THE PROBLEM EXISTS, a paper by and © copywrited to Steven M. Kolins
Email me
Systems Technician, Alamance-Burlington Schools, 2001+
Permission to use, and copy, provided that the above copyright notice appear in all copies and that both that copyright notice and this permission notice appear unless otherwise specifically approved in writing by me. All other rights reserved. Copywrites of references are held by their owners.

The Promise and the Challenge

"Nobody really needs convincing these days that the computer is an innovation of more than ordinary magnitude, a one-in- several-centuries innovation and not a one-in-a-century innovation or a one-in-ten- years innovation or one of those instant revolutions that are announced every day in the papers or on television. It is an event of major magnitude. "-- Herbert Simon, in an address to a research conference on 'Computers in Education: Realizing the Potential', August 1983.

"...the very real support service crisis emerging... Rapid advances in hardware technology fuel new opportunities in applications, raise unrealistic expectations, and create unstable and unmanageable technological environments. Support services are expected to deliver everything that's new, virtually instantly; continue support of legacy systems beyond their reasonable lifetimes; assure interoperability of disparate and sometimes incompatible applications; and do it all with resources that are rapidly dwindling in proportion to the work at hand."(1)

Introduction

That the problem of supporting the introduction of computers in the K12 environment has not been defined is perversely documentable by it's absence in self-proclaimed plans for the use of technology. In the 1998-99 North Carolina state technology plan for the public schools there was almost no mention of any need of techs support. Things have progressed some in that the 1999-2000 report which addresses these issues is in an appendix which points towards a technology facilitator and technician at every school but "new funding would be needed"( see NC DPI.) The NC State's ABC plan however hasn't changed much - if you read closely we're tucked into the last part perhaps.

And North Carolina isn't alone; as of 1999 California says everyone has to be online: "Technology for All Staff - Provide an array of appropriate information technology devices for non-instructional staff and managers in each of the four public educational segments. Time line: This recommendation must be fully implemented by August 1999." And who is going to pay for it? "We must encourage K-12 schools .... Incentives to use local funds to support the use of technology must be developed." And what are they hoping to provide? Help Desks(2). Dr. Larry Anderson through NTCP maintains a website with a very long list of technology plans and lot's of advice on planning. A random sample finds mostly references of the clear need and certain hardware requirements and occasionally a preference for one or another architecture of computer installations without justification. No actual level of staffing of personnel was found looking through several dozen plans 1999. Ditto for a federal government initiative (3). NETC maintains a list of references to planning technology implementation in schools but again, no details on staffing needs. Mississippi State U. had a development class which reviewed many details as they helped various schools plan their technology use and implementation, but again, no details on staffing levels for technical support. Basically it looks like you can do everything without ever hiring someone to fix or install anything ever.

The beginning of the recognition of the problem can be discerned in the work of the Ceo Forum on Education and Technology. While spending most of their energy on the overall picture of technology use in schools from surveys in 1997 (with updates), they do note issues of tech support in their scheme of low, mid, high, and target technology schools. As early as 1997 they spoke of these stages:

In budgetary terms "Getting America's Students Ready for the 21st Century: Meeting the Technology Literacy Challenge - A Report to the Nation on Technology and Education", by the United States Department of Education says "On an annualized cost basis, the cost of equipment, wiring, services and supplies can increase the per student cost of technology from the current national average of between 3/4 and 1 percent to between 3 and 5 percent of total educational expenditures, or from $45 to $60 per student to between $135 and $300 per student."

RATIO INFORMATION

This matter is of extreme importance to school personnel as mounting demands cause stress and a decreasing ability to respond to anything but emergencies. How many people are needed to support a computer installation?

