NYSE Technologies has several investment firms participating in a new pilot -- instead of downloading all the end of day trade and quote data they are taking the result of NYSE Technologies’ analytics. Actually, said Jennifer Nayar, head of global data products at the firm, most are taking both so they can test the value of the analysis NYSE Technologies performs.
“We have been in the business of offering end of day trade and quote data for the last 8 years or so, working with 1010data. That involves multiple customers coming in at the very end of the trading day and downloading lots and lots of data and putting in their own sites and running their own analytics against that data.” It was time consuming and costly, especially for data storage.
NYSE Technologies, through its Market Data Analytics Lab, thought to offer the analyzed data so customers wouldn’t have to download the same information and build the same analytics. The analytics include statistical modeling, functional grouping, multiple regress analysis and time series analysis, she said. The analytics are from 1010data, whose work NYSE Technologies white labels, she added.
“We have been met by pleasant surprise,” she added. Skeptical quants have found that the results they received from NYSE Technologies seem to be meeting their needs.
“But we need a larger sample size to make a general statement,” she hastened to add, sounding like a bit of a quant herself. “There will always be customers who like to run their own analytics and have their own secret sauce.”
Many users, however, find no benefit in downloading the data, normalizing it and running standard queries.
“If you want to get to the data results as quickly as you can, this allows you to get ahead of the game, and it saves time and money since you don’t have to bear all the storage costs as market data rates increase.”
She said 1010data has been a partner for nearly 10 years and acts as the company’s data warehouse with trade data going back 15 or 20 years. NYSE Technologies recently launched similar services for the Euronext market, she added.
Sandy Steier, CEO and co-founder of 1010data, said the company’s analytics are used across financial services, including data-intense areas like mortgage-backed securities. It delivers through the cloud, reducing capital expense for its customers and providing easy expanson as business grows.
“Most of our customers use us because we provide a raw analytical power and processing power that is unmatched. Companies like Goldman Sachs that have enormous resources in technology became our customers because there were certain things they were not able to do that we could.”
Although 1010data works in big data terrain, it does not use Hadoop which is too slow for most financial services firm.
“We can handle a lot of data, but our system is designed for higher performance, deep analytics and ease of use.” It is based on a columnar database, a technology he said resurfaced two or three years ago, just in time to be overshadowed by Hadoop.
“Our database is incredibly fast. In customer benchmarks we invariably outperform every competition in terms of how quickly data can be analyzed and how quickly we can get data in. We don't capture data in real time,” he added. “We are meant to analyze historical data and to refresh in less than real time, such as hourly or daily.”
Trading desks like 1010data because it offers a spreadsheet-like user experience along with much deeper analytics and the ability to access various kinds of data, mingle it together and cross-analyze it.
“In the mortgage-backed securities marketing in particular, we are pretty much the standard analytical platform.” It can look at data about individual mortgages, if payments are made on time and add analysis by geographic areas to show the trends, including unemployment, in a zip code.
The company also does extensive work in retail. AutoZone uses 1010data to check car registration data around each store.
AutoZone realized a store in an urban area would have a different clientele and different types of cars from a suburban outlet. By checking registrations they could determine what to stock and what to advertise in the circulars the stores send out.
When it began analyzing customer data, Dollar General found their concept of shoppers was far from reality. They served two distinct groups -- people who shopped for convenience and would buy one or two items per trip, and a much more loyal group who would buy one or two dozen items on each visit. Dollar General found they could sell more to this group.
Among other types of analytics Dollar General is using 1010data for affinity analysis -- if someone buys a 2 liter bottle of Pepsi, do they buy something else, and if so, what do they buy, explained Tim Negris, vice president of marketing. Once it knows that customers are apt to buy more, the store can put Pepsi on sale and generate additional sales of other stocked items. It can also decide how to target the most profitable customers with coupons, newspaper ads, direct mail or other forms of promotion.
“Retailers have a gold mine of customer information for manufacturers,” said Steier. “In some cases they can charge for that information -- data they were doing very little with just a couple of years ago.”
