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

Chapter 5
Business Communication
is Different

As 1980 came up out of the future, Telephony asked me if I would do an article for a special issue on business communications. I was glad to do so, because it seemed to me that there was more that should be said about the differences between business and residential service. In its present form, I have added one or two paragraphs to the article as it appeared in the July 14, 1980, Telephony. It does not take much reading between the lines to see that my hopes for the future, started in earlier chapters, are yet to be realized.

***

About 25% of all telephones in the United States are classified as business phones, and of these, only slightly more than half are served by PBX or Centrex systems. Yet these few PBX and Centrex phones are responsible for perhaps half the daytime telephone calls and most of the telephone industry's profits. The profits come from toll calls placed during the day when rates are high; these rates encourage residential callers to wait until after business hours, allowing the business-sized network to handle even more traffic at small incremental cost.

Considering the importance of large-business traffic to the overall economic health of the telephone industry, one would suppose that PBX design, in particular, would reflect the special requirements of business communications; sadly, close examination of many PBX systems shows that this is seldom the case. Most PBXs are more like small central offices and are oriented toward residential service. User requirements for central offices are, in general, far simpler than those for PBXs, and different in kind as well as complexity. Thus the user's freedom of choice is far more limited than the large number of PBXs (with incredibly long feature lists) on the market today would seem to suggest.

But what are the differences between business and residential communications? There are many, but some of the more important are shown in Table 1.

Inter vs. intra-state calling.

Because of its political significance, the first item may well be the most important. Residential calling is primarily intrastate, to friends and neighbors, nearby stores and businesses. Although business calls cover the same area, there will be a much higher percentage of toll calls and calls crossing state lines. Indeed, interstate WATS lines must carry toll calls and cannot handle intrastate calls at all. Thus they could completely bypass the Class 5 CO and enter the toll network directly, greatly reducing the traffic load on local switching. This is one of the several ways in which WATS service is fundamentally different from toll.

Consumer vs. professional management

Item 2, professional communication management in large companies, is another difference from residential service. Often (but not always), a communication manager is far more knowledgeable about his telephone service (and, indeed, about business communications in general) than the telco salesmen who play musical chairs handling his account; one almost never has the same telco sales rep for more than three or four months. Thus a major portion of a communication manager's time must be spent in educating telephone company personnel, making the past history of his needs and problems known to them, and getting them to a point where they can assist him properly.

Unfortunately, this is only the beginning. To combine items 1 and 2 in our table, many large companies operate on an interstate basis. It is quite typical to have a headquarters location in one state, a factory in another, warehouses in additional states, and sales offices all over the place. Thus the corporate communication manager must often work with a large number of telephone companies. It is here that the interstate nature of businesses other than telephony is driven home. Each state has its own tariffs, and each telephone company has different ways of doing things. It is quite typical to be told in one state that a given PBX can only offer certain features, and to be told in the next state that the very same PBX is technically incapable of supporting those features, but does something quite different. Bills are as bad as equipment. Each state has its own format, either for regular paper bills, punched cards, or magnetic tape. A professional communication manager who wants to analyze his toll traffic has to make allowances in his computer programs for this widespread variation.

In short, the fragmentation of the telephone industry, invisible to personnel of telephone companies operating under the tariffs of one state only, is painfully clear to the communication manager of any interstate company. He continually finds himself the only apparent link between different telephone companies, or different branches of the same telephone company. He finds his tie-trunks graded independently from both ends, sometimes with as many as a third of the circuits inaccessible from either end. He finds finger pointing between AT&T personnel and Bell Companies, with both, when possible, pointing at the Independent Telephone Company serving the customer. It is small wonder that advertising about "one big system" is not always met with enthusiasm by professional communication managers.

One must give AT&T credit for attempting to remedy this situation in recent years. Dimension tariffs tend to be uniform in method if not exact dollar values from state to state, and feature standardization is much better than for previous systems. A standard telephone bill is being studied, and standard Station Message Detail Recording tape formats are being debugged. But the system still tends to be the problem, and until it is generally recognized that there is no such thing as an "intrastate" telephone, interstate companies will continue to suffer from telco fragmentation.

