Spokesmen for digiMine, the early supplier of analytics as a service over the Web, make no bones about it. Their early customers were many of the now-failed dotcoms and they got off to a fast start providing them with subscription services.
The subscription model—based on a monthly fee instead of a one-time licensing deal–fit dotcoms because they needed an analytical service that could be up and running quickly at a low initial cost. digiMine, founded in 2000, filled this need at the peak of the dotcom spending and supplied its services over the Internet.
Then the recession that lead to the great dotcom die-off changed the climate to one in which large established companies began to consider online analytic services.
“The shift has been very clear to us,” said Omar Tawakol, digiMine’s vice president of product marketing, who has seen his firm add Dow Jones & Co., Starbucks Coffee Co., J.Crew Group and Microsoft’s Xbox unit’s www.Xbox.com to its stable of customers over the past 11 months.
In addition, it is providing analytical services to CNN, CBS Marketwatch, Real Networks and Nordstrom Inc. department stores. It has a major bookseller and an automobile manufacturer among its customers, although Tawakol said he was not allowed to name them.
Large companies are willing to adopt an online application vendor—in effect, outsourcing a specific data analysis need—because they are “pretty disciplined about doing analytics with a clear return on investment,” he continued. As they confront masses of customer data gleaned from their Web sites, they must originate data mining techniques for a new channel.
digiMine Vice Pres.Omar Tawakol
“The Web as a channel produces a lot more behavioral data” than predecessor channels, such as catalogue shopping or walking the aisles of a brick-and-mortor store. “CBS Marketwatch had so much data they weren’t able to process it in-house as fast as it came in,” he said.
Much of the data is unstructured, which makes it harder to store and analyze. Enterprises’ data warehouse operations typically deal with well defined and “scrubbed” data that has had ambiguities removed. Web site data is frequently “unstructured data,” often derived from a click stream with unpredictable elements to it. Analyzing that data amounts to a “technology disruption” for the in-house warehouse builders, so there is less opposition to farming out that function, Tawakol said.
digiMine and other data online analytic vendors have also sought to tailer their services to individual customers. digiMine has a custom engineering group that tries to adapt its service to the internal systems of a customer, relate Web to other customer data and produce greater flexibility in reporting. “We understand our own application programming interface,” so there’s a short learning curve for connecting digiMine’s analytic applications to customer data sources, he said.
digiMine was contacted by the Wall Street Journal a few weeks before they were due to launch their revamped site last January. “We had very little time, but we told them we could get up and running with them very quickly,” Tawakol recalls.
But its main asset has been the change in economic climate that makes using an online provider more attractive to large companies. “If we had mapped this out on a white board, it wouldn’t have happened,” Tawakol.
digiMine is privately held and declined to release details of recent financial results. It recently completed its third quarter, which Tawakol termed “our best third quarter ever,” without disclosing figures.
By: Charles Babcock