The earliest reference the writer has yet seen is from a Gartner Group report in 1990 (cited in the text of the 1992 report just below): the appropriate level of field support staffing is one staff person to approximately 35 users (1 user=1 computer). The 1992 report at Colombia University (4) argues for a ratio of 1/50 as normal 1/100 as possible and 1/200 as extreme. The Gartner Group, the dominant consultant in planing use of technology in the business world today (5), (6), (7), refers to tightly and loosely managed support. Their 1995 (8) study showed the lowest ratios are around 1/20. The average loose management was at 1/25. Loosely managed situations meant people whose main job was something else while also doing tech support and generally operating without a clear and limited budget. The most common ratio was around 1/60. Overall tightly managed ratios averaged 1/77. A very few ratios rose higher - as high as 1/250. The RAND nonprofit organization has a 1996 study of 8 schools (9). They note that if a system is putting out technology where none existed overall costs might be 70% higher. Laptops also represent a higher cost of about 57% (10). More recent published comments in the business world are consistently in the range from 1/25 to 1/50 with the most common references around 1/50 (11). A federal government review aims at a technology coordinator presumed to do some support plus a technician at a rate of 1 for every 400 computers (no account of lan/wan taken) but there is no evidence presented in support of this target.

Starting with the lowest ratio and applying these numbers to the school system would mean eventually as many technicians as teachers in school systems. Even the average ratio boggles the mind in the education field; in a model high technology school such a number points to one tech for every two or three teachers! This would be a major restaffing demand which local and state budgets would not be able to consider.

Are there any trends or capacities which can hope to extend technical support over more computers? The assumption is that the business world generally isn't using these already. To the extent that some use an informal support system holds out the promise that the higher support ratios can be supported with smart applications of strategies.

TCO - The Total Cost of Ownership.

In very recent years the concept and measuring stick of "total cost of ownership" has shed light on the developing technology and awareness that buying the box is perhaps not even the biggest part of owning a computer. Consider in your own experience how long a computer runs doing what it is assigned to do without some kind of intervention. Or how long before some kind of intervention is required for the computer to not be effectively useless. The total cost of ownership is usually defined as the total cost of ownership of a computer over the life time of the machine and breaking down the cost per year - this includes buying software, printing cartridges, technical support; just about everything but electricity. It is clear that the TCO in the educational setting is less than that in business. Part of a 1997 IDC study clearly shows this. Schools don't buy the high end computers and buy computers and software usually at significant discounts.

The single most effective means to cut TCO and extend tech support ratios has already been pointed out. Organize a tech support department whose job it is to do the work and motivate them to do a good job. In Gartner Grp's words "...for every $1 cut from the expense budget of the "formal" support functions, total cost of support company-wide increases by $2" (12). This is a 300% effect. This is the factor that turns ratios like 1/20 to 1/70. That the industry average is 1/50 or so indicates that most of the business world is implementing formal support systems but perhaps nothing more. The most often cited single component of a formal support approach is the help desk but we have already seen that the best tech support examined is so called "just time time" - getting it when you need it right at your computer. In "Establishing Help Desks in K-12 Schools: Lessons Education Can Learn From Business", Gary Quiring reviews the concepts and strategies of help desks. Training of technical support personnel is another factor.

The next highest effect is to train **users** for at least 3 days per year (under "Effect of End-User Training") (13). This shows about a 50% extension in technical support costs. A print reference agrees that teachers need 80 hours of staff development and use to move from the beginning level to integration and application level of technology usage (Henry Jay Becker, "A Truly Empowering Technology-Rich Education--How much will it cost?" Educational IBM Quarterly, Vol. 3, no 1., Fall 1993. U.S. Office of Technology Assessment, 1996.) The importance of this is so straight forward and simple that its brevity should not detract from its importance and indeed much attention is being drawn to staff development for teachers and administrators. On the other hand it should not out shine the importance of differentiating people whose job it is to fix and install the computers and their associated parts. A knowledgeable user is often a boon to a technician because they can provide clear feedback and solve minor problems themselves. But some knowledgeable users constantly push the envelope or do things "differently". David Cappuccio, vice president of the Network Computing Service at Gartner Group Inc., at (15) breaks down the ratio with respect to the kind of users. The idea is that "Type A" - aggressive power users - require a ratio like 1/30 while it's more like 1/60 to 1/100 with "Type B" users and as low as 1/125 for "Type C" users, ( he also recommends accounting for levels of infrastructure - add 1 for every 20 file servers, 1/12 web & email servers and about 1/5 database servers. ) The previously mentioned CEO Forum has an interactive self-evaluation form for a school or system you can fill out as well as one targeted at teacher preparedness. The From Now On Organization has papers and activities for highlighting successful staff training practices and there are many surveys about staff training levels and kinds (14)(15). However there are two forms of training that need to go on in the public school setting - one is to train users how to use the technology - literacy - and the other is how to use the technology to deliver the curriculum - integrating the technology into the teaching process. Neither is very useful without the other.