Time of busy hour

Turning to item 3, there is really little more to discuss. As has been mentioned, rates are high during business hours, and lower at other times. Between two-breadwinner families and this rate differential, business traffic pays the way and residential toll traffic is moved out of business hours to provide maximum service with minimum investment. Businesses do, however, tend to take advantage of off-hour rates to call west coast locations from the east after 5, or east coast locations from the west before 8. Automatic facsimile and data systems, too, take advantage of low evening and night-time rates. It is hard to get measured WATS costs below the cost of evening toll, but it is even harder to get measured WATS lines busied out after 5 and on weekends. WATS lines, of course, are primarily a business offering.

Single vs. multi-line service

Item 4 is much more interesting. Most residential telephone service is provided by a single pair, although several telephones may well be bridged across it. With only one line and one number, features such as "Call Waiting" can be used without ambiguity. In business, by contrast, multi-line operation is typical. Most PBXs are served by many "trunks," and even small key telephone systems have, by their very nature, several lines, usually in hunt, terminated on them. Finally, although it never ceases to be a shock and surprise to telephone design engineers, most extensions behind PBXs used to be and, in many cases still should be, terminated on key telephone sets. This was notably true with SXS PBXs, but the much higher cost of Centrex and electronic PBXs has forced single line sets to be used at the expense of service to hold down skyrocketing costs.

Note what happens, however. When one has multi-line service, Call Waiting and similar features (such as Automatic Call-back) cannot be used. The whole idea of multi-line service is to complete more than one call at a time in a logical, easy manner. Clearly, there is no way an ESS could apply Call Waiting to an ATB condition on a group of PBX trunks, and there is small need for a PBX to use Call Waiting when the boss and his secretary have several extensions in hunt. Further, with two or more lines available, it is very easy for even an executive to push the hold button to leave one call, select a different line, consult with someone else, and then return to the original call. Or, where permitted, to depress two buttons on his telephone simultaneously to effect a conference call. With all this available and so easy to use for the past 30 years, designers and marketers should begin to understand why so-called "modern" PBX features have few attractions for most business telephone users.

The telephone industry has counterattacked, however. A strong trend in advertising, customer training and other avenues of communication has led to the preposterous idea that everyone should answer all his own calls. Now, one big advantage of key telephone systems is the ability to provide secretarial screening. The secretary can answer a call on the boss's line or on the secretarial back up line, take a message, put the call on hold, use the intercom to announce the call to the boss, and let the boss pick up or reject the call as desired. I have never met anyone who could not do this intuitively, without formal training.*

[* FOOTNOTE: Another preposterous idea is station dial transfer which assumes that the boss's time is less valuable than that of the switchboard attendant even though the attendant, nine times out of ten, has the only up-to-date extension listing in the company.]

But consider the often recommended single line approach, based on simulation of residential service. A secretary cannot screen calls because her phone is on one line and she does not have direct access to calls on the boss's line or any other. If the boss's line is busy, hunting can still deliver calls to the secretary but, under this circumstance, the easiest way for her to announce the call is to hand-carry a note to the boss's desk.

Call Waiting could have let the boss know of a second call but, if the existing call is important, he may not be able to leave it. Certainly his secretary cannot answer it and take a message; she has no way of hearing the Call Waiting tone. The present caller, too, may be troubled by the intrusion of the Call Waiting tone. Finally, if the boss happens to be making an intra-PBX call and both parties hear the Call-Waiting tone, there may be some confusion as to whom the call is directed.

With single line sets, switch-hook flashes followed by increasingly complex manipulations of user dialed "feature codes" to select Call Forwarding, Call Hold, Call Pick-up, etc. (which cause no difficulty to design engineers but require extensive in-hours customer training and retraining programs) are seriously recommended to executives and secretaries alike to solve problems that didn't exist on SXS PBXs equipped with key telephone stations. Progress is clearly in the eye of the beholder.

There are good ways out of the woods, however. Northern Telecom was the first to see them and take appropriate action. The SL-1 telephone set, designed as an integral part of the SL-1 PBX, provides all the convenience of traditional key station equipment while taking full advantage of the stored program capabilities of the PBX control computer and two-pair wiring to phones. Line buttons with suitable illumination can be used for multi-line pick up as with key, causing the desired line to be connected via the switching matrix rather than in the set. Alternatively, the same buttons can be assigned to invoke PBX features. One need not flash the switch-hook and choose one of a dozen or more feature codes; one simply depresses a suitably labeled button.

