When you open a Stock Characteristic inside Analytical Platform, the result may look simple at first. A few charts. Several statistics. Historical performance. A list of stocks.
But behind every single result, there is a surprisingly large amount of data, processing, computation, and quality control.
Data Most Investors Cannot Easily Access
Every Stock Characteristic starts with high-quality historical data.
And that does not mean only stock prices.
We also work with financial statements, fundamental indicators, historical index constituents, and many other datasets that are either unavailable to most individual investors or extremely expensive to obtain and maintain.
Buying Data Is Not Enough
Every data source has its own structure, identifiers, update logic, and methodology.
Before the data can be used to calculate Stock Characteristics, it has to be cleaned, standardized, validated, and transformed into our workflow.
This is a major part of the work that users never see.
Computation at Scale
We currently support:
- 15 investment universes
- 400+ Stock Characteristics
- 5 different time ranges
Every month, we recalculate 30,000 model portfolios used to evaluate individual characteristics.
These are not meant to be final investment products. They are research portfolios that help users understand how each characteristic behaves across time, universes, and market environments.
On top of that, we also run additional calculations for user-created portfolios inside the application.
The hardest part is not simply running the numbers.
The hard part is making sure the numbers are right.
Accuracy Matters More Than Speed
In investment data, having the right formula is not enough.
You also need to know:
- which data was actually available at each point in time,
- how index constituents changed historically,
- how to reduce survivorship bias, look-ahead bias, and other distortions,
- and how to make sure the results reflect the real world as closely as possible.
That is why we upgraded our data methodology in 2026. The goal was simple: more reliable historical calculations and better protection against common biases in factor research.
Much of this work is directly related to our recent methodology upgrade, where we significantly improved historical universe construction and bias reduction across supported datasets.
Read more about the methodology upgrade →
Where AI Fits In
We also see more and more investors asking ChatGPT or other AI models what to invest in.
These tools can be very useful for explaining concepts, generating ideas, or finding connections. We use them every day ourselves.
But most AI models do not have access to the large datasets, historical calculations, and infrastructure required for serious factor analysis.
AI helps us with reasoning, interpretation, and data monitoring. But the actual investment outputs still depend mainly on high-quality data, correct methodology, and robust computation.
That is why we believe the future is not about asking AI to pick a single stock.
It is about giving AI the right data and properly calculated inputs.
The Result
The user pays €19.90 per month.
Behind the result they see in a few seconds, there are thousands of dollars per month in data sources, cloud infrastructure, and computational power.
But our goal was never to sell raw data.
Our goal is to transform a huge amount of complex information into simple, usable outputs.
So investors can focus on the most important question:
Explore 400+ Stock Characteristics, long-only factor portfolios, and historical performance across multiple universes.
Explore Stock Characteristics