Wednesday, February 27, 2008

My early days with Wal-Mart data processing

When I was in college, Wal-Mart was still a fairly small company, I'm guessing around 300 stores in perhaps a 6-state region, with only 5 warehouses (now called distribution centers) two of which were in the headquarters hometown of Bentonville, Arkansas (one attached to the general office).
I had job hunted in Houston, Texas, where my older brother once lived, and actually had a computer programming job offer from the Fluor Corporation. But during the last semester of school here at The University of Tennessee at Martin, I happened to attend a meeting of the ACM (the only meeting I ever attended, as my job as a cook at the local Shoney's ususally precluded my attendance), and the speaker was the vice president of data processing for Wal-Mart, Glenn Haburn. His talk about a large computer operation in a small town (Bentonville at the time was population 10,000) sounded appealing, so I got his business card and followed up with a letter, which got me invited to visit at their expense. I flew from Memphis to Bentonville on a small Skyways airplane (they used joke about it as "scareways"), landing in a small airport in Fayetteville. The department secretary drove me to the general office, and the route we took passed a lot of rural farmland with cows and chickens -- no sign of a metropolis in those days! This was also before the nearby company, Tyson, made it big selling chicken to MacDonalds.
I got a personal tour of the local warehouse by the directors of data processing, Frank Parker (who nearly always dressed in a cowboy shirt, and was one of the few sporting a beard, which was only allowed for data processing folks) and "John T." Williams (a former IBM-er assmebler-programming guru who hired on with Wal-Mart). Also a tour of the data processing area (it seemed like there were only around 3 dozen application programmers at the time, plus perhaps 4 systems programmers, a few operators, and an area called Data Communications that handled the transfer of data between the store computers and the general office; at that time through a telephone interface, as there was not yet satellite communication). Anyhow, we were mutually impressed with each other, and a job offer soon arrived.
I started work there on August 29, 1981. Right after graduation, I had already committed to serving as a Baptist summer missionary in Baton Rouge, Louisiana, and thankfully Frank Parker ok'd me to delay my job start until that was completed. My first assignment was to the Warehousing project. Project leader was Randy Mott (who later rose thru the ranks to become vice president of Information Technology, then later hired by Dell as CIO, and then CIO of HP). A really great guy, with his famous run-together phrase "gotolunch?" and stacks of green-bar printouts of source code and reports proudly bearing his username, PROGRM. Others on the team were Wes Beckham, Tina Peltz and Gary Evenson. Wal-Mart was a "big blue" (IBM) shop, running a good-sized IBM mainframe and the VM, VSE, and MVS operating systems (the VSE stuff was migrated to MVS), and CICS was the transaction processing system, which I learned on the job. It was actually a great opportunity, because back in those days most processing was "batch", and few programmers had yet learned "online" processing. The primary language was COBOL, although the Store Network project coded in IBM 370 Assembler.
Gary Evenson and I later moved into a new project, spun off from warehousing, called Traffic and Trucking. The trucking system, written for CICS in COBOL, was based on 3 DL/I databases (IBM hierarchical databases, the big thing at the time before relational stuff like DB2 came along), one each for Drivers, Trucks, and Trailers (keep in mind that the semi truck that pulls a trailer is distinct from whatever trailer it might be pulling behind it). The traffic system, on the other hand, was for providing online data entry and processing of the stuff transported by other trucking companies. Gary was project leader. He liked to smoke (still ok in those days -- he gave me a small desk fan to blow the smoke back his direction in case it bothered me) and whose main hobby was canoeing. Gary drove an old Ford Mustang.
With my knowledge of Assembler programming, I was able to become project leader of the Store Network Host project (which was the mainframe side, complemented by the Store Network project led by Rusty White. The stores at the time had IBM Series I computers, and Rusty's group coded in a special assembler macro language that I think they developed themselves. Our side used IBM 370 Assembler, and wrote macro-level CICS (compared to command-level CICS). Other names I remember from the Store Network Host project are Rick Bowen (who helped me build a cedar fence) and Karla Poye. I remember Rick learned programming while in the Marine Corps, and was an ace Assembler programmer.
Later I became the first Training Coordinator for data processing. Prior to that, we had training material and I had been given the ad hoc responsibility of coordinating it, while still a project leader. The then-VP of data processing, Billy Martin, announced the new position at a departmental meeting.
I had developed an interest in systems programming, and so later moved from being training coordinator to systems programming for the VM and TSO project under Steve Marack. We wrote in Assembler.
During a trip to Martin, Tennessee, while talking to Dr. Otha Britton, who was my advisor at UTM and now the director of the computer center, I was made aware of a job opening for a programmer at UTM. Although I was not looking to change jobs, it sounded interesting, as my wife's family was in Martin (compared to being 8 hours away in Bentonville). After much soul-seeking and prayer, I opted to take the offer, and started there on October 1, 1987. So my total time at Wal-Mart data processing was 6 years and 1 month.
