Wegmans Fixes the Checkout Productivity Problem

[Note: This was a major issue that took months to understand and impact.  I'm including several key excerpts here separated by:

   ...........             ]

Checkout Productivity Issue

The introduction still faced many challenges and the number was growing. Chains continued to point to insufficient source symbol marking. But a new concern was being increasingly expressed. Even in stores with higher U.P.C. symbol levels, the expected level of checkout productivity was not being reached. Wegmans, Giant Foods, SGC’s Pathmark store, Shop Rite in Fort Worth, Steinberg’s and others all were reporting they had not been able to reduce the checkout schedule to the degree they had expected. IBM’s initial reaction was to work with Giant Foods to video tape checkout in a scanning and a Mechanical Cash register store for analysis.

Mid-April found me at Wegmans getting more personally familiar with the scanning productivity issues they were concerned about. Erik Lunkenheimer had raised the flag for this being a significant problem and some response was required. En route I stopped by in Albany where Dave Goliber and I discussed strategies on getting Price Chopper to move forward.
Erik met me at the curb outside baggage claim at the airport in Rochester, NY. We drove around to the headquarters of Wegmans located adjacent to the airport and met briefly with Bob Wegman. He expressed his concern about the lack of checkout productivity, but also his understanding that these changes in procedures sometimes have to iron out the kinks. After Erik picked me up the next morning we spent our time in the store seeing if we could find anything obvious.

The only thing I observed was something that had been painfully obvious since before the first system shipped. The printer had a significant delay in time between when the item was scanned and when it started printing. Even after it was printing, it wasn’t very fast. It might take three quarters to well over a second between the time the scanner beeped that it had a good read before the printer started printing the item description and price at the receipt print station and then another second or more to complete the print line. It appeared that checkers mentally waited for the printer to complete printing. That little delay put a kind of stutter step into the rhythm of scanning and bagging groceries. Maybe that choo-choo train protocol wasn’t fast enough at 2400 baud on the supermarket loop? If not, we’d have to document that. While we were out we walked through and observed an ESIS store and a NCR class 5 register store. Then we returned to Wegmans' headquarters and shared our observations.

The Wegmans people agreed with us and had made similar observations on their own. I discussed what I knew about the work we were to do with Giant foods and video taping checkout in the store. Bob Wegman wanted to hear what we learned from Giant when the results were obtained and also expressed willingness to have us return, and if it would help, to do similar video recording in Wegmans’ stores.

Slow printing had been first identified as an issue in the early days of the Troika, early in 1974. The response at that time was that printing had to be that slow because Field Engineering would not sign off on the product, if the printer ran faster. They argued that at a faster speed, the life of the print wires would wear much faster and that raised their costs, so they would raise the price for service. The Data Processing marketing division was still concerned about the high cost, so increasing the print rate was deemed a bad trade-off. Now, it appeared the slow printing could sink the customer’s productivity benefit. Field Engineering might end up making a fine margin, but on no business.


Checkout Productivity Issue

When we returned on July 2nd to Wegmans to document store productivity, we video taped live checkout. Wegmans gave approval to install a camera over a regular checkout lane, lane 4, on a Friday, a very busy shopping day each week. Erik and I went into the store after closing on Thursday night. Using a ladder we pulled a ceiling tile back far enough to put the camera in over lane 4 looking down at about a 60 degree angle. It also included sound. We needed to hear the printer run and hear the scanner beep as well as see the scanning process. The camera cable was strung in the ceiling above the ceiling tiles about 65 feet to the locked room where the store controller was located. There it was connected to a cassette recorder with a small monitor. I had about 8 blank cassettes.

The next morning we returned about mid morning. We only wanted to record the activity in high demand times, so there was no point to being around before the store started to get busy. Wegmans’ management had briefed the store manager and he could not have been more cooperative. Erik left a short time later since doing this was essentially a one man job. In a conversation with the manager and the head checker, they agreed to have their best checkers and some average checkers put on checkout lane 4.

We wanted the fact that we were recording to be as far out of the checker’s mind as possible to get the truest representation of normal checkout. Of course the checkers knew we were there. Three or four times a day I’d go back to the break area in the back of the store to get some coffee from a vending machine. The checkers would recognize me as the unfamiliar face hanging around the store office and ask me what I was doing. I told them I was from IBM and we were measuring the equipment. I never saw the slightest hint of disbelief, distrust, or other response that would indicate they were uncomfortable about what was going on. They accepted it.

During the day 5-6 checkers were recorded working on the lane for a total of 10 to 12 hours. When head checker or manager would assign a checker to lane 4 I’d start watching on the monitor. About half the time the checker would give some initial indication that they knew this was the lane with the camera because they would be unusually stiff in their interactions with customers or they might stand a little more rigid than usual. Once, right after she’d been assigned to Lane 4, I caught a checker’s eye glance up to the ceiling tiles to check where the camera was located. But we didn’t start recording until well after these mannerisms had stopped. In all cases, customers started coming up, those cryptic short conversations that are normal between checkers and customers happened, and in 15 to 20 minutes the checker had pretty much put the camera out of conscious thought. That’s when recording started. There was no evidence of a customer ever being aware the checkout was being recorded. When a checker starts adjusting something that bothers her with her undergarments or shares a slightly off-color remark with a shopper, it signaled that she’d forgotten about me for now. We’d get a record of the sales on that register at the start and end of recording each checker to measure what was accomplished in the tape.

