Voice Communication in Business Volume 1
Essays on telecommunications, 1969-1980

Chapter 6
Planning & Installing a Business Communication System
A Case Study

When I left RCA to become an independent consultant, I was a little nervous about the whole thing. It is hard at the age of 49 to give up a regular paycheck and strike out on your own. But it was something I had always wanted to do and, when presented with the opportunity, I knew it was then or never.

At RCA, I had worked with some marvelously talented people. Indeed, it is from them and not from Bell Labs that I learned most of what I know about business communications. But, as at Bell Labs and elsewhere, I had been a member of a team; there was always somebody I could turn to for help, and I was responsible for only a limited number of activities.

One of my first consulting jobs was to assist a client with the selection and installation of a PBX. When the whole responsibility fell on me, I learned very quickly a number of things that even my years at RCA had not prepared me for. When you have to do it all, you learn or die. Learning experiences are, by and large, the better of the two alternatives.

A year or so later, a friend of mine was chairing a session at the IEEE's International Conference on Communications (1976, in Philadelphia). He needed speakers, and the time seemed right for me to document my adventures. Later, Business Communications Review reprinted the paper in the Nov./Dec., 1977, issue.

***

A telephone system for a moderate sized business organization is not inexpensive. Equipment will cost about $20 per month per extension, and tolls and message units will run another $25. The salaries of switchboard attendants add roughly $5 to $8 although Centrex can reduce this somewhat. There are other costs: charges for equipment moves reflecting personnel changes and office layout alterations, floor space costs, and management overhead for verifying the billing and placing orders with the telephone company. Thus it is not unusual for a 150-line PBX to generate expenses exceeding $100,000 a year.

In addition to expense, telephone systems generate a certain amount of anxiety for their customers. The average business executive charged with the responsibility for communications in his company usually has many other duties of more pressing importance than monitoring telephone system "state of the art" or the complex infighting that characterizes the "interconnect revolution."

To add to the problem, the old, well-understood Step-by-Step (SXS) PBXs with their cord-board attendant positions are being tariffed into obsolescence while the newer equipment that is being offered by telco and interconnect companies alike has two characteristics of great importance to the user: it is more expensive than SXS and is incapable of rendering certain services which users have come to expect.

On the other hand, these new systems add many features felt to be desirable by design engineers. By and large, the features make possible the replacement of cord-boards with consoles. A console, sleek and modern in appearance, is a very potent sales tool. Its one additional advantage—relieving the attendant of the need to "pull down the cords" when the call is over—might otherwise be more than counterbalanced by the larger size of the switching matrix required plus the need to invent features to replace moving cords to transfer calls, to override busy at the multiple, and to access specific trunks and tie lines for testing.

There are some very new systems, just now coming on the market, that appear to offer the user some real advantages. However, in the time frame of this case study, shown in detail in Table 1, they were not available. Because the company was moving from one location to another, we could not wait; we had to choose the best solution at hand, and obtain in the process a system that would meet the company's needs for many years to come.

Initial planning

A direct approach to market research is essential: start by reading bills. What a customer says or thinks he wants is important, but what he is actually paying for—how he "votes with his bucks"—is vital. My preliminary study in the case to be discussed showed the following: 171 extension lines were in use, and 16 of them served single-line instruments; the remaining 155 were terminated in 6-button key telephone sets. On a dollar basis, the PBX matrix, switchboard, and CO trunks came to a little more than $1200 per month. Station equipment had a basic cost of $333 which, if all instruments had been single line, would have been the full cost. However, adding the key features cost an additional $1400. Although these numbers may come as a surprise to many, they are quite typical. They show that most of the features needed by business are provided by key telephone equipment, and other PBX features, although interesting, may not benefit the customer as much as new approaches to key system features.

Analysis of the toll bill showed nearby bedroom communities receiving the most calls; beyond this region, traffic was widely scattered and no one area code received more than 6% of the total. Message units ran up a charge slightly more than one-third the size of the toll bill. There is no simple way to analyze message units.

Armed with this preliminary information, a day was spent going through the offices and talking to people who actually used the phones. A number of surprising things came to light. For instance, to transfer a call, the user would put a line on hold, select another line on his key telephone set, dial 0, and ask the switchboard attendant to transfer the call on the first line. It turned out that the cord lamps were very faint, and "flashing recall" was not available on the ancient 552 switchboard. Thus if the attendant didn't see the cord lamp flash when the user depressed the switch-hook, she would miss it altogether.

