|
From
March 2007
Off-the-Rack
Functionality
Ross Stores sold on
speed, utility of data warehouse
appliance
By Fiona Soltes
Say all you want about the Internet
speeding up business communications, it still won’t explain
the frustrations that employees of Ross Stores were experiencing
running time-consuming data reports online. Some took hours.
It was 2005, and something had to give. The current system just
wasn’t meeting the needs of the growing company, so the search
was on for a new solution for data availability. The choices
were fairly straightforward: re-tune what was already in place
or select a new platform. The costs would have been about the
same, so new won out, and Ross chose a platform from Framingham,
Mass.-based Netezza.
“Going into this, with Netezza being a relatively new player
in this space, we definitely wanted to do our homework,” says
Scott Lindblom, vice president of applications for Ross. “But
they came on-site, and we were impressed with that. They made
an investment and brought one of their boxes, took our data
and put it on the system and showed us what the capabilities
would be. They explained that we’d see the same results in production,
and the outcome really was as advertised.”
Andy Winans, vice president of retail and consumer packaged
goods for Netezza, lists a number of large companies that have
utilized the company’s data warehouse appliance, including Neiman
Marcus, Amazon, Cingular Wireless and Restoration Hardware.
Partnering with such large customers has been a great boon,
Winans says, but it has also had the unintended consequence
of creating “a misconception that our systems are only for organizations
with tons of data.
“We have many customers that have a terabyte or less of data,”
he says. “And they’ve been able to leverage the high performance
of our system for deep analytics without having to use a lot
of human and financial resources as [customers] would with other
data warehouse options.”
With just under 800 stores, Ross works with about nine terabytes
of data; some of the larger companies may have 20 terabytes
or more. “Ross is growing very quickly, with many stores in
geographically dispersed locations,” Winans says.
“And when that happens it becomes difficult for a buyer or allocation
manager without such a system to look at detailed data and make
the right decisions on assortment, allocation by store and other
business-critical areas.”
Pent-up optimism
In recent years, he says, some businesses have become disenchanted
with warehousing systems. The initial applications were easy
to build, Winans says, but “things start to slow down when people
want to add that second or third application, and really use
their system as an enterprise data warehouse. It becomes very
difficult to add different reporting capabilities and continue
to grow the warehouse.”
Netezza, he says, “brings to the table the pent-up optimism
that was in data warehousing 10-15 years ago.” And with the
amount of data being captured and analyzed continuing to grow,
it’s not a moment too soon.
“Businesses can change,” Winans says. “And
that changes the way people look at data, the way they do their
reporting and the way they want to look at their customers.
It could be that you get your merchandising up and running,
and you want to add the supply chain information, or give your
marketing people access to customer data. Today’s retailers
want more value from their investment.”
Ross implemented the Netezza appliance last summer and has been
concentrating on merchandising. As an off-price retailer that
is focused on close-outs, the company has a larger-than-average
number of SKUs and doesn’t always know what specific products
it’s going to sell.
“We’re not really doing a lot yet with store reporting,
although we have recently started reports on shortage or shrinkage,
which is a large cost to the company,” Lindblom says.
“We’re trying to practically manage that. And in
terms of store operations, we could definitely go through additional
analytics.”
All the same, he says, the company has already experienced a
significant improvement in data load, and has spent less time
on system maintenance and more on developing new capabilities
than it had in the past.
“There’s a lot of horsepower,” Lindblom says.
“If the old solution was a V-6 engine, now we’ve
got a V-10.”
Winans won’t argue with that analogy. “There’s
a lot of technology in there that allows it to process data
much quicker,” he says. “Moving the processing very
close to the data allows us to see 10- to 100-times performance
gains against our competitors.”
Seamless on Day 1
And the appliance packaging means the system is very easy to
install, maintain and run; the entire product is about the size
of a refrigerator. “We can roll a box off a loading dock,
and within hours, load it up, plug it in and add data to it,”
Winans says.
The implementation was “seamless,” Lindblom says.
“We got it to work Day 1.” Ross ran it parallel
to the previous system for a brief period to ensure the systems
matched, and then went live in April. Since then, it has upgraded
to a new version of the system for added responsiveness.
“It’s really a scalable system,” Lindblom
says. “Our business, as an off-price retailer, is a little
different from a traditional retailer.
“There’s a lot of focus on identifying trends and
trying to chase those trends to increase sales and profitability.
And having the data to be able to do that is key to our success.”
The way Winans sees it, however, Ross Stores has only “just
begun to scratch the surface” of what the platform can
do with the data.
“Some customers are interested in doing more one-on-one
marketing, and others are expanding out to other application
areas … to see how good the return on their investment
will be,” he says.
|
