Bloomingdale's And The CRM App That Associates Were Never Asked About

One of the oldest adages in IT is that corporate tends to issue edicts without checking how they will play in the stores, without asking the foot soldiers and store managers whether it's a good idea. A consultant this week offered a deliciously illustrative example, and it involves Bloomingdale's and a CRM project.

The coincidental number that was the program's weak spot? The required information-gathering with a customer took about 10 minutes, which is also roughly the same amount of time it takes most associates to make a new sale. They receive commission for those new sales and nothing for filling out the forms. Wonder why the program didn't work well?

The consultant, who asked that his name be withheld, said he was asked by Bloomingdale's to figure out why the program wasn't delivering. The program was designed to gather a lot more information about customers, to be added to that customer's CRM profile. The first problem was that the application sought far too much data from customers. Other than annoying customers (which would reduce cooperation and is bad for 100 other reasons) and wasting sales associates' time (wouldn't you rather they be selling?), much of the data was simply not needed.

"It wanted to know the personal assistant's phone number. A third E-mail address. Where they had certain houses for certain dates throughout the year," the consultant said. "They wanted so much dream data. They had these execs in a room who wrote down everything they wanted to know. Everything that the corporate team thought was important."

So was the data even used? Corporate "pretty much ignored 95 percent of the data."

Added the consultant: "It took 5 to 10 minutes to put one sale into the database, and they weren't paid any incentive. If I'm wasting this time, I might as well get another sale in." That was from the sales associates. Their direct managers had similar thoughts, because the managers are paid on productivity.

The system also glitched periodically, said the consultant, saying "the profile is not complete on John Doe" and related issues that made completing the reports in the field challenging.

We reached out to Bloomingdale's for comment and no one replied.

There are many things that corporate knows better than the field, but the best way to make local sales—especially with a business as personal as Bloomingdale's—is an area where the stores' view must be supreme.

Failing that, the next key question is how to get associates to carry out corporate edicts. The old school kneejerk answer—"They either comply or they're fired"—doesn't work with sales associates. They know that substantially beating their sales targets is job security, one that allows them to ignore all but the most important rules. Can they steal from the store? No. Can they assault a customer or another employee? No. Can they show up for work an hour late? If they're blowing past their numbers, most likely they can, and they know it.

What about filling out screens of data for corporate, something that even their manager doesn't care about? Nope.

More importantly, if corporate understands the issues, would even corporate want it?

More importantly, if corporate understands the issues, would even corporate want it? In other words, was there likely anyone in that executive meeting room who projected the likely sales, revenue and profit reductions if this program was enthusiastically embraced by associates nationwide?

Was the possibility of this program causing a sales drop even considered? Sometimes, a little disobedience from field troops is the best thing a general can hope for.

Asking executives for feature wish lists is a long and truly terrible tradition in retail IT. Everything on a wish list costs money, and this is one of the few places where taking a hard ROI view of IT makes absolute sense.

When a CIO is asked the question, "How much will it cost to ask just one more question?" the CIO tends to answer in terms of how much storage that data will require. (Programmers and DBAs will already have to do work on applications and databases, so the incremental cost of one more question is tiny.) But every piece of CRM data that's not used is another potential privacy and security problem, both after the data is stored and after questions are asked in the store. There's a risk associated with that process, and risks are a potential cost that can be calculated.

If executives are willing to budget a process that asks time-consuming questions at the POS, it still should be piloted—and then rolled out progressively, and the results of the progressive rollout carefully examined. Associates are great at working around bad ideas, especially if they have the tacit approval of their managers. For example, if those managers encourage associates to collect just one missing piece of information—but only one—eventually the record will get filled in with no disruption in the sales process. But if IT doesn't support that approach, the sales process becomes an endless series of manager overrides.

The biggest cost of adding in just one more question is that it makes it harder to do what you're really trying to do, which is to find answers that will boost sales. The more distracting, unnecessary data in the app, the harder it is to find the answers and—frighteningly—the less likely it will be that colleagues will keep trying to find those answers. It's the old noise-to-signal-ratio problem. Say what you will, but corporate just loves creating noise.

Another issue: How annoyed are customers likely to get at all these prying questions coming from an associate who is delaying closing the deal by asking for information that's not required to hand over money?

Again, this is a place where store associates have a much clearer understanding of how much customers will take. The last thing you want is to upset a high-end customer because an associate insists on asking for the phone number of a summer home before handing over the merchandise.

Unless you make a lot more money from CRM than from selling products, of course.