The switchboard attendants handled incoming calls, placed outgoing toll calls, and took messages for people who were out of the office (colored pins were used to mark the line jacks of people who wished their calls to be intercepted at the switchboard). Completing incoming calls was tricky because, although one person might have three or four different lines, these lines were not always numbered consecutively. Consecutive numbering, of course, puts related extensions on adjacent jacks in the switchboard, greatly simplifying busy-testing. Here, however, the attendants had to memorize scattered jack positions. It turned out that this randomizing was deliberate. Because of maintenance problems, users had developed the scheme of having numbers in different parts of the automatic switching equipment so that they wouldn't be isolated when one part broke down.

Even so, there were many consecutive-number groups. I was quite surprised to find that "line hunting," perhaps the most useful PBX feature, was not employed. In a hunt group, if one line is found to be busy, the system completes to the next idle line. In SXS equipment, consecutive numbering within a given tens-group is necessary for hunting to be implemented, but it is of great value in reducing "busies." The availability of hunting had not been made known to the customer.

The boss-secretary telephone arrangement, one of the main applications of line hunting, turned out to be rare at this location. Usually, the boss and his secretary are given consecutive numbers, and when the boss's line is busy, the system completes to the secretary's line. If the boss's line is idle and the secretary's line is rung, she knows the call is for her. Here, however, the more general case turned out to be one principal with several assistants. The assistants would answer incoming calls and only if they needed help would the principal be notified, via an intercom line, to pick up. The important factor was that a group needed not two lines, but three or four. With two lines, the modern feature group consisting of station dial transfer, consultation hold, and three-way conference can sometimes replace key equipment. But with three or more lines, key telephones are required.

One particular sales group was pointed out. Here four or five secretaries screened calls for from eight to ten sales people. Because the sales force moved from office to office almost at random, but still wanted the same secretary to pick up for them, a continuous and expensive sequence of rewiring the key telephones was always in progress. The situation called for uniformly wired call directors for the secretaries.

I wrote up a brief report detailing my findings and noting what the telco could offer. Although the client had explained to me he did not want Centrex and did not want to risk going with an interconnect company, the various systems available were compared, and open options were indicated.

Specifying the system

Had an interconnected system been considered, it would have been necessary to prepare a full-scale specification and get bids from competing companies. This can be very time consuming and may not elicit usable results. An earlier attempt at a PBX specification for another company, ultimately published in article form (Reference 1), had led to nothing because manufacturer response was unsatisfactory. Some necessary features could only be supplied at very high cost, while others could not be supplied at all.

To understand this state of affairs, one has only to look over the professional literature. Reference 2 is particularly interesting in this respect; it contains a large figure showing the sources of requirements, ideas and concepts in switching development. These are: operating companies, labs, industry coordination, and blue sky. Even though the author mentions that subscribers "will continue to force high quality service and sophistication in features," the absence of user suggestions as a source for user requirements is painfully evident. Table 2 contrasts needed user features with those currently offered in most PBXs. The table is intended to be suggestive, not all-inclusive.

Table 2
CONTRAST BETWEEN OFFERED AND NEEDED PBX FEATURES

Offered Features of Limited Value

1. Touch-tone signaling, color phones, consoles

2. "Series 300" features.

3. Abbreviated dialing

4. Call forwarding.

Offered Features of Considerable Utility

1. Trunk answer from any station (in small PBXs).

2. Call pick-up (when available)

3. Limited class-marking and call restriction

4. Limited line-hunting

Needed Features

1. Call recording for toll and message unit cost allocation, by extension.

2. Traffic recording, for user's administrative purposes

3. Internal and external trouble routining and reporting to user

4. Extensive class-marking and call restriction

5. Automatic alternate routing to minimize toll costs.

6. Reasonable tie-trunk switching, with through-switching on a four-wire basis

7. Use of system control to eliminate high initial and monthly key system costs, and to simplify moves and changes of station sets.

