Tag Archives: Camex

1993-1996 TSI: AdDept-Burdines Interface

Posted on by
2

The proverbial brass ring? Continue reading →

Even before I became a professional software developer, my friends and acquaintances often approached me with their ideas for computer programs. It started in the Army with Doc Malloy’s idea for a tennis game, continued through graduate school, and was nearly ever-present in my business life. It seemed peculiar to me that so many people seemed to imagine that I had a skill that they lacked but no idea how best to employ it.

I have a great idea for a software project!

In point of fact, the limiting factor in software development was almost always money. A new software system required a substantial investment to cover development and testing costs, as well as marketing expenses. Very seldom did the people who propose these project give any thought to helping to finance them. At least they never volunteered information about having a secret source of funds. They evidently thought that they should share in the imaginary profits because they provided the original idea. I sometimes told them, “ideas are a dime a dozen. Implementation is everything, and marketing brings it home.”

I had plenty of ideas of my own. A few of them, such as the idea of running several simultaneous “threads” for the cost accounting programs generated a bit of revenue for TSI. One of my ideas, the use of a butterfly-shaped website for insertion orders and emails for notifications, resulted in a very profitable product for TSI, AxN. The genesis of its design and the marketing concept that turned it into a financial winner is described here.

TSI’s clients also had a large number of ideas for programming, but they seldom expected us to work pro bono. I spent many hours researching and writing quotes for changes to our systems requested by our clients. I doubt that a month went by in which I failed to produce produce ten or twenty of them. I considered my most important responsibility at TSI to be providing a clear description of each requested project and assigning an appropriate cost figure.

Gilbert’s LinkedIn Photo.

Very seldom did someone approach me with a project that included funding. I can think of three times in forty years. Only one of these concerned software that we had already designed and coded. The person making the proposal was Gilbert Lorenzo1, who was one of the top bosses in the advertising department at Burdines, the Florida division of Federated Department Stores.

Gilbert telephoned TSI’s office in the early nineties. He had received one of our first mailings about AdDept, our administrative system for large retail advertising departments. He said that he would be in New England to meet with some people at Camex2, the company based in Boston that marketed a system for digitally producing page layouts for newspapers and large advertisers. He requested us to show him a demonstration of the AdDept system.

Most of this huge structure served IBM’s business partners.

We reserved some time in a demonstration room at the elegant IBM office in Waltham, MA. We had a relationship with this office, but we had never done an AdDept demo there before. I arrived there as early as I could to get the system set up for my presentation. It was a very nice facility that always impressed potential clients.

The AdDept system in those days was fundamentally sound, but many “bells and whistles” were added on in the subsequent decade. In almost every case they were suggested by and paid for by one AdDept client or another.

The most impressive thing about the demos in the early years was the speed with which the programs moved from one screen to the next. Once the tables were set up, a user could define all aspects of a new ad to run in dozens of papers in just a minute or so. This always generated the biggest “Wow!”

In our discussion after the meeting Gilbert said that he liked what he saw. He might have even said that he wanted to buy the system. However, I did not hear from him again for several years. This was consistent with what always seemed to happen with Federated’s divisions, a phenomenon that is explored in more detail here.

Meanwhile Burdines was—unbeknownst to me—experiencing explosive growth. In 1991 alone the number of stores increased from twenty-seven to fifty-eight through the assimilation of two Federated divisions— the Maas Brothers and Jordan Marsh stores in Florida. More stores were added throughout the rest of the nineties. By the end of the decade Burdines dominated the department store market throughout the entire state.

The purpose of Gilbert’s second contact with TSI was to invite me to Huntsville, AL, the home of a software company with which he was working at the time. I don’t remember the name of the company. I do recall that two of the team assigned to this project formerly worked as software developers at Camex before DuPont purchased the company and changed its focus. One of them was Mike Rafferty3, whom I had met at Camex’s headquarters when our common customers had requested that an interface should be constructed between the two systems, a project that was described here.

The software company in Huntsville had a very impressive headquarters. As I understood it, the company’s primary customers were NASA and companies that worked with NASA. That was de rigueur in Huntsville.

Gilbert explained that he was working with Mike and the others to develop a comprehensive software system for the advertising department at Burdines, and he hoped and expected that the other Federated divisions would also use it. He wanted TSI (or at least me) to participate in the project, and he insisted that he had the funding for it.

Mike described their approach to the project. They intended to use a home-grown database that resided on a server, but most of the programs would reside on the individual “clients”—PC’s or Macs. When I told him that TSI’s programs were written in BASIC, he suggested that we consider converting them to use Microsoft’s Visual Basic.

Most of the discussion concerned the scope of the project. They were interested in integrating something like AdDept into the unitary structure that they envisioned. No one addressed how TSI would be integrated into the development process. Maybe they expected me to fill in some of the details or to volunteer to research how difficult it would be. Maybe they knew that we seldom backed down from a project just because it was difficult.

The atmosphere was cordial and positive. I remember that we all went out to lunch together, and I ordered a Monte Cristo sandwich. Nevertheless, this meeting made me very uncomfortable. On the one hand, the prospect of installing a version of the AdDept system into all the remaining Federated divisions was way beyond tantalizing. It would be a dream come true. What they suggested would undoubtedly a big job, but TSI had a talented group of programmers who were quite familiar with both the subject matter, and the way that I liked to approach big challenges.

On the other hand, from my perspective the way that this project was described was adorned with “red flags”.

  • I had already researched the possibility of using Visual Basic. It might have been possible to convert some of the programs, but there were no tools designed to help. It would certainly have taken TSI several years to produce a workable system. We would be discarding all of the tools that we used in favor or ones that we had never used and, to my knowledge, had never been used by anyone in a data-intensive situation. TSI’s programmers would certainly need a lot of training. We would probably need to hire skilled employees or at least consultants to achieve any degree of efficiency.
  • Their whole architecture was different from what we used. In the AdDept system the data and all the programs resided on the AS/400 server. The “client-server” approach that they proposed located the data on a server, but the program were all distributed to the PC and Mac clients. To me this sounded like an administrative nightmare. All changes—including emergency fixes—must be installed on all of the clients.
  • I considered the AS/400 integral to AdDept’s success, and so did our customers. The operating system code was built on the database rather than the other way around. That meant that the system itself could never be used for programs with the the huge requirements for memory, disk, and processing speed that design and creation of advertising layouts required. The AS/400 was definitely not designed for that. However, it was ideal for administrative systems like AdDept. It competently handled so many problems with which all-purpose operating systems constantly struggled.
  • I trusted IBM and the AS/400’s database. I knew how to get the latter to function efficiently, and IBM’s support was unmatched in the industry. The idea of converting to a home-grown database seemed just preposterous.
  • By the time of the meeting in Huntsville TSI had finally turned the corner. The AdDept product had a solid client base and a good number of prospects outside of Federated. How could we continue to pursue AdDept development for those companies—which was relatively certain to generate revenue and good will—while devoting a great deal of time and attention to the massive Federated project? It did not seem possible to me.
  • Gilbert had said that he had funding, but he never provided any details about who, how, or how much.

