Fingerhut's Updated Data Warehouse Increases Profits

Share this article:
Fingerhut Cos. Inc., Minnetonka, MN, is finding that it's updated 7-million-count customer data warehouse system is bringing in record return on investment.


Every day, Fingerhut, which has annual sales of $1.5 billion, mails an average of 1 million catalogs -- selling everything from automobiles to clothing -- at an annual cost of $600 million.


Its warehouse amasses up to 1,400 individual pieces of information -- including birthdays, anniversaries, payment preferences, product purchases and product interests -- for each customer. This information is gleaned from a variety of channels such as the phone, Internet and customer questionnaires sent with catalogs and returned by mail.


To manage this data, Fingerhut originally relied on a Promotional Scoring System, a production system that scores Fingerhut's customer list each week with customized segmentation models for each individual catalog. PSS would score the value of customers and create catalog distribution lists based on these scores.


However, Fingerhut wanted to optimize customer selection over the many catalogs customers receive -- Fingerhut has more than 100 catalogs. PSS, which was designed and built inhouse in the mid-1980s, would decide for each of the 100 catalogs which customers were profitable.


"Obviously, we do not want to send that many [catalogs] to any customer, so we try to choose the right catalogs from all the possible ones and send only those to the customer," said Randy Erdahl, group manager for corporate research at Fingerhut. "We needed to find a system that would help us better eliminate those catalogs from being sent to a customer if it did not add incremental revenue or profit from the other catalogs being sent to that customer."


Fingerhut developed a Mail Stream Optimization program, which consists of IBM SP2 hardware and Orchestrate software from Torrent Systems Inc., that allows companies to process more data in less time by partitioning data across multiple processors and then streaming the data through multiple, parallel instances of each application step. This solution allows Fingerhut to process multiple customers simultaneously during the scoring cycle of MSO.


Without parallelization, Fingerhut projected it would take more than 20 days to run the MSO system. Now, the system runs in less than a day on four processors with parallelization.


"MSO enables us to choose a better subset of catalogs from the 100 choices to send to the customer," said Erdahl. "The mailstream [or sequence of catalogs] the customer receives better matches their interests and needs than what they would have received under PSS."


Currently, Fingerhut's system cross checks the needs of 7 million customers that are scored against 20-40 catalogs each week. The scoring process produces more than 700 million scores each week, and the optimization process must find the single best solution from more than 1 trillion possible candidates.


Instead of taking 72 hours to run the system -- Fingerhut's original goal -- the system finished in less than 12 hours.


MSO cost Fingerhut more than $1 million to design and build. But this year Fingerhut is projecting that it will realize millions of dollars in income benefits from MSO because of the advertising reduction.
Share this article:

Sign up to our newsletters

Follow us on Twitter @dmnews

Latest Jobs:

Featured Listings

More in Database Marketing

What's H-appending? DiscoverOrg Taps Marketo's Webhooks

What's H-appending? DiscoverOrg Taps Marketo's Webhooks

Cloud-based marketing automation behemoth Marketo joins forces with marketing intelligence company DiscoverOrg to improve its data collection capabilities.

A Toast to Marketing Attribution

A Toast to Marketing Attribution

Vino accessories and storage company Wine Enthusiast indentifies top and underperforming affiliates using algorithmic marketing attribution.

Q&A: When (and How) to Bust Down the Data Door

Q&A: When (and How) to Bust Down the ...

Some people run into issues with trying to build the perfect solution when often an 80% solution will do, says MailChimp's chief data scientist.