8. Elaborate line hunting (circular and terminal, hunt on busy and no answer, pilot address plus different individual line address, etc )

In any event, because the intersection of the set of telco offerings and the set of solutions acceptable to the customer was very small, there was no need for a complete specification. But, as time was very short, it was emphasized that a proposal should be requested from the telephone company at the earliest possible. date. About four and a half months later, a response was obtained. Over the customer's objections, the telco recommended Centrex, but also suggested that a PBX might possibly be used. The customer requested the PBX. Nothing happened for a month or so, and I was contacted again. A meeting had been scheduled with the telco, and the customer wanted me to be present.

The meeting took place on May 6. A lot of time was spent explaining just why Centrex was not desired. For the record, the reasons are as follows:

1. Centrex is appreciably more expensive than a PBX. A recent analysis of a somewhat larger system had shown the bill to the same telephone company would double in going from PBX to Centrex, although negligible savings in floor space and some small savings in switchboard-attendant salaries reduced slightly the differential.

2. Centrex is not suitable for a location where 90% of the incoming calls are completed by name rather than extension number. The telco insisted that a company was a good candidate for Centrex if calls were completed by extension number or name. Other telephone companies dispute this.

3. Centrex is not suitable where there is a highly mobile work force. On an average day, when a third of the staff is on external assignments, the ring-no answer problem is not inconsiderable. In a highly competitive business, a phone not answered in two or three rings may be enough to divert business to a competitor. And a phone not answered on five or six rings can have a detrimental effect on everyone within hearing. At this location, call screening and message recording by the switchboard attendants were used effectively, particularly with regard to the "principal + assistants" operating pattern without secretarial screening.

The telco sales manager listened carefully and stated clearly his preference for Centrex, although he made a sincere effort to please the customer. He admitted that a 770A PBX could be made available. The architect, responsible for designing the new office layout, immediately asked for size, weight, air-conditioning and power information so that he could complete his drawings. Preliminary information on the 770 from my files was available. The telco would still have to provide official data, but we could at least get started. The cutover was less than four months away.

Ultimately, we specified a 770A PBX, equipped for 200 lines. Because there were occasions when people worked nights, the trunk-answer-from-any-station feature was considered desirable; to get it, we also had to take station dial transfer, consultation hold and three-way conference, the "Series 300" features, as part of the package. It turned out that station dial transfer was mutually exclusive with attendant transfer; however, it seemed to me that, in view of the awkward method of transfer already in use, station dial transfer would not encounter too much resistance. This ultimately turned out to be correct, but for some weeks after cutover, there was a certain amount of discontent.

Still another choice had to be made concerning the feature group. We could have it on incoming calls, or on both incoming and outgoing calls. Since we weren't eager to have it at all, we elected not to have it on outgoing calls. This saved a certain amount of money and kept us from having to have switched loop consoles; both were gains. However, we ultimately got the feature group on outgoing toll calls without additional expense. The "dial 9" trunks were toll diverted at the central office; thus, toll calls had to be placed through the attendants. Unfortunately, the 770 cannot advance calls directly. Rather, the attendant has to obtain the calling extension number, ask the user to hang up, and then call him back to connect him to an attendant trunk (usually used for incoming calls) since attendant trunks are not toll diverted. Because outgoing toll calls went via incoming trunks, they had the feature group available to them. The features are seldom used, and the call-back procedure remains very unpopular.

For PBX Consoles, we specified Direct Station Selection (DSS) so that the attendants would be able to complete calls quickly, and also so that they could mark the extensions of users away from their desks for prolonged periods. DSS gives a good station-busy display, and plastic rings rather than the colored plugs used on the manual switchboard make marking easy.

With most of the required features supplied by key systems, station design is a very important part of the specification procedure. It is also very time consuming, so an attempt was made to have the telephone company handle this part of the job. Because time was getting short, however, this turned out to be impractical. The client and I worked together to generate the specification. The job took several days; first, a floor plan of the new office space was obtained and the client's secretary tracked down who was to sit at each desk. Using the old location's telephone directory as a guide, I worked out a tentative numbering plan and station design, and checked each station with the client who, when there was a point in question, verified arrangements with each department. We used single line instruments where we could, non-illuminated key sets where only two lines terminated on a phone, and illuminated key sets where three or more lines or other factors made visual discrimination necessary. The one sales group mentioned earlier was provided with the new 20 button call directors at secretarial positions and a dial intercom to tie together the fourteen phones in the group. Only the highest executives got colored phones with Touch-Tone signaling. We ended up with less than 150 extensions at the new location, 31 of which were single line instruments and 17 of which were non-illuminated (and, thus, less expensive) six-button key sets. Conveying this information to the telephone company consumed approximately four hours.