Something about the project sounded fishy to me. They were interested in my participation, but they never specified how. Did they want to buy TSI? No one mentioned anything like that. Did they want to hire some or all of us? Did they just want me to consult with them as to the system design? Or was there something else?

At the end of the meeting, Mike asked me what format TSI preferred for exchange of information. Both of the programmers were very surprised when I told them that our offices were connected to all of the clients’ AS/400s via phone lines. We used the AS/400’s built-in messaging and word processing. No one had ever asked us to communicate outside of that.4 I told them that TSI’s employees had PC’s, and the company had a few modems, but we mostly used the PC’s as terminals to the AS/400.

The group did not come to any agreement about how the project was to proceed. I had an impression that they thought that I (who was well into my fifth decade on the planet) was a fossil. I, on the other hand, thought that they, who had dealt almost exclusively with production of ads for newspapers, dramatically underestimated the difficulty of designing a single multi-user database that was capable of handling all aspects of scheduling and managing the financial aspects of all media. The planning and cost accounting modules were even more challenging.

After the meeting I had a little bit of private time with one of the principals of the software company. He asked me what I thought about the project. I told him that it was interesting, but I did not see the ROI (return on investment) for combining the two systems. I remember his exact words. “ROI. Oh, yeah, where’s the ROI?”

I did not hear from any of them again, and I did not press for inclusion in their project. In all honesty I had too many other things demanding my attention. After a year or two I sometimes wondered whether Gilbert had abandoned the project or had gone ahead with it. The answer, it turned out, was somewhere in between. I spent no time searching for information about the project, but little hints turned up occasionally.

Our liaison at Lord & Taylor, Tom Caputo, described to me his experience interviewing for a job in the advertising department at Bloomingdale’s, a Federated division in New York City. He asked the people there about their computer system. They showed him boxes that contained the software for the FedAd system, which Federated had sent them and told them to use. The people at Bloomies had never unsealed the boxes.

When I installed the AdDept system at Macy’s South5 in December of 2005, TSI’s liaison there was Amy Diehl. Her official title was “FedAd Coordinator.” By then I knew that FedAd was the culmination of the project begun by Gilbert Lorenzo more than a decade earlier.

I soon learned that the advertising department at Macy’s South was not actually using the FedAd system at all because the programmers had admitted that it could not handle the department’s planning process. Instead they had been using parts of a previous version called Assets for a few tasks. I was astounded to learn that the Assets system used a Microsoft Access database. They had sent a boy to do a man’s job! Federated Systems Group no longer supported it.

Later we heard that Macy’s East was using the FedAd system, which by then had been given a different name. At the time its advertising department was still using the Loan Room system that TSI had written and implemented for them in the early nineties. That meant that for years the details of every ad were being entered into at least two separate systems.

I even quoted a bizarre request from Macy’s systems people to write an interface between their system and AxN. I provided them with a quote, but nothing came of it.

In all of that time—more than two decades—I never heard anyone say anything good about FedAd. As far as I know it generated a great deal of expense and not a penny of revenue for the company. I only knew of one department that used it. TSI, in contrast, sold and installed thirty-five AdDept systems, each of which was customized to the needs of the individual departments.

On the other hand, I might have been able to carve out a career as the guru for Macy’s concerning administrative software for advertising. That would have certainly been something to crow about. After all, when the game was finally ended, Macy’s had all the marbles.

I doubt that they would have let me—and whatever portion of TSI was involved—participate from Enfield or East Windsor, and I doubt that they would have let us continue to perform or oversee work for their competitors. They might have allowed me to program for the AS/400—I saw several of them at the FSG data center in the Atlanta area. However, it was more likely the Gilbert would have required everyone in the process to use the same database. He seemed to be calling all the shots.

So, I probably would have had to sell my soul to Macy’s. I might have made a lot of money, but I think that I would have been miserable. Almost everyone in my acquaintance who had worked for one of our clients and then worked for Macy’s or a Federated division quit in the first few years and was openly bitter about the experience.

Finally, I must add that I suspect that there was a good possibility that the invitation to Huntsville was just a ruse to get me to expose the totality of the AdDept system to people who might be able to replicate it.

Epilogue: While researching the blog entry for TSI’s relationship with Federated Department Stores (posted here), I discovered that Val Walser’s LinkedIn page prominently features how she “directed development of a sophisticated, integrated software product” in the division run from Seattle. It must be referring to the system that Gilbert and Mike envisioned so many years earlier. I never heard anyone mention any other such system.

1. For some reason Gilbert Lorenzo has two LinkedIn pages. They are available here and here.

2. The Camex system was used by both of the first two AdDept users, Macy’s East, and the P.A. Bergner Co.

3. Mike Rafferty’s LinkedIn page is here. It did not provide much information about him when I discovered it in 2022.

4. Keep in mind that the Internet was in its infancy. At that time Microsoft had not yet completed its domination of the word processing and spreadsheet markets. Technical people used “message boards”, not email, for communication. AOL did not hit the web until 1997.

5. The installation at Macy’s South is described in detail here.

1989-1993 TSI: AdDept-Camex Interface

Posted on by
1

An exciting new feature. Continue reading →

The first time that I ever heard the word “Camex” was when I was researching the requirements for the first installation of TSI’s AdDept system1 at Macy’s East2. One of the job titles that Macy’s used for employees assigned to production of a newspaper ad was “Camex operator”. I asked Alan Spett, a vice president at Macy’s who was our principal contact during this period, what a Camex was. Alan told me that Camex was a company in Boston that had designed and implemented software and networking for workstations from Sun Microsystems. The workstations were used by many newspapers and some of the largest newspaper advertisers to help with the design of pages to appear in newspapers.

I later did a little research on the company. It was founded in 1974 by George White and a partner whose name I never discovered. The first customer was the Boston Globe. Once that installation stabilized, the company grew dramatically by selling expensive systems (roughly $2 million each) to newspapers around the country and then to large retailers.

In 1989 the company was sold to DuPont, the chemical giant. George White stayed with the company until 1993.

Camex had a booth near TSI’s at the Retail Advertising Conference in Chicago that Tom Moran3 and I attended in 1991 or 1992. Our booth was the smallest allowed. Camex’s booth, which was near ours, was perhaps ten times larger. They must have brought twenty salesmen and lots of workstations and network servers.

Alan may have voiced an interest in creating an interface between AdDept and Camex in those early days. However, it was included in neither the original installation nor the first set of enhancements.

Dan Stroman, TSI’s contact for the AdDept installation at P.A. Bergner & Co.4, told me that Bergner’s wanted to implement an interface between AdDept and Burgner’s Camex computers. I told Dan that Macy’s might also be interested in such an interface, and I gave him Alan’s telephone number. The two of them agreed to make a joint project of the interface.

My recollection of the details is fuzzy. I think that AdDept was supposed to create a file that contained all relevant production job information for jobs in specified ad types that had not yet been released. Camex would create files for AdDept that indicated job steps had been completed on those jobs.