Western Electric followed suit soon thereafter with Dimension's Electronic Custom Telephone Sets similar to those used on Horizon. Bell Labs had experimented earlier with a three-pair telephone set on a stand-alone key system; once the possibility of using a PBX with universal low-pair-count wiring to multi-button sets came clear, there was no problem making the combination.

This approach is not to be confused with that of systems using a button labeled "hold" to replace the switch-hook flash. The hold button is a good idea, however, because it is difficult for laymen to flash correctly (too short evades the switch control, and too long dumps the call; further, if timing is increased to reduce lost calls, a new call placed shortly after the termination of an earlier one may turn out to be a conference call). Once the "hold" button has called in dial tone, feature codes can be sent. Some telephones have feature codes labeled on the set or a plastic sheet that can be attached around the dial or buttons. Obviously, this helps the caller find the feature codes when required.

The newest trend, however, follows an idea to be found in my Design Background for Telephone Switching. Many years ago, when repertory dialers from independent suppliers first came on the market, it became obvious that there was no need to limit number storage to address signals; feature codes could be stored equally well. But to do the job right, more is actually needed: the ability to insert a switch-hook flash and call progress tone detection at appropriate points in the dialed number sequence. With a switch-hook flash stored in one digit memory, dial tone detection (or time-out to save money at the expense of performance) stored in another, and a feature code stored in additional digit positions, a properly designed repertory dialer can now make PBX features available at the push of a single button without any modification to the PBX or any appreciable training for the station user. Such an approach doesn't do much for multi-line pick up or secretarial screening, but it is an improvement over turning the telephone user into a telegraph operator.

Metallic vs. electronic switching

Item 5 is about switching matrices for connecting lines to each other and to trunks. In the public network, metallic switching is used, while most modern PBXs use electronic switching. (Note that Nos. 1, 2 and 3 ESS are electromechanical switches controlled by a computer complex. In Europe, this approach is called quasi-electronic.) The reason metallic switching is used in the public network is to permit ringing and coin control voltages to be applied through the switching matrix from common equipment, and to permit automatic line insulation testing (ALIT).

Electronic PBXs can and sometimes do use tone ringing or ringing controlled by digital signals; conventional ringing is totally inappropriate in that it cannot be applied through an electronic matrix and must be connected at considerable expense at each line circuit. PBXs have no need to do much in the way of coin control, and ALIT is a feature that PBXs, with their largely protected indoor wiring, do not usually need. However, the differences between metallic and electronic switching have some additional repercussions which will be discussed subsequently.

Long vs. short loops

Item 6, concerning loop length, is of great interest today, now that specifications and standards for digital PBXs and telephone sets are being established. Most operating telephone companies, interconnect vendors, distributors and designers want to take advantage of the standard 500 and 2500 type telephone sets to minimize inventory and training. To allow a $40 item to be the tail that wags the dog in a $1000 per line installation seems to me to be somewhat odd, but there are serious technical difficulties as well.

Note that almost all PBX extensions are within, say, 100 yards of the PBX, while almost all residential and small business telephones are at distances from their switch measured in miles. Further, compensation built into the basic 500 type telephone set to be used at the end of a very wide range of loop lengths (resistance loop design) puts serious demands on line circuits for very short loops.

To be specific, the 500 type telephone's transmission network has varistors that shunt down the transmitter and receiver when high current flows as a result of a short loop. On long loops, where DC current flow approaches the minimum, these elements do NOT reduce the transmit or receive gain. Thus the signal level at the CO is reasonably independent of loop length, and no manual adjustment is required in the telephone set for transmission similar to the bell-tap adjustment required for the ringer.

With 2500 type telephone sets, the DTMF (Touch-Tone) signaling pad also has its output level controlled to some extent by loop current, again minimizing the effects of loop length and permitting telephones to be installed without field adjustment. All this is very good design for residential telephone service but, when the same telephones are used on a PBX (or, for that matter, a station carrier system) where very short loop lengths are almost always encountered between the phone and the battery feed, the problem surfaces.

The standard 400 ohm battery feed, when used on short loops, will allow very high line currents to flow during off-hook conditions. This deals properly with the transmission network and the DTMF pad, but causes heating in the line circuit and a needless waste of power. However, if one limits the flow of current from a solid-state line circuit, the transmission and DTMF signaling arrive at the switching matrix at a very high level which, among other things, can overload codecs in digital PBXs.