Physical problems

From the May 6th meeting on, we had to deal with physical problems. The 770A is a modern crossbar PBX. As such, it turned out to be more than twice the size of the SXS switch it replaced. More important, it was much heavier. Floor loading for the 770 is between 175 and 180 pounds per square foot, and special approval was required from the building's structural engineers.

We wanted to put the console attendants in the same room with the switching equipment. The equipment was enclosed in modern, attractive cabinets, and it was quiet, generating less noise than a typewriter during its busiest time of operation. Heat load might have been a problem, but additional ducting for air conditioning could have been added before the walls and ceilings were closed. The telco, however, could not agree. It was felt that the room was too small to permit maintenance to be performed on the switching equipment if the attendants were in the same room. As it turned out, cutover would have been very difficult if the attendants had been located in the same room with the switch during the first few days, but after the first frenzy of de-bugging was over, the plan would probably have been satisfactory. In any event, a separate room was provided for the attendants, and the overall increase in floor space over that needed at the old location was about 20%. It would have been more, but a rather different and more space-efficient form of main distributing frame was used. Centrex could have saved some space, but the consoles, cross-connect frame for station wiring and the racks for key telephone equipment would still have been needed. Thus the space required with Centrex might have been 20% less than that needed at the old location, rather than 20% more.

Conduit for the station wiring was also a problem. We requested that all telephone locations be wired with 25 pair cable to permit the use of key sets at any point; the only exception was the sales area where 50 pair cables were needed for the call directors. Unfortunately, the conduit was already in and the walls closed before we were able to make the request. Thus we had to make a change at some expense. The architect, just to be safe, specified conduit for 50 pair cables at certain other points where, at some future time, call directors might be needed. Putting it in before the walls were closed made only a negligible increase in cost.

Placing the order

Actually placing the order was not easy. The telephone company wanted the whole order complete from us before they placed their own order on Western Electric for the PBX and station apparatus. However, they indicated that nothing could be considered until they completed a traffic study, and the traffic study could only be made after a switching vehicle had been agreed upon. The May 6 meeting plus several later discussions settled that point, and June 5 was scheduled for the traffic study. Unfortunately, this was a day when most of the work force would be out of the office; we had the choice of delaying the study and missing the move, which would have been disastrous for the client, or taking the study and getting the wrong quantities of circuits. We chose the latter option.

After a week or more of "analyzing" the data, the telco gave us recommendations for CO trunks, intraswitch trunks, etc. All quantities appeared to be too low when compared with the existing system. We increased them where necessary, going in with 13 dial-9 toll diverted outgoing trunks, and 16 attendant trunks for all incoming calls and outgoing toll calls. It turned out that we needed four more dial 9's, based on observations after cutover, but there were only three more ports on the trunk side of the matrix. We changed the order to three, and it will probably be enough, but one cannot help wondering about the need for the five week delay to get the traffic study.

After the study, the next step was to place the PBX order. The telephone company then called a problem to our attention. Under the new two-tier pricing arrangement, the contract with the customer could not be prepared until an accurate cost figure was available, and costing could not be carried out until the contract was signed. This "Catch 22" threatened to stop the whole project, since the order on Western Electric could not be placed until the customer order was signed. Ultimately, however, the telco found its way through the maze when it was pointed out to them that they, themselves, had constructed the tariff, and thus it was up to them to find the solution.

During all this debate, it was suggested that they could at least be pulling cable for station sets at the new location. The telco, however, was adamant; nothing was to proceed until the complete order, including station apparatus, was in hand. It was at this point that the client and I worked out the station equipment.

Implementation

All of the information was made available to the telco on June 14, and the PBX was ordered that day. I also spent part of the 16th on further details of the station equipment order, and everything was given in writing by the 18th.

The installation people from the telephone company went to work as soon as the order was in, and in the best traditions of the telephone industry, they did everything in their power to make sure the date was met. Some even delayed their vacations. And they made the cutover date; the weekend of August 23-24 was a little hectic, but Monday, August 25, the system worked. From order to cutover between June 14 and August 25 may not be a record, but it is close. Part of the credit must also be given to Western Electric at Denver; their approach to factory wiring and testing of the entire system, checking it against the order, minimized hand-wiring on-site. For example, even the hunt groups were wired in Denver.