There were many issues to resolve.

  • What was the naming convention for jobs on Camex? That field would need to be added to the AdDept database.
  • Should AdDept require that each Camex job name be unique? If not, will uniqueness be enforced at the time of the interface? If not, how will Camex handle two jobs with the same name?
  • Should the source for the interface file from AdDept be the live database or the history records?
  • Are history records for the interface itself necessary?
  • How should AdDept tell Camex about jobs that have been killed?
  • What if the size or shape of the ad had changed?
  • And so on.

There were also technical details about the nature of the interface files. IBM’s PC Support program for the AS/400 could be used to transfer data to and from a PC, but it probably would not work with a Sun workstation. So, a PC would probably be necessary between the Sun Workstation and the AS/400.

My recollection is that two meetings were held at Camex’s headquarters at 75 Kneeland St. in Boston. The first was mostly just to get acquainted and set an agenda. I have no notes from these meetings, and I don’t remember anyone’s name. I seem to remember that Alan may have attended. Sue Comparetto and I drove up to Boston. Camex’s office was very close to Chinatown, and the Camex people treated us to lunch at one of the nearby Chinese restaurants. My other vivid recollection is of my wonder at what Camex had accomplished in a fairly short time.

I am pretty sure that Dan attended the second meeting. Sue and I definitely drove up from Enfield. This meeting was disrupted by a fire alarm. Everyone was asked to abandon the building and stand around in a nearby parking lot. It does not take much for me to be cold, and I was absolutely freezing. I am sure that we were outside for at least an hour. We finally did get to go inside for an hour or two. My recollection is that we made a little progress, but at the end I still did not know what data Camex was planning to send to AdDept. This was because the people with whom we were talking knew very little about the database portion of their system, and the database programmer was not available.

I wrote up a programming quotation for the portion of the interface that AdDept would initiate. I submitted the quote to Dan and Alan. They both approved it. I even started work on it.

Then a very strange thing happened. The person who served as the client liaison at Camex called Dan and told him that Camex had decided not to participate in the interface. They also volunteered to refund deposits that Bergner’s had made for additional equipment.

We got paid for the work that we did—luckily avoiding the bankruptcies of both Macy’s and Bergner’s. Shortly thereafter DuPont split Camex up into pieces and spun them off as separate companies. Camex did not last long after that.

I cannot remember where this happened, but I overheard someone at a large retailer talking with their rep from Camex. The rep said that Camex no longer recommended the Sun workstations. Instead they recommended that advertisers just buy Macintoshes and off-the-shelf software. I was both dumbstruck and disappointed. I had envisioned our relationship with Camex as a possible entrée to many excellent AdDept prospects. Sic transit gloria mundi.

The episode has an epilogue. A few years later I had a meeting in in Mobile AL with some programmers who previously worked at Camex and Gilbert Lorenzo, the advertising director at Burdines department store. That meeting is described here.

1. The design of the AdDept system is described in some detail here.

2. The AdDept installation at Macy’s East is described here.

3. Tom’s time at TSI, including our time at the RAC, is described here.

4. The ups and downs of the AdDept installation at Bergner’s are detailed here.

1986-2005 TSI: Marketing Employees

Posted on by
19

TSI’s salesmen. Continue reading →

By the mid-eighties Sue and I really needed help with marketing. We had some good products to sell, and our service was fantastic. However, our salesmanship was poor. I could often persuade people that I could develop a solution to a difficult problem, but I was not very good at persuading them that TSI’s product and approach were better than those of our competitors.

The first person whom we engaged to represent us was Joe Danko, who lived on Cape Cod. At first the relationship was on a commission-only basis. Later we considered hiring him as our salesman, but we decided against it. The details are described here. Joe was never actually an employee, and we never paid him for his services. I don’t know how much effort, if any, he put in on our behalf.

Trust me; Paul was nothing like this guy.

We hired some consultants to help us. They, in turn, hired a graduate student named Paul Schrenker, to sit in Sue’s office in Rockville when she was on the road. We provided a list of presidents of ad agencies and their phone numbers. In only a few cases was it a direct line, but, even so, quite a few people agreed to talk with him. Ad agency executives were all about relationships. Whether Paul was a potential client or a potential vendor did not matter that much; many agency heads were always on the lookout for connections. So, a surprising number of advertising executives accepted a cold call from a graduate student who knew a lot about biology but very little about any aspect of the business world.

The Patriots debacle was not O&P’s finest hour.

One of the ad agencies, O’Neal & Prelle in Hartford, agreed to an appointment, and we eventually closed the sale. Paul did not participate in closing the sale, but he did make the first appointment.

TSI severed its relationship with the consulting firm. We decided instead to hire a full-time salesman, and we approached it in the same way we had recruited programmers and administrative people—by placing an ad in the newspapers. I think that we interviewed a couple of people. One stood out, Michael Symolon. He seemed excited about the job, and he was quite well-spoken. He was a graduate of Central Connecticut. He had worked in marketing for five years at Triad Systems, a company that specialized in software for dentists.

What about TSI?

I think that we hired Michael at some point in 1987. His LinkedIn page, which can be found here, was no help in this determination. Although he included previous and subsequent employers, he left TSI off of his list of experiences. We paid him a pretty good salary as well as commissions.

I remember that when he first began to work at TSI Michael was gung ho about setting up a nationwide sales organization. He advised me to schedule annual trips to exciting destinations exclusively for the most productive reps of our software systems.

Michael.

This attitude shocked me a little, but he eventually revised his expectations when he discovered how complicated the GrandAd product was. Our competitors could undercut us on price on the hardware, and there was not much that we could do about it. The key to selling was almost always our willingness to customize the system for the prospective client. The idea of setting up a network of sales agents seemed unworkable to me. If I could not deal with the people personally, how could I assess what changes were necessary and feasible?

We gave Michael room to be creative in his approaches, but I was not ready to discuss how to celebrate sales generated by imaginary salesmen.

9.5 rounded up.

Terri Provost left the company shortly after Michael was hired. Michael interviewed and hired Linda Fieldhouse to take her place as administrative assistant/bookkeeper. Both of them are described here. Michael assured me that Linda was “at least a nine and a half.”

I am pretty sure that Michael and I went on a couple of ad agency sales calls together. I remember driving up to Vermont with someone—it probably was Michael. When I got out of the car I realized that I was wearing the pants for my pin-stripe suit with my blue blazer. We did not get the sale, but I don’t think that my fashion faux pas was the cause. Vermont is not known for haute couture.

I also remember that Michael accompanied me to Keiler Advertising once. Evidently he had once dated Shelly, who at that time was in charge of bookkeeping there. Michael was very embarrassed by the incident. I did not ask him for historical details.

I don’t remember him closing sales of any new GrandAd clients.

We took Amtrak from Hartford’s Union Station to NYC.

Michael also came to New York City with me for at least one very important presentation to Macy’s in 1988. He was almost a hero, as is described here.

