Printer Friendly

Benchmark survey: technical support.

BENCHMARK SURVEY: TECHNICAL SUPPORT Lately, the software industry has begun to take a fresh look at a perennial problem--technical support. Developers have been trimming their warranty periods and rewriting corporate contracts, exploring 900 line service, installing advanced voice routing and database systems, and offering 24-hour and weekend service. Hundreds of developers now sell annual support contracts, and new ideas about support-based service packages are beginning to get serious attention among company strategists.

The driving force behind many of these changes is the fact that tech support more than ever looks like a financial sinkhole for the entire industry. Hiring new technicians and installing extra telephone lines adds new costs without adding new revenues. Worse, improvements in the quality of support often seem to attract additional callers with questions that have little to do with a company's own product line. Developers often report that more than two-thirds of their calls come from people with questions about operating systems, hardware setup, or compatibility problems created by other products. As the software and hardware environment becomes increasingly complex, this particular sinkhole is bound to grow deeper.

Yet at the same time, tech support also looks increasingly attractive as a way to build closer ties with customers. WordPerfect is the classic example of a company that aggressively promoted an open-ended, highly responsive support program as a reason to become a WordPerfect customer. That policy gave WordPerfect a distinct edge in dealing with competitors--especially MicroPro--whose support departments were chronically understaffed and unresponsive. In the long run, support managers argue, their departments may turn out to be a significant profit center after all.

Despite this ongoing debate over support services, however, there has been remarkably little concrete information about industry practices and operating statistics. Typically, research on tech support has focused on the concerns of end-users and mainframe software vendors; the industry's shared experience is largely anecdotal and based on skimpy data.

For this reason, we decided to take a thorough look at technical support operations. Like our other Benchmark surveys, this research report is designed to supply practical guidelines for companies of all sizes, in very diverse markets. In developing the survey, we polled a cross-section of support managers and company executives to get a sense of which topics we needed to survey, then sent a detailed questionnaire to our subscriber list and to our database of micro-based software companies. By the January 30 cutoff date, we had received usable data from 282 companies--a response level that clearly provides a good deal of statistical depth. (As usual, we didn't accept anonymous survey responses, though the actual names of all respondents are confidential.)

In addition to questions about technical support, our survey questionnaire included several demographic questions about company size, development platform, size of installed base, etc. that we used to define important subgroups within the total survey sample. The demographics of our sample are close to the overall industry profile we've derived from other Benchmark studies:

* Company size. 256 of the 282 companies in our survey sample provided us with data about sales levels. Of this group, 42% report annual sales of less than $1 million, 33% have sales in the $1-$5 million range, 8% fall in the $5-$10 million range, and 17% are above $10 million in sales. (For analytical purposes, we treat the two middle groups as a single category.)

* Development platform. Predictably, DOS shows up as the most popular platform among our survey respondents, with 233 companies that develop titles for DOS machines. The Mac follows with 96; Unix, 46; OS/2, 43; Apple II, 39; Atari and Amiga, 20 each. The sample also suggests that multi-platform development strategies have become increasingly widespread; relatively few companies indicate that they support only a single platform.

* Installed base. Because support obligations often extend indefinitely, the size of the installed base has some impact on total support costs. The median installed base for our sample is a relatively modest 12,000 copies; the 50% range extends from 2,000 copies to 80,000. Typically, the installed base represents more than one title. Only 10% of our respondents market a single title; 46% support two to six titles; 28% support seven to 20 titles; 13%, 21 to 100; and 4%, more than 100.

* Flagship product. Most of our tech support survey data measures company-wide support policies and activity levels, not single-title activity. However, we did collect some basic data on each company's "most popular title," which sometimes helped us spot correlations based on product prices or market segments. We find, for example, that the median price for the "most popular" products in our sample is $299 (50% range, $99.95-$795), and the median installed base is 6,000 copies (50% range, 1,275-28,000 copies). The size of the "most popular" installed base suggests that the typical flagship title represents about half of a company's total installed base).

In terms of product categories, the flagship titles in the survey are dominated by business applications and vertical market titles. Ranked by frequency of response, the distribution looks like this:

As we've noted in earlier Benchmark surveys, there are so many possible ways to analyze and present this kind of survey data that we've had to narrow our focus to provide views that will have the broadest application to mainstream companies.

But we also recognize that many readers may want to explore narrower slices of our database. So we've assembled the raw data from our questionnaires (excluding company names) in spreadsheet form. Copies of this spreadsheet can be ordered from Soft.letter for $25 prepaid. Available disk formats include Lotus 1-2-3 (PC 5.25" and 3.5") and Excel (Mac, PC 5.25" and PC 3.5").