There were a few problems. The switchboard attendants had no "live" practice with the console, since there was no training facility available for this type of equipment. The telco service advisor showed them the console on Friday, August 22, and read the instruction manual since she, herself, wasn't familiar with the 770. The attendants who had been working cord-boards for years were justifiably nervous, but they learned quite quickly as the calls started coming in on Monday morning.

The station dial transfer feature was something of a problem. It took a while for us to discover that a person transferring a call had to stay with the connection until the party receiving the transfer answered. Otherwise, the call would be lost. It seemed reasonable that, if a console attendant could leave the call immediately, a station user should be able to do the same. However, the design problem was how to handle a no-answer in such a situation. The designers intended for the user to flash the switch-hook to terminate ringing and return to the original party, but we only found this out after a certain amount of frustration and the correction of one open relay coil in a transfer circuit.

It was necessary for us to show the secretaries in the sales area how the dial intercom worked on their call directors, and other details required attention. Since many people had decided to sit at desks different from the ones indicated on our floor plan of a few weeks earlier, a number of instant directory changes had to be made. However, all these problems were minor.

Some comments on tie-lines

There are a few tie-lines running from this PBX to another in a nearby city. On the old PBX, they had been used only for "inside" traffic. It seemed reasonable to consider using them for off-net calls into the two cities, thus saving a certain amount of toll.

When using tie-lines between two PBXs for off-net calling, it is desirable to provide separate CO trunks for tie-line use. There are two reasons for this. First, it keeps the billing separate; the local management does not have to pay for the message units (and tolls) incurred by the people calling from the distant PBX. And second, different restriction patterns are usually needed. In our case, we could have shared the dial 9 trunks with the far PBX because dial 9 was toll diverted. At the other end, however, the dial 9 trunks were totally open; control was exercised by restricting most of the extensions. Only the select few had access to the outside world, but they could dial as far as they liked.

After some investigation, we decided to give the project up. There was no way we could separate the message unit billing at our end and provide different restriction patterns for local users and tie-lines at the other end. With an old-fashioned SXS PBX such as the 701, there would have been no problem on either count.

The final act

On November 20, 1975, we finally received a new Service Charge Detail, telling us the monthly cost of the new PBX. There was a reason for the delay. Just before we cut the PBX over, there had been a massive tariff change, and the billing was in doubt. I checked the equipment against the bill and found numerous small errors. But, most important, charging was based on rules in the old tariff, just as if the change had occurred after cutover rather than before. The principal item was the cost of key telephone systems. We had rejected a feature called "wink hold," partially because it makes it difficult to tell a ringing line from a held line when both are simultaneously present, and partially because it adds a dollar per line per month for small benefit. Under the new tariff's "package pricing," wink hold was required at the higher cost. The telco, upon reviewing the bill later, indicated that we should be charged for wink hold since our equipment went in after the tariff changed. Our response was that since we did not have wink hold, we would not pay for it, and if it was added, we would go to the PUC. As of now, the problem is not resolved.

Another point of interest. Under the two-tier pricing arrangement, we have a portion of the PBX itself protected against tariff increases by paying a certain amount of money "up front." The total bill for equipment is still on the order of $3000 per month, but the protected part comes to just a little over $500. Any protection against rising costs is welcome, but the supposed effectiveness of two-tier pricing should not be over-emphasized.

Conclusions

Planning and installing a business communication system in the real world is a challenging and interesting operation. It differs totally from the concerns of the system designer. Needless to say, my friends who are still in the design business do not take my experiences very seriously. They are wrapped up in a world of LSI, stored program control, "state of the art," and other ephemera. It would be very useful if this continuous technological innovation were balanced by direct exposure to the customer who both uses and pays for the end product.

References

1. Voice Communication in Business—Chapt. 3.

2 . F. D. Reese, "A USITA 'Look Ahead' at Switching Technology," Communications Society, Vol. 12, No. 5, Sept. 1974, p. 8.

[ Top ] [ Next ] [ Table of Contents ]

Copyright 2006 Lee Goeller. All Rights Reserved.