Many education researchers and policy analysts have stressed teacher training as perhaps the single most important ingredient to effective implementation of the new technologies. See, John Winkler, et al., The Rand Corporation, 'Administrative Policies for Increasing the Use of Microcomputers,' July 1986; Karen Sheingold, et al., Center for Children and Technology, Bank Street College of Education, Preparing Urban Teachers for the Technological Future,'Technical Report No. 36, 1985; and Brian Stecher, 'Improving Computer Inservice Training Programs for Teachers,' AEDS Journal, Winter 1984. Sherry Turkle, a sociologist who specializes inhuman interactions with machines, has argued for "socialization" of teachers, broadening the concept of trainingto include a wide range of behavioral and intellectual norms believed essential for effective integration of computers in education. (Footnote 9 - U.S. Congress, Office of Technology Assessment, Trends and Status of Computers in Schools: Use in Chapter 1 Programs and Use With Limited English Proficient Students, NTIS order #PB87-176723 (Washington, DC: U.S. Government Printing Office, March 1987).)

The importance of a developed technical support picture and an effective staff development training program have been recognized by the Southern Technology Council who undertook a study, chapter 5 of which examines the success and value of technology support to the adoption of computer technology in schools. Specifically "[there is a] higher correlation between this combined training/technical support index and general technology use than either of the two indices alone, in fact, this correlation was the highest across all the implementation practice areas" and rank this so important that "...both practitioners and policy makers need to keep this foremost in their minds when making planning or funding decisions." Kenneth C. Green, founder/director of The Campus Computing Project and a visiting scholar at The Center for Educational Studies of Claremont Graduate University in Claremont, CA, says "user support and instructional integration are the flip side of the same coin" here.

Fairly universal references for Wintel machines point to about $10k as the TCO from beginning to end. This (14) 1997 Gartner Group report considers a variety of platforms and architectures that are some of the developments in computer technology. The primary means examined is some kind of network structure used to minimize TCO. There are two basic models of using networks to support computers: the first uses networks to manage computers (sometimes called a "thick client") and the second makes computers into dumb client machines and everything runs over a network from a server (or "thin client".) Fat clients offer greater flexibility and can run programs purely locally. Thin clients entirely depend on a network connection. No connection = dead box with a power light on. Fat clients maybe easier to support on servers while thin clients may require more specialized support. The first cuts local TCO about 25% while the second cuts cuts TCO upto 35%. However the second system may not be appropriate in school systems - teachers alone know the value of backup plans. The business world can charge rates to support short term or redundancy systems. This is pretty clearly far too expensive for school systems. Field tech support totals tend to track proportionately with total costs so a 25% cut in TCO is also a 25% extension in technical support ratios. So this is another trend which can extend technical support ratios in the field. This does add network technical support costs but the increase is clearly less than the decrease in field support costs. In addition the author already uses capacities of network based support of computers that have come about since the above study was done. This should be kept in mind with respect to TCO but also specifically to technical support ratios. One general class of a kind of fat client is workstation configuration software - on Macs the main example is Apple's Network Assistant but there are a few other competing softwares. On the PC side less capable software like Image Cast and Ghost exist. The general idea is to leverage a network connection to configure, fix, upgrade software or role out an entire installation to a number of computers at the same time. Of course such powerful software sometimes causes problems or brings with it new problems but the picture seems clear that they make some things easier. However, bandwidth is needed and WAN speeds may not be sufficient especially if they are competing with normal WAN needs. In house bandwidth is often more than enough though. The impact these technologies has had cannot be under estimated. Instead of fixing one computer at a time or running around with ten copies running continuous installations a technician may sit at one computer and configure them all. A less established and more radical approach is Apple's Netbooting. This is a fat client approach of a new kind - the computer need not have a hard drive locally to boot off of at all. Everything comes from a server but runs locally. bandwidth demands are such that wan connections are out of the question and even LANs require excellent performance. But the one of the biggest sources of problems in the technical support picture is not absolutely counted on now - the hard drive. In addition while Network Assistant is on version 4.0 as of late 1999 Netbooting is only on version 1.1 - and there have been problems in the early software compounded by advances in other areas at the same time.