We asked respondents how much they spend on "total support costs" as a percentage of their net sales. The median came out about where we expected, at 6% of sales. However, a more interesting point emerges when we compare companies that support DOS, Mac, and Unix platforms. The popular assumption is that Mac software imposes a significantly lower support burden, especially in comparison with a less-friendly environment like Unix. In reality, support overhead doesn't seem to differ much from one platform to another.

We also asked our respondents how they allocated support costs within their internal budget categories. Here, the lack of consistency is dramatic: 30% or our sample allocate support to marketing and sales, 29% allocate it to G&A, 15% to cost of goods, 12% to R&D, and 14% to "other."


Our data shows that East Coast software companies tend to work a slightly later shift than their West Coast counterparts, which is a reasonable way of dealing with a nation-wide customer base:

The survey also shows that relatively few companies currently offer weekend support. Monday-Friday openings averaged 266 responses; Saturday, 35, and Sunday, 24. the median number of holidays closed is nine; the 50% range is 6-10 holidays.


In analyzing our survey data on call volume and other measures of support activity, we broke out responses by company size and also looked at the support overhead for each company's "most popular title."

* Call volume. Small companies (those with sales under $1 million) report that the median number of calls per month they receive for all titles is 60, compared to 400 for mid-sized companies ($1-$10 million) and 4,000 for large companies (over $10 million).

However, software companies often find that call volume can vary considerably, depending on seasonal factors, upgrade releases, etc. to measure these variations in volume, we polled our respondents on activity levels per day during "average, low-volume, and peak-volume periods." (Note that the percentages of completed and abandoned calls often do not add up to 100%, which suggests that many companies are overestimating their performance in answering the phone.)

* Telephone requirements. Tech support groups in small companies generally operate with three or fewer incoming lines and about the same number of outgoing lines. As companies grow bigger, the number of lines increases, at a rate which seems roughly proportional to total sales volume:

* Staffing requirements. Larger companies clearly enjoy some economies of scale in tech support staffing. In small companies, the median number of calls handled per day is four; in large companies, the number is 25:

* Volume per copy. When we look just at each company's flagship or "most popular" product, we get a somewhat different picture that provides useful guidelines for looking at the support requirements of single products. For example, looking at our sample of companies as a whole, we see that the "most popular" product generates a median of 150 calls per month, with a 50% range of 22 to 490 calls. We also find that each sale typically generates one support call, and that two-thirds of calls can be expected during the first 90 days of sale:

* Reasons for calling. Finally, we looked at activity levels in terms of why customers required support. Overall, installation turns out to be the leading source of support requests, followed by product features. However, there are some significant variations between hardware platforms. We broke out two sub-groups--one whose products are primarily DOS-based, and a second whose products are primarily Mac-based products. Mac titles clearly require less support for installation procedures, but we were surprised to see relatively little difference in printer and operating system support requirements:


Almost without exception, software companies base their tech support programs primarily on telephone call-in service. We asked our respondents a group of questions that were designed to explore how they sell and deliver basic phone support services.

* Unlimited support? Despite the proliferation of paid support programs, the majority of companies in our sample (60%) still offer unlimited telephone support. About 17% of companies set some limit to the free support period, most commonly 90 days after purchase or--in some cases--after the first call. A relatively small number of companies (2%) limit support to the current version, while 11% require customers to pay for support.

* Organizational issues. We also looked at how our respondents handled such issues as toll-free lines, registration requirements, special support plans for large customers and resellers, internal databases, upgrade sales, and money-back guarantees. Again, we broke out responses by company size. (The chart shows the number of respondents who marked Yes, No, or Future check boxes.)


Although support is traditionally a people-intensive service, we find that the typical tech support department is relatively small. The median number of employees in all support-related departments is three, with a 50% range for the whole sample of two to eight people:

* Scheduling. Because tech support employees often handle several different tasks, we asked our survey respondents to break out the average number of hours per week that their department allocates to various functions. We also asked respondents to identify how many of these scheduled hours were performed by full-time and part-time employees. (As usual, many respondents seemed to have trouble dealing with statistics on part-time employees, so our confidence level in these benchmarks is lower than for other data in the survey.)

* Programmers. The survey shows that programmers in 67% of small companies handle routine tech support calls, but the percentage drops dramatically in mid-sized companies to 23%. Only one company over $10 million (2%) in our sample says its programmers get involved in routine support.

* Hours on the phone. We asked respondents how many hours their technicians are expected to spend on the phone during an eight-hour shift. The answers suggest that tech support in many small companies is a part-time function (judging from the previous question, programmers often handle some of the support burden). For mid-sized and larger companies, however, support technicians typically spend 5-6 hours per day handling calls.

