6 min read
Predicting the Impossible: How Investors Can Predict Profit Metrics with Alternative Data
Most fund managers who use Alternative Data lean on it for building forecasts of “top line” metrics such as revenue or same store sales. This is intuitive, as many alt data sources provide a sample of revenue and sales (e.g., credit card data). Using Alternative Data to predict more complex profitability metrics has long been a pipe dream for investors. Unlike revenue, profit metrics are influenced by numerous factors, many of which cannot be easily tracked outside of the company. For years, investors have sought ways to use data to predict a company’s profitability. The wait is finally over.
After several years of painstaking R&D, our team has begun rolling out our first wave of profitability models for Maiden Century clients. Coverage is very broad and spans nearly our entire universe of 2,000+ public companies, spread across all major verticals and geographies. We are starting with Gross Profits as our first “bottom line” forecasting metric.
Starting here was intuitive. Using our Model Relevance indicator (which tracks stock price sensitivity to surprises around various reported KPIs), we found Gross Profits to be the leading indicator of share price performance in roughly one-third of the companies examined. Intuitively, this makes sense: revenue growth in the absence of persistent improvements in profitability will fail to translate into enterprise value improvement.
Forecasting profitability metrics required us to take a fundamentally different path to modeling. Gross profits are a mix of both revenue and CoGS, which means accurate modeling requires multiple inputs. Analyzing costs in the absence of a view on revenues is meaningless. Hence, we have built entirely new models for predicting Gross Profits, leaving behind the legacy approach used for revenue forecasting.
We remain big proponents of the ‘model of models’ approach that we have used successfully to build revenue models. Unsurprisingly, MaidenCentury’s Gross Profitability modeling leverages multiple models. Each represents critical metrics for estimating the operating leverage in the business, the means by which changes in revenue can drive moves in gross profit, as well as how trends in inventory turnover and input costs portend changes in margin. Taken together, they provide an effective approach toward modeling gross profitability. Clients will be receiving a deeper primer into how each of these models work to drive our Gross Profitability estimates. Request the primer by reaching out.
Having tested our Gross Profitability estimates, we are witnessing hit rates at ~65% across the 2,000+ companies we model, as of 1Q24 earnings. We expect that hit rate to converge with that of our revenue models (~70%) by year-end.
Our ultimate goal is to be able to predict real-time changes to company valuations based on incoming signals from alternative data. Working backward, this means having a view on how the Street’s long-term FCF projections are likely to change in the coming weeks or months as a result of surprises already predicted by alternative data, and then using DCF modeling to back into expected changes in share prices. In that context, our initial Gross Profit forecasts are a baby step towards the grand vision.
Predicting non-revenue metrics with alternative data has long been an elusive goal for investors, and we think the work we’ve just rolled out is a very promising first step towards that goal. Find out more in a live demo: contact us.