Michael invited Sue and me to supper one evening at his house in Farmington. We got to meet his wife and kids. It was a very nice house, but I don’t remember any details.

I am sorry to report that Michael was at the center of TSI’s first great crisis, which is described here.

I ran into Michael at Bradley International one day in late 1988. He told me that he was working for a company that sold advertising software to magazines. I told him that Macy’s had finally signed the contract, that I had been working my tail off to get all the software written and installed, and that TSI would send him his commission check as soon as we got the final check from Macy’s. There did not seem to be any hard feelings.

For a couple of years TSI muddled along without a salesman and with very little effort at marketing. Those were very difficult years in a number of ways. By the spring of 1991 the AdDept system had two pretty substantial accounts, and we felt that it was time to start marketing it seriously nationwide.

Meanwhile, our ad agency clients seemed perfectly content with their current hardware and showed no interest in converting to the AS/400, the system that IBM had introduced in 1988. It is described here.

We hired a young man named Tom Moran to help with marketing. He was a very nice guy, but he knew next to nothing about computers, advertising, retail, or, for that matter, marketing. He was definitely eager to learn, and he was willing to follow up on leads, which was the most important thing. Plus, both Sue and I liked him.

I remember going on two trips with Tom. The first was for a meeting with Hecht’s in Arlington, VA. Sue, Tom, and I drove down to the Washington area. A Motel 6 on the Maryland side of DC kept the light on for us, and I am happy to report that no murders were committed (or at least none reported) there that night. It was the first and last time that I stayed at a Motel 6.

The three of us met with Barbara Shane Jackson, who was in charge of Hecht’s patchwork PC system and her boss, the advertising director, whose name I don’t remember. Tom did not contribute much, but it was a good meeting on the whole. In the end we got the Hecht’s account.

The RAC was held at the Hilton in downtown Chicago.

Tom and I also attended RAC, the Retail Advertising Conference, in Chicago. It was a huge pain to get everything prepared for our booth there. We had to rent an AS/400 from IBM and to hire union employees to set everything up. Nevertheless, we did manage to get our demo computer system working by the time that the attendees came to visit the vendor area.

Some vendors who were familiar to us were there. Camex, the company from Boston that specialized in programming and selling heavy-duty Sun workstations for the production of ads, had an exhibit that was ten times as large as ours and had a dozen or more people. Tapscan, the broadcast software company. was right across the aisle from our booth. One young lady who worked there must have accidentally left her skirt at home. It appeared that over her black pantyhose and high heels, she was wearing a wash cloth that she purloined from her hotel room.

Most of the conventioneers were drunk or at least tipsy by the time that they reached our area. We made one contact with the ad director of Hess’s, a department store chain with headquarters in Pennsylvania. Tom gave him a copy of our sales materials and got all of his contact information. Unfortunately, almost as soon as we had begun correspondence with him, Hess’s was acquired by another retailer, and his position was eliminated.

The convention would have been a complete fiasco except for two things. The first was that I got to introduce Tom to the indescribable pleasure of Italian beef sandwiches purchased from street vendors in the Windy City.

The other redeeming event was the appointment that I had made to do a demo at the convention for Val Walser, the Director of the Advertising Business Office at The Bon Marché, a department store chain in the northwest. The programs worked without a hitch, and she was very impressed with what the system could do. She even invited us out to Seattle for a presentation to the relevant parties at the IBM office there.

Tom accompanied me on that trip, too. Our plane landed in Seattle very late, well after midnight. We checked into our hotel, but we only managed to get a couple of hours sleep. We went to the IBM office, where I checked that all of the software was working correctly. By this time I had been chain-drinking coffee for several hours, and still I felt very sleepy. This was an important presentation, and I had to be at my best.

The demo seemed to go pretty well. Everyone was attentive. The people from the IT department were asking tough questions, which usually boded well for us. I was so tired that I could barely concentrate. As we were putting away our materials I realized that I had been drinking decaffeinated coffee all day.

Nevertheless, I convinced Val and the other important parties. We put together a hardware and software proposal, and they submitted a requisition to the IT department, which also approved it. However, the powers that be at Federated Department Stores1, the mother ship, vetoed it.

This episode taught me that TSI needed someone who could navigate his way through the bureaucratic structure to find out what the hold-up was. Tom was not ready for this kind of responsibility. In the end, we decided that we could not afford someone who just tagged along for demos. In fact, we were really in the position where we could not afford anything.

Fortunately, we were able to use the Hecht’s installation as an entrée into the May Company, which at the time had about ten divisions. Not long after that I persuaded Foley’s in Houston to install the system, too. I also convinced Neiman Marcus in Dallas to get the system.

A grainy photo of Doug in an airport.

Those sales gave TSI both a solid base of accounts and enough revenue that we again looked for a marketing person in 1993. We found what we were looking for in Doug Pease, who had actually worked in the advertising department at G. Fox, the local May Company chain.

At first I had hoped that Doug could do some of the demos, but I soon gave up on that idea. I knew exactly what the system did, what it could potentially do, and what was beyond us. The programmers were generating a lot of code every week, and so these lists were in a constant state of flux. Besides, I had a great deal of experience at public speaking, and Doug did not. I don’t think that I would ever have trusted anyone with the demos.

Doug was a real bulldog once he had a hot lead. He was extremely good at following up on everything. In his first year we closed extremely profitable sales to Lord & Taylor, Filene’s Basement, and Michaels Stores.

Susan Sikorski

In April of 1994 I received an email from a woman named Susan Sikorski, who worked at Ross Roy Communications, Inc. in Bloomfield Hills, MI. The company at the time had eight hundred employees (!) and seven satellite offices. They wanted a production billing system that would feed their Software 2000 accounting system and some internally developed applications.

A few years earlier I would have considered this opportunity a godsend. We had already written interfaces for Software 2000 accounting systems for two AdDept clients. We loved to do interfaces, and the more complicated they were the better. However, we were so busy with programming for clients that Doug had landed that this was my response:

Unfortunately, as I looked over your package, I realized that our system does not really measure up to your requirements. We would have to make very substantial modifications to meet even the minimal requirements. Since we specialize in custom programming, this would not ordinarily be a great issue to us, but at this time we would not even be able to schedule the work for many months. So, I guess that we will have to mass.

And it was almost certainly a good thing that I was forced to make that decision. In 1995 Ross Roy Communications was purchased by the mega-agency called Omnicom Group. If TSI had been chosen for the project, I strongly suspect that the plug would have been pulled on it before the system became fully operational. Susan found a new job at Volkswagen in 1996.

Meanwhile, in the next few years Doug managed to get TSI’s AdDept system into all of the remaining May Company divisions, as well as Elder-Beerman, the Bon-Ton, Stage Stores, two Tandy divisions, Gottschalks in California, and all but one of the five divisions of Proffitt’s Inc., which later became Saks Inc..

Doug and I took many sales trips together. The most memorable one was in December of 1997 to Honolulu to pitch Liberty House3, the largest retailer on the islands.

