Building economic superintelligence by pooling demand data across brands — the way LLMs pooled text.
cross-brand training
Every brand is an island. No one has enough data to see the whole picture.
Each brand sits on years of demand data trapped in its own silo. No single brand has enough data to understand what actually drives demand — weather, promotions, competition, or baseline shifts.
But pool the data across brands — the same way LLMs pooled text — and the model gets dramatically smarter for everyone. We’re building the shared corpus that makes this possible.
For the first time, the data can be pooled — and the science works.
Scaling laws were proven for economic time series this year (NeurIPS 2024). Commerce APIs opened brand data. Google TimesFM, Amazon Chronos, and Salesforce Moirai proved multi-source training works for time series. The window to build the corpus is open.
Pooling works. Cross-brand training produces dramatically better predictions.
A model trained on pooled data (retail, macro, weather, competitive signals) outperforms a single-brand baseline by +16% on held-out demand prediction. Tested on M5 — the world’s largest retail forecasting benchmark: 100,000 Walmart SKU-level time series. The pooled model carries knowledge from thousands of sources to every brand on day one.
For a $10M brand, this is the difference between overstocking by 12% vs. 3%.
Every brand that joins makes every other brand smarter.
What the network unlocks.
SKU-level demand with uncertainty, refreshed weekly
What’s driving demand: weather, promos, baseline, competition
When to act and when to wait
No single brand can generate these outputs alone. The cross-brand corpus decomposes each SKU’s demand into signals only visible when you pool enough data.
This isn’t analytics software. It’s a shared data network.
Retail demand planning software (Gartner, 2025). Beachhead: 80,000 Shopify brands underserved by enterprise tools — high data-pooling upside, fast sales cycles.
SaaS subscription scaling with SKU count. Target $100K ACV.
Co-founded Homesick Candles (8-figure exit to WIN Brands). Founded NorSari (44% CAGR, 8 years). Built WeatherVane’s causal ML architecture and benchmark proof solo. 10+ years in DTC demand.