February 17, 2014

Opening Up the Estimates Business

Many of the new founders have worked in finance where they pinpointed a gap or weakness in the technology or workflow of financial analysis. Drogen previously was a trader for hedge fund Geller Capital, and then worked at StockTwits, before he set out to build a crowdsourced model for earnings estimates that would be more accurate than traditional sell-side analysts. Today Estimize covers 900 stocks per quarter and is more accurate than sell side analysts 69.7 percent of the time.

While Estimize concentrated on gaining the buy side and independent analysts to contribute estimates, it's bringing on sell-side estimates as well. Now the firm is adding earnings estimates from Needham, Suntrust, Oppenheimer and Susquehanna. "We're basically going to capture the whole data set as well so we'll own that too.

Drogen says he founded the company to transform "the general ecosystem around data and estimates." Launched in December of 2011, Estimize has grown into a community of 4,000 contributing analysts and 28,000 registered members, up from 3,000 contributing analysts and 16,000 registered users last September.

One characteristic of the startups is that they grow fast and are constantly adding new types of data. In line with this trend, Estimize plans are to roll out macroeconomics data, such as GDP, CPI, jobs numbers, oil inventories, retail sales and other data. It's also getting attention from investor relations departments who typically call the sell-side analysts to get color on their companies.

While anyone can join Estimize and contribute estimates for free, Estimize is adding premium services on top of the public data. It charges the hedge funds that want to receive the estimates into their models.

Making Time Series Data Searchable

Recently, Kensho sparked media coverage for developing "Warren" a market assistant named "Warren" that can answer complex financial questions posed in English about global asset prices. While "Warren" uses advanced natural language processing, New York-based Eidosearch has brought a content search engine to query the world's time series data throughout history. "The whole idea is that people sit on a mountain of time series data, and there's fascinating relationships they would like to uncover," says David Kedmey, president and co-founder of EidoSearch. "But unless you give people a natural way to query for that information those relationships remain untapped," says Kedmey.