One of the "grassroots" things I did at Wal-Mart (which was later copied by many departments) was putting up a bulletin board of photos (I got a Polaroid camera for it) of everyone in the department, as we were growing enough that I could not always remember everyone's name.
Another thing I took advantage of was Wal-Mart's "open door policy" which basically meant you were free to talk to anyone in the company, without going through a chain-of-command first, at any time. I opted to talk to the president, Jack Shewmaker, and later on president David Glass, using this policy. I just went over to "executive row" and asked if they were in. Of the two, I was most impressed with David Glass for talking one-on-one; Jack Shewmaker was best at talking to larger audiences.
Wal-Mart stock was doubling and splitting every year. Sam was still around (I was in close proximity a couple of times, but can't recall if I ever actually spoke to him). We first beat top-dog K-Mart at sales per square feet, and later beat them at total sales. The first Sam's Wholesale Club opened in Oklahoma City, just 6 months after Sam said he wanted to start that up. I remember another project in data processing scrambling to retrofit some programming to support the new division, since up to that time only regular Wal-Mart stores existed (there were not any Super Centers yet, either). I also remember Wal-Mart making its first billion dollars of sales in a single year. I used to refer to K-Mart "the K word" and would not actually step into one of their stores for many years.
When I first came to Bentonville, what is now a museum was actually still a (very) small Wal-Mart store located at the court square (Bentonville had a small civil war park, and traffic went around it; county office buildings and shops surrounded it). There was also a nearby "sample store" that Wal-Mart associates (don't ever say "employee") could shop at for merchandise that was not generally sellable (perhaps, for example, a damaged box, but the contents were fine). Sort of like an internal flea market. This later moved to a much larger location. But all of this reminds me of how Wal-Mart started out very small and humble. It was a fun place to work, with a lot of good folks, including many practicing Christians (Jack Shewmaker was a member of First Baptist Church in Bentonville, where I also attended). Sam's success was in hiring good folks dedicated to seeing the company grow and succeed, and at least in the early days, there was very little turn-over. (Turnover in data processing was almost non-existent).
A point of trivia -- the first Wal-Mart is not in Bentonville, but rather in neighboring Rogers. Bentonville is store 100 (not sure if the 100th store opened, or if perhaps that number was reserved for it). You can tell the relative age of a Wal-Mart by its store number -- Martin, TN for example is 107. I remember when data processing scrambled to handle store numbers over 999 (sort of like the "Y2K" thing of handling years past 1999 when they were stored as 2 digits like 99).
One interesting thing I worked on was actually the first experimental "online" credit card approval system for Wal-Mart. I wrote something in Assembler that would take a request from a store and then communicate to the clearinghouse bank over a connection called VTAM. We tried out this home-grown programming with one of the stores. I don't recall what became of the project, but at least it was a "proof of concept". Of course nowadays this sort of thing is taken for granted, but at the time, Wal-Mart did not do "real time" approval of credit cards -- if you used a credit card for a purchase over 50 dollars, the cashier had to call over a manager who would call in for approval. This was also back when all items were individually stickered -- I also remember when UPC bar code scanning was a new thing. Another "first" for me was being the first at Wal-Mart to learn and use relational database programming (SQL for DB2). It was another proof-of-concept assignment, but I did indeed self-learn the basics of SQL (and the QMF utility).
The first IBM PC came out in 1981, the same year I started at Wal-Mart. It was interesting to see these new microcomputers, since everything up to that time was mainframe (we used 3270 terminals that were hard-wired to the mainframe through miles of coaxial cables). Outside of regular hours, I wrote a small program, in BASIC, for a local CPA to process tax returns on the IBM PC under the DOS operating system (this was a few years prior to Windows). I also attempted to port it to CP/M for another CPA who was using that instead of DOS (CP/M was an early alternative to DOS on the original IBM PC's, but it went the way of the dodo bird and betamax tapes). Speaking of betamax, I also recall the videotape standards war between them and the VCR tape everyone now uses. If you bought a betamax player, your investment soon became as useful as an 8-track player (remember those?) I did not see a Macintosh until I went to work for UTM in 1987.
When I started, Wal-Mart's stockholder meeting was still held in the local high school gym, and then Sam would have out-of-town stockholders over to his house for BBQ (not being from out of town, I never got to go to that). Later, after the HQ built a good-sized auditorium, the meeting moved there. Then later it moved to a larger facility I think at the University of Arkansas in Fayetteville.
I visited Bentonville recently (2004) and nearly could not recognize it. The small town and neighboring area is now wall-to-wall stores, restaurants, hotels. Like taking the busy part of Paducah, KY and plopping it on top of Martin, TN (which has a similar population to the Bentonville I once knew). But no matter how big it gets, I will always remember the early years, the "fun years", when we were not the king of the hill yet, but were getting there. As long as Wal-Mart folks remember their roots, there will be good years ahead. To the associates past and present, God bless y'all.

Tuesday, February 5, 2008

My first blog entry

Well, I was visiting another blog (good afternoon, rainramblings!) and ended up setting up a blog! More to come.