In 1965 I was a rookie programmer and systems designer for Radio Shack. Charles Tandy had purchased the failing electronics retailer in 1963 for (to the best of my memory) $300,000, The company should have been already bankrupt, but Tandy had an idea for an aggressive marketing program, and he thought that the electronics business was about to boom. Tandy was buying a business with about 400 stores still operational, a couple of warehouses, and a strong reputation among people who built their own ham radios, TV’s, and the like. He also bought a company with a poor back office operation. Tandy found that he had no real idea what was in the warehouses or in the stores. And he had no idea about the true situation of the hundreds of stores. Some should have been closed, but many were probably well-positioned to flourish under the right home office planning and management. He also had no idea if the back office had any way to implement his new, aggressive ideas. He turned to computers to get him the answers. Radio Shack depended on wired boards to do most of its data processing. It had a 1401, IBM’s still new, widely used, processor for businesses. It had only 4K of memory and Radio Shack’s library of programs was written in RPG, SPS, and Autocoder. All three languages were still being used. There were no disk drives in those days so the system depended on four tape drives which were slow and plagued with “data checks.”
Very few of the existing programs supported Tandy’s marketing ideas. Tandy was a hard-driving man bursting with ideas and his energy level was so high you could see waves of electricity emanating from his head like you would see in the comic books of the day. His office was upstairs and the back office was downstairs in the headquarters located at 2727 W. Seventh in Fort Worth.
When I was hired I was immediately sent to an IBM programming class on Basic Assembler Language (BAL) in Dallas. Many people still today speak of “ALC” as if it were a language—but it is the name of a class, “Assembler Language Coding,” that taught “BAL.” This language was to be the basic language of the new family of computers called the System 360. Tandy had, I was told, the first non-defense 360 to be delivered in Texas. It was supposedly much more powerful than the 1401 and could support Tandy’s ideas. It was essentially a faster 1401. BAL was virtually identical to Autocoder and the Autocoder programs could be run in “1401 emulation” on the 360. The 360 was faster than the 1401 and had 16K memory, with much faster tape drives. I was hired to use Tandy’s 360/30 (one of the smaller models) to answer his questions: What was in the warehouses and what was the retail value of the contents. Likewise for the existing contents of the stores. And to answer the crucial question: could the new computer enable Tandy to implement his ideas.
While we waited for the 360 to arrive I learned all I could about the business from the existing systems. The big day arrived and Tandy was impatient to get started. One day he stopped in our area on his way to lunch and dropped a couple of sheets of paper on the front desk. It described an elaborate mailer for immediate implementation. He wanted us to find all of the customers who had made a purchase of fifty dollars or more in the last sixty days. In addition he wanted us to find the names and addresses of the next-door neighbors as well as the three neighbors directly across the street from the recent customer. The mailing piece was big, and expensive. Depending on the class of merchandise purchased by the recent customer we were to customize the language of the mailer to offer a coupon to each recipient which could be redeemed at the same store visited by the recent customer.
For that day, the program was complicated. Finding names and addresses was especially difficult because the transaction file contained no zip codes,. It took about two weeks to write and test the programs. Tandy was reported to be upset but I never saw any sign of it. I saw a man who was on a mission and he was determined to succeed come hell or high water (as us old timers used to say in the good old days.)
We ran the program finally and it took days to completely process the entire master file, composed of dozens of reels of tape of customer transactions. Then we data processing personnel fanned out all over Fort Worth with reels of tape, boxes of the mailer to be printed, and a simple program to read the selected records from our tapes and print the mailers. Those many companies had purchased computers but had not yet programmed them. As a result they had idle time which they would rent out. Tandy would rent as much time as he could get and computer printers would be spewing out Radio Shack mailing pieces all over Fort Worth for about a week. The mailer was a success and Tandy was ready with still more ideas.
I realized that we had to find a faster way to implement his ideas. The quicker we could find the records he wanted (always a part of each marketing effort) the quicker we could determine (by processing a reel (or a select few) we could tell him how long it would take to run the entire master file and how many “hits” we might find. So, I designed a program that would read punched cards encoded with the logic of his selection criteria. The resulting codes would then be processed by a program I cleverly called “Selector” to find Tandy’s customers. Within the capabilities of the time, this was state of the art.
Not long after this success, the USPS finally sent American businesses a zip code tape which enabled us to put zip codes into our customer records. Big deal, really big deal, and it still is. The program ran very slowly so we put it on the 1401 to let it grind the hours away. One of the operators, working late at night in the quiet of the computer room, said he could hear the 1401 CPU “buzz” so we called it “Buzzer.” So, in 1965, this old man, then a young man, wrote one of the first google-type search programs and we called it “Selector.” It answered a lot of questions for Charles Tandy.
While all this was going on I was trying to find out what was in the three major warehouses and in the stores. It required a lot of manual inventory and marking numbers with a grease pencil on pre-printed inventory cards at each warehouse and store and then shipping the marked cards to headquarters where they would be processed. I also wrote projection programs to forecast sales by store and by regional warehouse. The programs worked, and Tandy was pleased and showed me his appreciation.
While I was with Tandy, he purchased Pier One and Wolfe Nursery both of which produced profits for decades. So, while Radio Shack is gone, almost, Charles Tandy created an operation that made lots of money for stockholders,paid taxes, gave good service to customers, and employed thousands of people for more than fifty years. IBM’s System 360 was the leading computer system in the US and in the world.