DOWNTIME

Another major factor affecting the technical support picture is the amount of down time to be tolerated. Businesses typically accept downtimes of only hours whereas schools tend to run into days.  Apple commissioned a study by International Data Corp.(IDC) in 1997 which studied many aspects of TCO in education looking at the priorities outlined by the CEO Forum mentioned near the top.  They note that down time in the business world is usually a matter of hours whereas in education it is often a matter of days. A study called Progress of Technology in the Schools: Report on 21 States, by the Milken Family Foundation reviews several stats for states.  Table 21 reviews down time. State averages range from 2.5 hrs to 15 days. People who bothered to report down time in hours fairly smoothly went from 2 hrs to 7 hrs but it tended to jump up to the 10-15 hr range afterwards. Most that responded in days said it took 2-5 days, a few 5-7 days, and a rare few noted 10-15 days. While not quantified anywhere so far surely the longer the down time tolerated the larger ratio of computer technicians to computers though at some point the down time would be critical and nothing more than emergencies would get tended to (and smaller problems would mount until they also became emergencies.)

PLATFORM INFLUENCES

There are obvious and documented influences that are platform dependent in reducing TCO and extending tech support ratios. The Gartner Group itself notes that the Macintosh platform saves 25% or higher (16) in TCO "out of the box". The same study also points out "The deployment of Windows systems for unfettered knowledge [users] increases the risk of higher technical support costs. Redeploying these systems into heads-down, well defined, production style use (such as data entry) reduces the risk that users of these Windows computers will generate support calls, thus reducing technical support costs." While it is true educationally that the skills one learns on one or another platform are largely the same (17) differing platforms are clearly not the same in terms of tech support issues. Using Macs is about as cost saving as the radical dumb client/server architecture but preserves the usefulness of the computer when the network or servers are down.

The above IDC study rates Macintoshes against the Wintel platform in many many ways. The study surveyed respondents on their ratings of how good the platform they used satisfied their needs (NOT comparing their experience on Macs or Wintel to the other platform.) Usually the data is along the lines of x% reported very good or excellent satisfaction doing y and then matching these y-actions on different platforms and reporting the satisfaction %. Here are some selections (nothing reported a higher satisfaction rate among Wintel users):

Overall satisfaction was almost 90% very good or excellent for the pure Macintosh environment vs 75% for pure Wintel. Mixed environments tended to have more negative ratings overall (perhaps constantly having to take extra steps to deal with the differences of the platforms.) Walking through the graph from Wintel to Mac: 75% of people who work in a 100% pure Wintel environment think they have it good. Add a small minority of Macs and satisfaction drops to it's lowest of about 65%. Keep adding Macs and at about 80% Macs the satisfaction rate has equaled the 100% pure Wintel environment. Then the satisfaction level rises to 90% in a 100% pure Mac environment. Traveling the other way through the graph and you start at the highest level of satisfaction and it drops the hardest in one step adding just 20% Wintel computers. But the satisfaction rate keeps dropping until you have almost no Macs left. Then satisfaction rises somewhat with no Macs left.