* Time on the job. Tech support departments are notorious for burnout problems. Yet our survey data suggests that this reputation isn't entirely deserved. We asked respondents whether "most tecnical support employees leave after one year or less, one to two years, or more than two years." Except in very large companies, the most common answer was more than two years:


Our questionnaire explored other support services that many companies currently offer, including bulletin boards, personalized mail responses, newletters, consulting, and specialized printer support.

* Bulleting boards. Tech support bulletin boards are in place in a surprising number of companies, we find. Overall, 34% of our respondents operate a bulletin board; 26% plan to set one up in the future. When we look at larger developers, the percentage rises even higher: 59% of companies over $10 million are already in the bulletin board business:

How active are these bulletin boards? The median number of callins per day for all bulletin baord operators is ten, with a 50% range of 5-35; the median connect time is ten minutes, with a 50% range of 5-20 minutes. However, activity levels vary a great deal from one product category to another (utilities seem to generate particularly high numbers of inquiries, while programming tools and languages involve the longest average connect times).

Finally, we looked at who actually operates these bulletin boards. The answer: 75% of respondents rely on in-house staff, while 25% use a commercial on-line service.

* Mail responses. Especially as fax machines proliferate, tech support departments are beginning to receive a growing volume of inquiries that arrive in written form rather than by telephone. We asked our respondents how many letters, faxes, and disks they receive per week, and whether they encourage or discourage this activity. The median number of written communications is ten items a week for the sample as a whole, with a 50% variation of 5-30 items; these levels do not vary much by product category:

About 9% of our respondents say they discourage written inquiries, 20% say they prefer questions in writing, and a large majority--70%--have no policy on this issue.

* Newsletters. A fairly large number of support departments seem to publish their own newsletters; in fact, newsletters are about as common as bulletin boards:

The typical support newsletter is free to registered users (75% free, 12% by paid subscription, 13% "other"). Publication frequency is usually 4-6 issues per year (predictably, large companies tend to get out more issues than small companies).

* Consulting services. We asked our respondents if they provide consulting for "systems analysis, application design, or other special services." Overall, almost half--including a good many larger companies--say they do offer such services.

The median rates for consulting services tend to reflect company size, though the overall rate is $600 per day or $60 per hour:

* Printer support. Questions about printers and other output devices account for a large number of all tech support inquiries, and many of these questions require fairly advanced technical assistance. Very few companies (2% overall) have set up dedicated printer support groups to deal with such problems, but we find that about 17% of our respondents are willing to write customized drivers to help users work with unusual printers.


Although paid support plans are still a fairly new phenomenon to many software companies, we were able to identify a few basic trends in this area. Our survey data shows that 36% of our respondents currently sell paid support contracts, and another 10% plan to offer such plans in the future:

* Pricing. The line between "individual" and "corporate" support contracts isn't always clear, so pricing comparisons are hard to make. Nevertheless, customers seem willing to pay relatively high rates: The median annual contract fee is $245, with a 50% range of $99-$595.

* Contract components. We also find a good deal of diversity in the services that are provided as part of these contracts. Of the 99 survey respondents with paid support plans, 89 provide unlimited telephone support; 46 of these also offer a toll-free number. In addition, 56 give their contract customers a discount on upgrades, 53 send a newsletter, and 33 offer a bulletin board.

* Sign-up rates. Paid support contracts seem to be relatively attractive to users. On the average, 37% of new customers sign up for paid contracts, along with 34% of older customers. Information about first-year renewal rates for these contracts is still a bit sketchy (only 57 respondents were able to provide data), but the average is about 64%.

* Profitability. Paid support implies a somewhat higher standard of service and performance--and higher costs--than free support.

Thus, almost half (48%) of the companies with paid support report that their programs are "marginal or breakeven," and another 19% say the effort has been "unprofitable, but less costly than free support." Nevertheless, paid support can make money: 33% say paid support is "a significant profit center" for their companies. And the downside risk is apparently minimal--no respondent in the survey reports that paid support is "more costly" to the company than free support.
COPYRIGHT 1990 Soft-letter
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:survey of software companies' technical support policies
Author:Tarter, Jeffrey
Date:Feb 18, 1990
Previous Article:Postscript.
Next Article:Multimedia: what's the compelling application?

Related Articles
Benchmark survey: manufacturing operations.
Benchmark survey: registration card procedures.
Benchmark survey: customer service.
The right way to track customer satisfaction.
Benchmark report: technical support cost ratios.
Data points: support staffing ratios.
Benchmark Survey: tech support cost ratios.
Benchmarks: Revenue and employment ratios.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters