➔
Chart showing the relative performance of flagship LLMs in extracting correct ticker information from unstructured text

How good are LLMs at tickerization?

For data to be marketable to investors, it usually must be mapped to asset-level identifiers, commonly known as tickers. An enterprise working to refine its datasets into valuable alternative data (alt data) products typically encounters the tickerization problem: given a body of unstructured data mentioning companies, such as social media posts, transcripts, or articles, extract […]

What Quants Fear Most

Why alternative data trials fail

The reality of selling alt data to quant funds Almost every data provider’s pitch includes a chart that looks like this: Unfortunately, live results rarely match backtested performance. In our conversations with quant teams at both large funds and small ones, we find that promising datasets often stall on packaging, not signal quality. The quant […]