Doug using a client’s AS/400 for something.

We had a little free time while we were there. Doug and I used it to climb to the top of Diamond Head together. He was an enthusiastic mountain biker, he had been a soccer player in college, and he was quite a bit younger than I was. I was in pretty good shape from jogging. So, neither of us held up the other.

Sue accompanied us to Honolulu, and after Doug returned home, she and I had a great time on four different islands, as is described here.

The other trip that was the most memorable for me was when we flew to Fresno, CA, to pitch Gottschalks, a chain of department stores in the central valley.

In those days you could save a lot of money by flying on Saturday rather than Sunday—more than enough to pay for a day’s food and lodging and a car rental. Doug and I considered going to Yosemite on our free day, but there was a problem with the roads there. Instead we decided to drive along the coastal highway from north to south to maximize our views of the coastline.

Somebody else’s photo.

I did not have a camera, but Doug did. His was a real camera of some sort. I was not yet into photography, and I had not brought a disposable camera on the trip. Doug took lots of photos. In fact, he ran out of film. When we stopped for lunch he bought some more film.

Doug took a lot more photos on the rest of the journey, or so he thought. When we got to Fresno he discovered that he had no photos at all after lunch. I don’t remember whether he forgot to load the camera after he took out the film. Maybe he did not wind it, or there was a technical problem. That was not the worst of it. He also somehow lost the first roll of film when we stopped for lunch, and it also contained the photos of his newborn child taken before we left.

But, hey, we got the account.

I guess that Doug is unloading new equipment in Enfield.

Doug and I almost never disagreed about what the company should be doing. However, near the end of his tenure he came up with an idea that I just could not sanction. He wanted us to start a new line of business in which we contracted for large chunks of advertising space from newspapers at a discount and then resold it to small businesses at a profit. Maybe he could have sold a lot of space; maybe he couldn’t. In any case such an undertaking would leverage no TSI products or services and none of the skills that the rest of us possessed. In short, he was asking me to backstop a new source of revenue for him. I declined to do so.

Doug and I made a great team. I gathered specs and did the demos. He attended, met the players, and subsequently followed up on everything. When the prospect had signed the contract, he made sure that all the i’s were dotted and the t’s crossed and ordered the hardware if they bought from TSI or a business partner.4 By 1999 we had more work than the programmers and I could handle. I told him to stop selling new software systems until the programming backlog could be reduced to a more manageable level, which would not be for at least a year. He made the imminently reasonable decision to look for another job.

After TSI moved to East Windsor in late 1999, we hired one more AdDept salesman, Jim Lowe. His previous experience was with a company that marketed hard cider. The challenge was to get retailers to give them adequate shelf space. It was retail experience, but not exactly the kind that we had dealt with.

Jim was a smart guy, and he could have been a good salesman for us. We went on a trip together to Wherehouse Music in Torrance, CA. Wherehouse was a large chain of music stores in California. Jim and I stayed in a nearby Holiday Inn the first night. We used MapQuest to find to the Wherehouse headquarters the next morning. At the very first turn MapQuest advised us to turn right. This seemed wrong to me, and I turned left instead. We reached the building in less than ten minutes. I don’t know when we would have arrived if I had turned right.

It was a very strange meeting. Rusty Hansen, whom I knew from Robinsons-May, had told them about us. We never got to meet with him or anyone else who seemed to know what they wanted. We did get to meet the president of the company, who was wearing jeans and a tee shirt. I never did figure out what this whole episode was about. The company went out of business within a couple of years.

Jim only worked for us for a few months. He took an offer that was very similar to his old job. Before he left he helped me with a mailing that produced some good leads. I sold the last few AdDept systems to some of those retailers by myself.

Jim’s advice to me when he left was that TSI should concentrate on AxN, which is described here. I don’t think that he ever really understood that the horse must precede the cart. We needed retailers to be sending us insertion orders in order to be able to send them to newspapers.

Bob in Denise’s office.

Bob Wroblewski was, as I recall, a relative of Denise’s husband. In November of 2003 Denise came up with the idea of paying Bob to get the newspapers signed up.

I got to know Bob on a trip taken by the two of us to California to persuade Rob-May and Gottschalks to use AxN. We both misjudged how well the two demos went. The people at Gottschalks seemed excited; Rob-May was somewhat cool. However, Rob-May soon came around, and I never did persuade Stephanie at Gottschalks to use AxN.

Here is how the marketing process worked. After a retailer’s advertising department that scheduled its newspaper ads in AdDept agreed to use AxN for insertion orders, it provided us with a list of its newspapers with contact information. I wrote a letter to each paper asking them to subscribe to the service. The letter was printed on the retailer’s letterhead and was signed by the advertising director or ROP manager at the newspaper. However, it was sent by us along with a contract that I had signed. The monthly rate was approximately what the newspaper charged for one column inch in one issue. This was a negligible fraction of what the advertiser spent. Then Bob called each one and persuaded them to sign up.

I don’t know (and I don’t want to know) what Bob said to the papers, but he had a very high success rate. He also earned quite a bit for himself in commissions. At one time we had over four hundred newspapers that subscribed for the service!

Bob’s wife died while he was still working with us. I drove to Providence, which is where he lived, for the wake.

1. Federated Department Stores owned many large chains that were all very promising potential AdDept clients. The rejection of The Bon Marché’s request may have been a blessing in disguise. In January of 1990, shortly after this meeting, Federated filed for Chapter 11 bankruptcy protection. It could have been really ugly.

2. Susan Sikorski is apparently working as a consultant for Avaya in 2021. She is featured as a graduate of Wayne State on this webpage.

3. We learned later that the advertising department at Liberty House had approved the purchase of the AdDept system, but the order was never placed because in March of 1998 Liberty House filed for Chapter 11, and the funds for new systems were frozen.

4. TSI was throughout its existence a certified member of IBM’s Business Partner program. However, because of the size of the company we were bit allowed to sell IBM hardware directly. Instead, we needed to pair up with a “managing Business Partner” who actually could place orders. We dealt extensively with several of these companies—Rich Baran, BPS, Savoir, and Avnet. There may have been others.

1990-1995 TSI: AdDept: The Installation at P.A. Bergner & Co.

Posted on by
4

She knows there’s no success like failure, and failure is no success at all.—Bob Dylan’s “Love Minus Zero/No Limit” Continue reading →

In the early years of the marketing of TSI’s GrandAd system for advertising agencies I distinctly remember being disappointed that the second installation (at Potter Hazlehurst) was even more difficult than the first. This tendency was definitely repeated with the system for retail advertising departments, AdDept (design described here). The second installation, which was began in the spring of 1990, was much more problematic than the first.

I had never heard of P.A. Bergner & Co. I now know that the chain’s original store was in Peoria, IL. After it acquired the Boston Store chain in 1985, the company moved its headquarters to the upper floors of the Boston Store in downtown Milwaukee. In 1989 the company also acquired the Carson Pirie Scott stores. The corporate management of all of these stores, including all aspects of advertising, was handled in Milwaukee.

