Consumer / B2C
Anonymised credit/debit transaction panels are the gold standard. Aggregated card-tx data correlates with quarterly revenue at R² > 0.9 for mid-market retail. Augment with foot-traffic geospatial and app DAU.
AltVal re-marks private-market portfolios daily using alt-data — transaction panels, web traffic, headcount, geospatial. We anchor at the GP's lagged quarterly mark and project forward via sector-specific factor models, with confidence bands and driver decomposition. Best backtest cohort R² 0.76, ~50 days of lead over the official report cycle.
Private-market portfolios are still revalued quarterly, on a 45–60 day appraisal cycle, with smoothing built in. The economic state of each company has already moved — usually visibly, in card-transaction panels, web traffic, headcount, port and facility activity. AltVal closes that gap.
Every portfolio company gets a fresh mark each day, anchored on the GP’s last reported quarter and projected forward via sector-specific factor models. Confidence bands and driver decomposition included.
Quantifies how much reported NAV is artificially smoothed by appraisal-based marking. Outputs an info-asymmetry alpha (bps captured by leading the official mark) and a volatility-adjustment vs the reported series.
Different sectors leak information through different signals. AltVal uses sector-specific factor models trained on the most predictive alt-data feeds for each.
Anonymised credit/debit transaction panels are the gold standard. Aggregated card-tx data correlates with quarterly revenue at R² > 0.9 for mid-market retail. Augment with foot-traffic geospatial and app DAU.
Web traffic + technographics + LinkedIn talent flows. A 10% engineering-headcount drop is a leading indicator of operational distress months before it shows on the P&L. Tech-stack churn flags impending customer loss.
Geospatial intelligence — satellite imagery of parking lots, oil-tank floating lids, port and rail activity — gives physical “ground truth” that precedes any financial reporting. Best signal-to-noise of the three sectors in our backtests.
Out-of-sample walk-forward backtest vs simulated ground-truth revenue across 25 portfolio companies in 3 simulated funds. 30-day step, 18 periods. Industrial / geospatial is the cleanest signal cohort; consumer-tx tracks well; SaaS is noisier and cohort-dependent.
Disclaimer: Backtest figures are computed against simulated ground-truth revenue inside the demo’s synthetic universe; they illustrate engine behaviour, not real-world accuracy on third-party portfolios. Past performance, simulated or live, is not indicative of future results. Nothing on this page is investment advice or an offer to sell securities.
Our second product. A coordinated fleet of agents trading equity options with an emphasis on volatility surface, dispersion, and event-driven setups. Currently in walk-forward paper trading; live capital gated on live-paper convergence.
The AltVal terminal demo is live now — launch it. Drop your email below for engine release notes, R&D updates, and a heads-up when AltVal opens to live private-market portfolios. No marketing emails.