Inspiration
When e-commerce operators sign up for multiple promotions, they can't judge a campaign by its projected sales alone. Shared inventory, ad budget and schedule conflicts change what revenue is actually achievable. A promo that looks great on a spreadsheet can oversell stock, blow through budget, or collide with another event — and the damage usually only shows up after the campaign is live.
What it does
PromoCheck compares two candidate promotions under one consistent set of rules — shared inventory, budget, scheduling and mutual-exclusion constraints — and tells you which plan is actually worth running and which one carries execution risk.
It does not just pick the "highest sales" promo. If a plan would drain inventory, exceed budget, or conflict with another event, PromoCheck flags it as infeasible or higher-risk. Only plans where both the upside and the resource constraints actually hold are worth prioritizing.
Key features:
- Campaign configuration risk checks
- Stock-gap view and a running inventory ledger
- Budget constraint tracking
- Schedule and promo-conflict warnings
- Side-by-side projected profit contribution chart
- Human-reviewable recommendations, with the evidence behind each call
- JSON save / restore and evidence export
How we built it
A local React + Vite web prototype, scoped to a single SKU and two concurrent promotions. Rules and user-entered parameters drive the comparison, and the risks and trade-offs are visualized with charts, a campaign timeline and an inventory ledger.
Challenges we ran into
The hard part isn't calling the highest projected revenue the "winner" — it's weighing shared inventory, budget and timing at the same time, and making the reason an infeasible plan fails intuitive enough for an operator to judge without a data scientist in the loop. We also had to resist the urge to overclaim: every profit figure on screen is an editable assumption, not a backtested result.
Accomplishments that we're proud of
- Turned a vague "promo scheduling goes wrong" pain point into a working, interactive prototype
- Made four usually-silent constraints — inventory, budget, timing and mutual exclusion — visible at the same time, instead of hiding them behind a single recommended number
- Built every recommendation to come with its evidence, and made every assumption editable, so an operator can challenge the tool rather than blindly trust it
- The whole app runs locally with no backend, and plans can be saved and restored as JSON for later review
What we learned
Business decisions have to evaluate upside and executability together. Showing the assumptions, constraints and risks explicitly builds more trust than a single black-box recommendation. "Highest projected profit" is only a meaningful answer once you've checked that the plan can actually be executed.
What's next for PromoCheck
- Connect to real e-commerce marketplace order/inventory APIs — right now the demo uses synthetic sample data and is not wired to any seller backend, and the "higher ROI" shown is a projection from the assumptions you enter, not a measured lift
- Extend from single SKU / two promos to multi-SKU bundles and more concurrent campaigns
- A/B the tool's recommendations against post-campaign results, and turn the projected ROI into a measured one
- Add multi-user review so a team can sign off on a plan before it goes live
Built With
- javascript?react?vite?css