The promise of personalization?where stores in one neighborhood would have entirely different merchandise than one in a nearby neighborhood with a different customer mix?is nothing new. Retailers have been adding technology to their stores for years to test it out, in many different ways.
But almost all of those efforts were isolated and piecemeal. They were also ineffective, costing the retailers time and money but the band-aid approaches were too small to deliver the true potential of customization, says a report this week from Forrester Research.
"Retailers have delivered advances in capability, such as assortment planning, by bolting on best-of-breed applications because that?s much less expensive and risky than contemplating an entire planning and order management replacement," the report said. "But adding further point applications exacerbates the problem for retailers. They carry the costs and risks of a heterogeneous application portfolio that they can?t easily adapt to new opportunities."
The problem is especially difficult because customization?by its very nature?touches a huge number of different systems and areas, requiring programming hooks into so many different regions that is must be centrally coordinated.
The report cited one supply chain example. "In retail, product planning has become so complex and intensive that most retailers no longer consider merchandising part of the supply chain. Rather, it?s a standalone organization that interacts with supply chain. But the isolation of merchandising decisions from supply chain consequences risks unplanned cost overruns and service degradation," the report said. "Without a holistic view of the supply chain ? from product planning through logistics ? and without merchant accountability for the full ramiﬁcations of decisions, there is no incentive to minimize supply chain impact. Even worse, there is a risk of chaotic stock-outs followed by urgent inventory shipments and interstore transfers."
Even worse, true product mix customization also impacts a wide range of business systems. "As planners tweak the assortment and rationalize it to appeal to target customer segments, they have to see both the budgetary and the supply chain impact as well as the workﬂow they trigger, such as purchase orders or pick orders launched to deliver the assortment," the report said. "Each store must be able to execute their plans, taking into account the store?s space and demographic eccentricities. Despite recent consolidation, few vendors are able to oﬀer a true end-to-end suite today."
The report examined 13 vendors?4R Systems. 7th Online, Aldata Solutions, DemandTec, Escalate Retail, Galleria, i2, JDA, Manhattan Associates, Oracle, SAP, SAS Institute and Tomax?to see how close any came to offering a comprehensive line. Forrester selected 11 areas that they thought an idea system would handle: demand forecasting, financial planning, assortment planning, price optimization, size optimization, store clustering, inventory optimization, allocation, replenishment forecasts, replenishment and space optimization.
Few vendors came close, with DemandTec only supporting two areas (price optimization and store clustering) and Galleria only four (allocation, replenishment forecast, replenishment and space optimization).
Oracle and JDA were the only two vendors to include all 11, but the primary author of the report?Nikki Baird?cautioned against seeing that alone as a reason to consider either of those vendors.
"Just because Oracle manages to check off all of the boxes doesn't mean that they can work together well," said the veteran retail analyst, who left Forrester shortly before this report was finalized and is now an analyst with the Retail Systems Alert Group.
"If you think about a traditional ERP application, it's very structured," Baird said. "But in merchandising, it's such a fluid process."
Baird argues that such an extensive system requires a full?and expensive?full system overhaul. "It's time for merchandising to pay the piper. With this whole customer centric thing, you can't get away with not cleaning up past expediencies."
The challenge is that such a radical system change is risky in two ways. It's risky from an IT perspective if a company moves too soon and incorporates immature and untested applications, but it's just as risky from a business perspective if most of a retailer's rivals are customizing merchandise before they can.
"You don't want to be the first person to do it," said report co-author George Lawrie, "but you also don't want to be in the last dozen."
A compromise, Baird said, involves the 20-80 rule. In this case, that rule suggests that retailers don't really need to customize all of their products to give their customers most of the localization benefits. Such retailers, Baird said, can probably realize about 80 percent of the customized assortment value by personalizing about 20 percent of the merchandise.