I was not yet familiar enough with the nature and history of American department stores to understand the importance of the timing of that last acquisition. Bergner’s advertising department was presumably given a budget to acquire a computer system to enable it to handle the absorption of the Carson’s stores. The money probably had to be spent in 1990. Any portion that was not spent vanished. Little or nothing for future years was budgeted. I came to understand these things later. I wish that the situation had been spelled out to me.

Moreover, Sue and I continually tried to portray TSI as a strong stable organization that was especially good at big projects. That was true a few years later, but in 1990 the company consisted of:

  • Me,
  • Sue Comparetto (who basically wrote checks, closed the books, and did presidential stuff),
  • Sandy Sant’Angelo, who could do simple programming, of which there was still quite a lot,
  • Kate Behart, who helped with marketing and served as day-to-day contact with some clients,
  • Denise Bessette, who worked only part time, all of which was devoted to projects for Macy’s East,
  • One or two administrative people.

The only way that the whole Bergner’s project could be implemented on time was for me to do almost all the programming work myself. The closest that I came to admitting this to Bergner’s was when I told them that Sue was terrible about time. Anything with a deadline had to come through me. Dan Stroman was quite taken aback when I said this. I was also pretty shocked when I had discovered this, but there was no sense in denying it.

Maybe I should try to grow a big white mustache.

Several IBM reps were involved in the sale of the AS/400 for the AdDept system to Bergner’s. The one whom I remember best was Sue Mueller. One of our last meetings took place shortly after the trip that Sue Comparetto and I took to England (described here). Sue brought to Milwaukee a copy of the magazine. When Sue Mueller saw my picture on the cover, she literally did not believe that it was a photo of me.

My recollection is that IBM sold Bergner’s a model B30 of the AS/400, the same machine that Macy’s was using. I think that it also had the faxing software and the modem that it required, and it definitely included OfficeVision/400, the word processing software.

Our primary contract person at Bergner’s during the negotiations was Dan Stroman1, who was either the advertising director or the production manager. He arranged for me to meet with the individual managers—production, ROP, direct mail, and business office. Either I did not do a very good job of collecting the requirements from them, or they did not do a good job of describing what they needed.

We tried to emphasize to Dan the importance of having one person designated as the liaison between TSI and the users. At first he wanted the liaison to be a person from the IT department. However, everyone agreed that the person whom the IT department assigned to the project, Kee-Huat Chua, would never have worked for many reasons. So, the job was given to Sheree Marlow Wicklund2, who at the time was the manager of the merchandise loan room.

I don’t think that in 1990 TSI yet owned the tools to produce detailed design documents. If we did, we certainly did not use them for that purpose in this installation. I merely described in the proposal what we planned to do.

The project was focused on quite a few things that required significant design changes.

  • Bergner’s stores had four or five logos (the name on the store’s signs). So, a table for logos had to be created, and separate versions of each ad were needed for each version of an ad.
  • On some pages of some direct mail catalogs and newspaper inserts the merchandise varied by market. That is, one “block” on a page might show items from one department in one market and different ones in others. Bergner’s called these “swing pages”. Providing a way to handle them meant, among other things, that the entire approach to measuring books needed to be rethought.
  • Bergner’s wanted to use the system to “traffic” production jobs. We proposed to do this by defining lists of steps, which we called “production schedules” and “timetables”—date relationships to the release date (the date that the materials were delivered to the printer or the newspaper). This approach worked well for setting up the original schedule, but the traffic coordinators did not like to use it, and I never did figure out how we could make it more useful for them. Sheree was certainly of little help.
  • Macy’s newspaper coordinators ordered their space reservations by phone. Joyce Nelson3, the newspaper manager at Bergner’s, sent a schedule to each paper once per week with all the ads, positioning requests, and other notes for every ad to be run in that paper over the next week. She wanted the system to fax them all. I was very happy to add this feature.
  • Bergner’s employees wanted to be able to record comments in many places. These requests were often difficult to accommodate. I devised a trick to handle some of them. IBM’s office-vision software had the ability to run programs within documents and to display the output sent to an output queue within the document instead! The parameters for the program could even be included in the document. So, a form letter could be created for sending detailed information about an ad to a group of people. The text surrounding the report could be changed every time. I especially liked this approach because I was fairly certain that it could not be replicated on any other system.
I was signed on to our B10 all day every day.

By the time that the software was delivered, TSI had a small AS/400 in its office, and we quickly established peer-to-peer communication with Bergner’s system. This allowed us to sign on to their system from TSI’s office and to send programs directly from our AS/400 to theirs over the phone lines. I worked all day on programs in TSI’s office and installed the ones that I deemed to be working early in the morning. This was feasible because our office was an hour ahead of theirs.

I communicated with Sheree via the AS/400’s messaging system, which supported both plain text and word processing documents. Kate communicated with Sheree by telephone. I expected Sheree to interact with the users and managers personally. Instead they had weekly meetings that all the managers attended. They appear to have been mostly gripe sessions. Sheree took notes and sent me a list of the issues reported. She did not send one document per area; all areas were included on one document.

I liked the idea that we received a written record of the issues, but my responses expressed my exasperation at some of them. Often users were still complaining about things that I was positive had been fixed. In other cases they had changed their minds about decisions that they had previously made and we had implemented. I responded to all of the issues that Sheree had sent in one document. I did not anticipate that Sheree would share all of my responses with all of the managers. I guess that I should have anticipated that possibility, but I had not worked at a large company for seventeen years, and the meetings at Macy’s were usually limited to one area at a time.

At any rate, I ended up on several shit lists because I responded negatively in print to some managers’ comments, and all of the other managers got to read. It never occurred to Sheree that if I had wanted everyone to read my answers, I could easily have sent the document with the responses to each of them.

More than a scribe.

Privately I was outraged that Sheree had violated what I considered to be a special relationship between the developer and the liaison. I expected her to be a strong advocate for the system when working with the users and an equally strong advocate for the users when speaking with us. She evidently considered herself more of a project manager whose responsibility was to organize and keep everyone informed. She did this. However, for such a complex system with such a compressed timeline, this was just not enough.

I talked with Dan about the situation, and I pleaded for a different liaison. He said that it was out of his hands. Ouch.

The installation at Bergner’s should have been a great success story. Some of the areas—notably the newspaper and loan room areas—were working very well in record time. The other areas proceeded much more slowly. I felt extremely frustrated by the incredibly inefficient process upon which Bergner’s insisted.

Bergner’s put the X back in Xmas.

Nevertheless, by the time of my visit in December of 1990 the ill will had dissipated to the point that they invited me to the department’s Christmas party. I could hardly believe it. A live band was playing. There was a mosh pit! The young people who worked in the creative and production areas were going crazy. The old fogeys with whom I dealt were much more subdued.