Back in Raleigh the next week with my boxes of tapes we found some unused offices near my own and set up a cassette tape player, monitor and a repeating timer. The tapes were analyzed using a technique called “work sampling.” Forms were created that identified the various work elements of checkout. As the observer, although I received plenty of offers to help after people, I would mark what activity the checker was doing each time the timer beeped every 10 seconds. That resulted in more than 2000 data points. From those you could with a level of statistical significance, determine the time spent on each work element.

But after the analysis was completed we decided it wasn’t enough. These initial results confirmed that our scanning checkout productivity was not much different, if at all different, from what we thought could be done with old mechanical cash registers. We wanted more data and on more different types of checkouts, so we arranged to return to Wegmans at the end of the following week.


Checkout Productivity Issue

Interest was growing in IBM’s analysis of checkout productivity. Tom Wilson had been talking with Bill Carey and getting summaries of what we were finding. He requested that I come up and brief McKinsey on what we had discovered. So at the end of the fourth week in September I stopped by the McKinsey offices in New York City on my way to present our analysis to Wegmans in Rochester. We showed Tom the Wegmans presentation. We knew that the lag in the printing sound contributed to the lower productivity. But, we also knew that the system operated faster than the checkers were doing in our videos. The net was there was no simple solution from us at this time.

In Rochester it was a command performance. After it was over, the Wegmans’ management indicated they appreciated the quantification of what they were intuitively feeling in their own stores. We both agreed we would continue to work on improving performance in our own spheres and to provide the video tape recordings to Wegmans. A package with those tapes was sent from Raleigh the next week. I left Rochester to help the Store Systems staff in Boston put together a presentation for Stop & Shop.

While we were all letting the results settle in and considering what options IBM might have, I got a call from Erik Lunkenheimer in Rochester. He told me Wegmans was “throwing IBM out.” “What does that mean?” I implored. “They want us to stay out of their stores for about six weeks and then we’ll see what is next,” he replied. This sounded terrible, but realistic. Erik added, “They said we can’t fix the problem and it’s time for them to try.” While this was quite challenging to my “either IBM or I can help anything” ego, think as I might, I had no better alternative than to honor their request.


Checkout Productivity Issue

As I stopped by my office the first week in December, I had a message from Erik Lunkenheimer. I returned it to find it was the good call from Erik that I had been waiting for. Wegmans was inviting us back to re-measure their stores. They had quietly gone off by themselves and worked on the scanning productivity challenge for about 8 weeks. They reported that it appeared the productivity had increased. Bill Carey told Tom Wilson of McKinsey about the call and Tom asked if one of his staff could accompany us when we re-measured the stores. That would be great since, if we documented good news, it would have more of an industry affect if McKinsey and Company had participated.


Checkout Productivity Issue

...I met the young McKinsey staff person in the evening in Rochester the second week in December 1975. We had dinner and then went around to the store to re-install the camera over Lane 4 with the other recording equipment. My casual observation that evening with only one lane operating before the store closed, suggested there was a new focus on checkout. It looked like Wegmans had broken their checkers habit of waiting for the printer to print. The checkers moved ahead to the next item based on hearing the beep when the scanner registered a good symbol scan.

The next day the McKinsey consultant and I arrived at the store again after it had started to get busy and went to the controller room. After talking with the manager, we tried to stay out-of-sight to the checkout staff. This time the store manager took more interest while we were taping, often standing with us in the cramped controller room where we had located the video tape recorder and the monitor. The room happened to be next to the locked area where checkers turned in their cash drawers when returning from the checkout lanes.

Once when the manager was with us, for some reason the door was not closed tight. The head checker was turning in her cash drawer after working a lane and caught sight of the monitor. “Oh, is that…” she exclaimed, but she was cut off by the manager who interrupted with, “Yes, how could she be working more productively?” The checker responded with some hasty suggestions, none of which really mattered. What mattered was that the checker’s sense of management’s priorities had been reinforced on how important productivity was to her store manager. I realized I had just observed another of the ways that Wegmans had addressed the productivity challenge.

The manager was true to his word and kept the pressure up on the lane where we were recording. Checker’s seemed just as talkative with the customers, but the conversation seemed to take place at the start and during the tendering portion of checkout. When they were ringing up items, they stuck to business.

It appeared during the recording that productivity had improved. When we got back to Raleigh and did the “work sampling” on the videos, we found Wegmans had in fact brought scanning up to where we thought it would be all along. The McKinsey consultant was with us in Raleigh for about two days; long enough to see where it would come out. After my return to Raleigh, I started my days by dropping by the A&P test store, but since things seemed smooth and they had somehow lasted without me the prior week, I’d go over IBM Bldg 602 and start the work sampling on the tapes.

As we considered the results we had to give high marks to Wegman’s managers. They had changed the checker’s perspective on what was the right work rate. They had broken the interdependency on hearing the printer run before scanning the next item or rather replaced the printer’s sound with the beep the scanner provided after a good read. It provided perspective to IBM marketing and development management that the equipment didn’t determine the productivity, checker’s attitude did. And more specifically how those attitudes were influenced by their managers. The scanner could be as fast as the fastest electronics, but it was the checker who paced the operation.

It was a good Christmas present for 1975. Someone had shown you could lick the productivity bugaboo. Wegmans was happy with the results, IBM was happy with the results and McKinsey was happy with the results. There was a success story to tell.