In older PBXs with metallic switching, high current on short loops was not a problem for supervisory relays; further, CO trunks cut through to the central office battery on outside calls, adjusting the telephone set's gain automatically. This operation cannot, of course, take place with electronic switching matrices, and leads to new standards which require a PBX to insert several dB of loss on internal calls. Electronic Central Office, however, are required to insert 0 loss on a precisely equivalent connection, providing a blind duplication of metallic specifications at the expense of potential instability.

Because digital PBXs are four-wire internally, and because very short loops impose almost no penalty for two pair wiring as opposed to one pair wiring which is much more reasonable for CO loops measured in miles, a new telephone set seems highly desirable for PBX use. After all, the basic 500 type set was developed prior to transistors (introduced in 1951 or so), and it's time to upgrade. The most obvious place to maximize improvements in transmission with minimum investment is in PBX systems; the very small number of PBX telephones generates disproportionately large toll traffic. Once a four-wire low current telephone set is standardized for PBX use, it can also be applied to station carrier and other small quantity situations where development effort might not be justified. In any event, it should be obvious that present telephone sets that are suitable for residential service have a number of deficiencies when used with a PBX.

Voice and non-voice traffic

Item 7, voice only vs non-voice traffic, is becoming more and more important all the time. Businesses have been using digital transmission for data for years, and have a great deal more traffic that could be digital if facilities were available (facsimile, for instance, could be greatly improved with different transmission media). Residential customers, however, seldom make other than voice calls and, when a home computer hobbyist wants to send data, a modem is quite satisfactory at the present time. For other home communications, notably cable TV and services that might be derived from it, analog transmission appears to be the way the industry will go for many years.

But business communication is different. Time sharing computer terminals, word processing stations, data base displays and many other mechanisms abound already and show every probability of increasing. That these devices should use the PBX system rather than parallel the PBX wiring seems obvious. The Danray PBX (now part of the Northern Telecom product line) first demonstrated simultaneous voice/data via telephone wiring, using the telephone set's signaling pad to set up data connections. The Danray switch uses four-wire analog transmission for voice, but handles data (full duplex at 9.6 kb/s) over the third pair to the telephone set, the power pair, and switches digital data via a separate digital matrix. The NT SL-1 carries this approach a step further. The SL-1 accepts asynchronous digital data up to 9.6 kb/s at the telephone set, carries it to a separate data appearance at the line card using the telephone set's signaling pair, and switches the data via a separate path through the same digital matrix used for voice. One can dial up computers, modems on long-haul analog private lines, or other data terminals. Rather than have this kind of business go to someone else, it seems to me that Northern Telecom is quite right in integrating it with traditional voice traffic in the PBX.

We can expect non-voice traffic to increase precisely in proportion to the channels available to it at reasonable cost. Some PBXs can readily incorporate non-voice communications and, indeed, have a very strong motivation to do so. Whether or not residential Class 5 offices have this need or, due to the much longer loops usually involved, can reach residential customers economically for anything other than voice traffic, remains to be seen. Few residential users care. But non-voice business traffic is here today and is already being inhibited by lack of proper transmission and switching.

Customer needs

Our last item deals with overall customer service. In the United States, we have the best residential telephone service in the world. Direct dialing and detailed toll billing, good transmission and speedy call set-up leave little to be desired. But when we turn to business, we find problem after problem. Not that any other country is any better, but in the United States we have bigger needs. Nobody does business like American businessmen, and business communication should reflect this.

The multi-line short-loop requirement unique to PBXs, combined with the need for detailed cost allocation within multi-line systems, is sufficient reason to take a new look at overall business system design. Further, WATS lines, FX lines, tie-trunks and many other standard business offerings which have no analogy in residential service place additional demands on PBX operation. Non-voice transmission may well dominate in business by the end of the decade unless it continues to be choked off. To say it again, PBX design should actively diverge from patterns established by Class 5 Central Offices.

The 1980s will be the decade of decision. We must recognize the difference in business and residential communication and do what we can to serve our customers, or a variety of other industries will parallel our networks, duplicate our station equipment, and take our business. The digital world of business communications is here now. If we can't live with it, we'll die without it.

[ Top ] [ Next ] [ Table of Contents ]

Copyright 2006 Lee Goeller. All Rights Reserved.