That evening Joyce Nelson asked me to give her a demonstration of my technique of throwing playing cards. She produced a deck of bridge cards, which are not as easy to throw as poker cards. However, I was sure that I could spin one at least twenty yards. Unfortunately, the ceiling was too low. To get a card to go a long way, its flight must describe a loop. I could not possibly throw a card at high velocity without it crashing into the ceiling or floor after a few yards. I had to demur, and I never got another chance to display my prowess.

Not for jugglers.

I located in our basement a few pages of memos written by me or Sheree about the installation. Kate had written notes in the margins. At a distance of thirty years (!) I could not understand the details of what was being discussed, but it was evident that I was trying to juggle a large number of balls at once. However, the use of the test system and the attitude of some managers made it seem like the balls were all coated with fly paper.

At some point in 1991 Bergner’s agreed that we had met the terms of the contract, which, by the way, was TSI’s simple version with only a few amendments. They made the last payment and also indicated that they also wanted to prepay TSI’s software maintenance for, I think it was, a year. They asked us to send them an invoice for this, but they wanted the description to be something like “Miscellaneous programming projects”.

The proverbial wolf was at TSI’s door again, and we had never turned down a check. So we did what they asked.

Bergner’s also approved quite a few enhancements that TSI had quoted. These projects were not covered by the original contract. So, I was still very busy with writing and installing all of the new code.

We used part of the money to hire Tom Moran to help with marketing. His role at TSI is discussed here. Perhaps I should have hired a programmer instead. I was more confident in my ability to produce an abundance of good code than my skill at closing sales. Also, I could not afford to devote the time necessary to train a programmer.

Bankruptcy: By August of 1991 Bergner’s had not sent TSI any checks in several months. We learned that the company had filed for Chapter 11 bankruptcy. My recollection is that they owed us more than $10,000. We had no experience with this; none of our previous clients had ever done this.

Our first experience.

Dan told me on the phone that Bergner’s would not be allowed to pay any invoice that was received prior to the default date and was still open. I realized immediately that they must have suspected that this was coming. That “miscellaneous” check was done that way so that we would be obliged to provide support for programming projects that they never paid for.

Maybe no letter?

I wrote them a letter that said that we intended to apply the money from the miscellaneous check to the projects that we had delivered but they had not paid for, and we would begin to bill them for software maintenance. The Senior VP of advertising, Ed Carroll3, was furious at me personally. Dan Stroman called me about the situation. He said that I didn’t know who our friends were.4 Well, he might have been a friend, but I don’t think that many people at Bergner’s qualified for that distinction. I told him that our legal advisers (Me, Myself, and I) said that money given on deposit like that could be applied to open invoices. Dan did not respond to that. Instead, he asked if he wanted us to continue as a customer. I told him that I wanted them to become a paying customer.

I don’t remember the details of the rest of this episode. We continued to support the users at Bergner’s. They asked for a few more projects. We felt compelled to bill them at higher rates to try to make up for the money that we had lost.

I learned later that much of the retail advertising community considered Ed Carroll a sleazeball. What his advertising department did with that invoice was probably illegal or at least against company policy. He resented that what I proposed would make him look bad.

Test Environment: I found a letter dated May 13, 1992 that Kee-Huat Chua sent to TSI. It began with these two sentence fragments.

As you are aware of PAB’s intention to set up separate program environments for the AS/400 system. What it entails: one environment for intensive testing by users and one environment for developer/programmer to perform units [sic] test on the program module(s).

It then describes electronic forms that we needed to fill out before we made changes to the system. After we sent the form to him, he would copy the entire production library and the entire data library to the test environment. When this would happen is not designated in the letter. It could not be done if anyone was using the system. So, presumably it would be done at night.

After the libraries were duplicated, we would be allowed to sign on and implement the changes. We were also supposed to test them and send him a memo documenting that we tested them.

I guess that the users would then test them, but the letter does not mention it. It also does not mention how the changes will get into the production environment.

However, it does emphasize (in bold print, capital letters, and underlines), “NO COMPILING OF PROGRAMS ALLOWED DURING OFFICE HOURS UNLESS UNDER SPECIAL CIRCUMSTANCES; when making program modification.” Those special circumstances are not described.

Kee also included a list or reasons why this approach would be good for “TSI/Programmer:”

– To better perform & concentrate on their development.
– Less interruption and delay for development tasks.
– Secure production application: AdDept.

Although I considered this additional requirement a breach of our contract, I did not feel that I could refuse outright. I was sure that no matter what the missing details were, it would slow down our delivery of projects that they had requested by 50 percent or more. It also pretty much excluded them from ever getting new releases of our software.

This time Dan called me. I told him that we could have done it using test libraries from the beginning, but the project would have been hopelessly behind schedule if we had. He replied that he thought that that was what I would say.

I didn’t say this to Dan, but I also knew that this approach guaranteed that we would be doing a lot of extra work for which we would not be compensated. I cannot imagine who would really benefit from this. The alleged benefits for Bergner’s were equally bogus:

– To stabilize the “live”/production system both in the short run and the long run.
– To adhere with PAB’s standard of implementing computer systems established by Management Information Systems
– To encourage more users involvement with the implementation of the application.

I was mostly thankful that they did not come up with the idea of inserting this guy from the IT department between us and their system while we were still working on the base system specified in the contract.

I made lots of mistakes, but I also fixed problems very rapidly.

This was a good example of why it always took IT departments ages to develop and install new software and why so many companies ended up outsourcing the IT department itself. My shoot-from-the-hip approach admittedly required some pain and frustration for a short period of time. However, it had worked extremely well in the ad agency environment, and it would also eventually prove effective in many other retail advertising departments.

If only it were that easy.

Documentation and Ease of Use: Macy’s East’s long and involved contract did not require TSI to provide written documentation of how the programs worked. Although they demanded a commission for any AdDept systems sold in the first year after it was signed, they considered the whole project a custom job.

I no longer possess a copy of our contract with Bergner’s, and so I cannot check on whether it specified anything about documentation or not. I doubt it. Producing a user’s handbook that covered every eventuality of every aspect of the system would have been an overwhelming undertaking that got worse with every program change or new module. Instead, I made a conscientious effort to try to make the system fairly easy to use.

  • The layout of all of the screens was consistent.
    • The top three lines of data entry screens described the purpose. The screen number was in the upper right corner.
    • If a new item was being created, the word “New” was also in the upper right corner, and it blinked.
    • If the item already existed, the user ID, date, and time of the creation and last change were displayed.
    • All binary fields required the entry of Y or N. The 5250 standards used by all of the terminals and PC’s did not support check boxes.
    • The options for all fields that would have had radio buttons on a GUI (graphical user interface)5 were listed on the screen.
    • Two function keys were available at all fields that were verified against the database. F2 produced a detailed description of the field. F4 called a program that allowed the user to list and from the existing entries. Exiting these programs returned the user to the original screen with the cursor position maintained.
    • All dates were entered in the form MMDDYY.
    • All fields were verified when the Enter key was pressed. Any entries that were not valid were highlighted, and the cursor was positioned there.
  • The format for all reports was as consistent as possible. The report number was almost always in the upper right corner.
  • A list of the names of the most important tables and their key fields was provided for querying purposes. The names of the fields was consistent from table to table.

