directions. Competing priorities, unclear ownership, layers of approval that exist to protect process, not progress. You learn quickly that saying no to powerful stakeholders is half the job.
You run workshops - real ones, not post-it-note theater. Internally, to force alignment on what actually matters. Externally, with customers, to understand how they buy, what breaks, what they'd use if it existed. Real workshops make emotions surface. People argue, defend, let go. It's not just another meeting you hope ends sooner so you can go back to pushing spreadsheets and emails.
You also end up picking up the projects nobody wants to touch. Half-finished, stuck in limbo, politically toxic. Someone has to close them out. That someone is usually whoever cares more about the outcome than the credit.
Then you build. In this case:
- A fully self-service B2B e-commerce solution replacing the entire spreadsheet-and-email workflow
- An invoicing system from scratch - the old process was manual AR teams fighting SAP and bureaucratic tooling that actively resisted change
- Improved checkout, fraud prevention and detection, security hardening
- A better admin portal for internal teams who'd been suffering in silence
What comes out the other side
- $5M in new revenue unlocked within 2 years, $30M+ projected over the next 5
- Entire B2B customer base moved to self-service ordering
- Engineering team went from floating chaos to clear ownership
- A few bureaucratic teams and ambitious middle-managers thoroughly annoyed
What you learn
Organizations at this scale can't move fast. That's not a failure - it's physics. Too many stakeholders, too much surface area, too much history baked into every process.
The job isn't to fight that. It's to be patient, keep pushing, and protect the work from drowning in meetings, approvals, and consensus theater.
You also learn where your own limits are. I got too invested in battles that weren't mine. Same outcome, less stress - that's the version I'd run next time.
This project was delivered through Toptal.