The biggest gripes were the lack of a user handbook and that the system was not “intuitive”. I considered trying to write a handbook, but I could not figure out a way to do it. Furthermore, I could not even start on it until all the programming required in the contract had been completed.

In all honest, I have never seen a system that was intuitive. The text-based user interface, which the only thing IBM provided at the time, probably seemed difficult to people who had never learned the keyboard.

I thought of a few ways to provide better documentation later, and we provided them to both new and existing clients. By then, however, Bergner’s had decided to try to find a different system.

Camex Interface: We were not the only software developers who felt Bergner’s wrath. When the AdDept installation began, Bergner’s produced its ads using Camex software on very expensive workstations from Sun Microsystems. Macy’s East did, too. Both of them wanted Camex to work with us for a two-way interface between the two systems. The story of how this project never got off the ground is recounted here.

Epilogue: The story of the installation has at least four epilogues:

  1. I think that after the bankruptcy brouhaha Bergner’s hired Gary Beberman, who had done a fantastic job as our liaison at Macy’s East, to help them. He called me about it. I told him that they had a good system installed, but we did not get along with some of the people there. I told him about the restrictive process that IT had set up for me to use. He was sympathetic, but he had no advice about how to turn around. He also warned me not to use Bergner’s as a reference account.
  2. Bergner’s advertising department tried unsuccessfully to replace AdDept for years, maybe even decades. I went back there for something some time in (I think) 1992. Joyce told me that they now had two systems, a pretty one and an ugly one. She said that she preferred the ugly one that worked.
  3. Bergner’s emerged from bankruptcy in 1993. They paid us pennies on the dollar. In 1998 the company, then known as Carson’s, was acquired by Proffitt’s Inc., which later became Saks Inc. All of the other Saks Inc. divisions used AdDept. Carson’s was merged with Younkers and Herberger’s, which both had successfully used AdDept. The resulting entity was called the “northern division”. Saks Inc. sold the division to the Bon-Ton in 2005, which closed down all 267 stores in 2018. During that entire disastrous sequence of events Ed Carroll ran the advertising department in Bergner’s. Carson’s, the northern division of Saks Inc., and the entire Bon-Ton.
  4. Some time early in the 2000’s I unexpectedly ran into Ed and Sheree (?!) at Saks Inc. in Birmingham, AL. Steve VeZain, who was in charge of the corporate marketing group, had asked me to come there for some other reason. Sheree was cordial to me, but when Ed saw me he barked, “What are you doing here?” More about this event is documented here. I never saw Dan Stroman again. I have no idea what happened to him.

Miscellaneous: Here are a few things that I remember about the installation:

Lovable Bucky?
  • Dan and his wife had Sue and me over for supper on evening. She was a big University of Wisconsin fan. I remember her saying “How could you not love Bucky Badger?” I thought that Bucky Badger was ridiculous then. I have subsequently learned to hate him. The Stromans’ house was on a large pond. Dan had a boat. I don’t know why, but he revealed to me that he had had a vasectomy.
  • Bergner’s insisted that I stay at the Mark Plaza Hotel when I was in Milwaukee. It had a doorman. He held the door for me even though I always had a free hand. I guess that that was his job, but it annoyed me. I have always hated it when people offered an unsolicited service and then expected a tip.
  • One evening I tried to get some work done in TSI’s office, where I had “passed through” to Bergner’s system. After a while I became too drowsy to be productive. So, I intended to end my remote session. Unfortunately I keyed in PWRDWNSYS on a command line rather than ENDPASTHR. I accidentally shut down their system! I had to call the IT support line in Milwaukee and talk someone through starting it up again from the system console.
  • The German food in Milwaukee was unbeatable.
  • One day Dan Stroman, a guy from production whose name I don’t remember, and I went to lunch at a Thai place. I didn’t think much of it.
  • I had fun jogging down by Lake Michigan in the evenings.

Failure: This entry has made me think about the nature of failure.

For about thirty-five years I designed, wrote, implemented, documented, and supported software systems. In the fifteen-year period that we actively marketed AdDept, we had thirty-five installations. Of all of those, the only one that I would characterize as a failure was the one at P.A. Bergner & Co.

I probably made mistakes. I tried to work on everything at once. I wanted to provide each stakeholder in the department with something to show that we were making progress.

Perhaps I should have insisted on doing one media at a time. Would Bergner’s have accepted that strategy? Who can say? They put a lot of pressure on us to get the entire project delivered. They seemed to have underestimated how much new programming they wanted. I am certain that I never said that we had already written software that was not actually working, but we were definitely trying to be optimistic. We desperately wanted the job. Maybe some statements that we made got misinterpreted.

I often took the “one area at a time” approach in other installations. We usually started with ROP and quickly achieved successful faxing of insertion orders and printing of schedules. That served as a confidence booster. If we had achieved important objectives quickly, perhaps no one would have suggested that wretched test environment. That approach would probably have also eliminated the weekly committee meetings. If my reports had been to one manager at a time, I would have been more careful in my comments.

In some future installations I designed one-time programs to help with the initial loading of tables by reading in user-supplied spreadsheet files. It did not occur to me to try this at Bergner’s, and no one suggested it. I am not sure that they actually had any files that I could have used.

I learned my lesson. In all subsequent AdDept presentations, I emphasized what I called “our dirty little secret”, to whit, the biggest factor in determining the success of an installation was the person who acted as liaison. I also took pains to specify what precisely the liaison’s responsibility and traits should be.

I also recognized that I should never have put in writing anything even slightly derogatory about a third party. Thereafter I always assumed that everyone would read every word that I wrote.

I learned one other lesson from this experience. I needed to investigate prospective clients more closely. From then on I asked prospective clients why they had decided to pursue automation of their advertising department at this point in time.

I can’t be too hard on myself, however. I don’t know of anyone else who managed to automate two advertising departments of large retailers. Over the years I installed thirty-five AdDept systems.

1. I never saw Dan Stroman again. I have no idea what happened to him. I suspect that he must fallen out of Ed’s favor. He may have been blamed for selecting TSI to design and implement the administrative system and for the collapse of the Camex interface project.

2. Sheree (pronounced like Sherry) Wicklund now goes by Sheree Marlow. Her LinkedIn page is here.

Ed
Joyce

3. Joyce Nelson worked for the company until 2004. Her LinkedIn page is here.

4. Ed Carroll, one of the very few people in my life who really hated me, died in 2016. His obituary is here.

5. IBM did not provide OS/400 with a real GUI until more than a decade later, and even then what they suggested was slow, resource-intensive, and lame compared to what people were accustomed to on PC’s and Macs. TSI’s screens maintained the same format throughout the history of the company because we never found a way to convert the hundreds of screens that we had to produce something that was reasonably attractive. The approach that came closest was to write everything over in a language that used CGI for HTML files. That was how AxN worked.