This kind of influence is substantiated in published anecdotes (18) (see also here).

"Let me give you a real world example: My consulting practice supplies both Mac and PC Support Technicians to a Fortune 500 client that currently has about 1000 Macs and 3500 PCs. The 1000 Macs are cared for by four Support Technicians. That's 250 computers per Tech.... The average PC tech at this company services about 80 end users. The PC users hate the MIS department and think the support they get is substandard.... Let's assume they pay these techs $50,000 a year. That's $200,000 a year to service 1000 Macs. What does it cost them to service 1000 PCs? If each PC tech can only take care of 80 computer users (and let's forget for the moment that they aren't even able to keep the users happy), this company needs to hire 12.5 techs. If we assume the same salaries, this company will spend $625,000 in salaries alone.... This isn't about zealotry. This is about tools that people use to get jobs done."

The Seattle Times reported on June 18, 1995, "At Intel, where many employees are true computer experts, the [Data Processing] department figures on one support person for every 30 Windows computers. The DP department was astonished to learn that one Intel division had 120 Macs and got along fine with a single support person."

Norris and Wong, an independent consulting firm, found the following: "The Macintosh still has a much better architecture than the PC with Windows95, and its software has far fewer complex interdependencies. These differences translate into significant cost and productivity benefits for Macintosh users (Maintenance Comparison: Macintosh vs. Windows95, Norris and Wong, November, 1995)."

Contrarily an online exchange among Windows NT techs reports 60hr wrk weeks and high turn over rate at a ratio of 1:267.

This is old news. The latest evidence of the effectiveness of the Macintosh platform is perhaps best underscored by the fact that there are no Y2K issues with the Macintosh hardware, operating system, and most of the programs. No multimeg fixes - even the fixes for Wintel vs Mac for the same software name differ markedly.

Minding that cuts in TCO track with extending technical support to more computers for the same amount of money that means a well trained installation of Macintoshes that also use network based support to manage them as much as possible should approach a 100% advantage over a mildly networked poorly trained installation of Wintel computers. Add this to an organized technical support scheme and you approach a 400% improvement over a loosely managed Wintel installation.

A 1997 Study (summarized) by the Gistics group examines the return on investment ("ROI" = TCO + productivity gains) of Macs vs Wintel and other platforms over several years. The above TCO benefit of the Macintosh platform probably relates to the reliability of the hardware and the integration of the operating system and the hardware (a much more mature process of "plug 'n play".) This ROI extends the measure to the effectiveness of the user experience in getting things done as well as exploring new strategies (an advanced skill resulting from training - "self-directed learner") that discover more effective means without a penalty for implications of faulty installations or de-installations of software or hardware. According to this measure Macs are more than 3 times the value of Wintel machines. Upgrading systems so that the educated user has the most power and flexibility on both platforms improves the ratio somewhat.

SUMMARY
From the two Gartner Group studies & the influences on relative TCO (and thus extending the tech support ratio) and assuming we begin with an adequately staffed department whose job it is provide tech support and recalling the most common ratio of 1/60 in the business world.... Add in network management and user training and you approach a ratio of about 1/130 on average. For the Mac platform you approach 1/160 on average. There are anecdotal incidents pointing to ratios around 1/200 to 1/250 but it is unknown if the factors allowing these extremes are applicable to the educational situation. For example in a public reference to the Bank of Montreal Counts on Network Computing with OS/2 Warp (about 80% down the page) there is a reference to a totally locked down (users cannot install their own software and the software is the same over a prolonged period probably never even saving to floppy) approach to using computers and their ratio of support was 1:250 with an uptime exceeding 99.5%. The only remaining factor not taken into account above was the effect of enough training that the user becomes a self-directed learner on a platform that minimizes the cost or likelihood of a faulty installation or de-installation which clearly favors the Macintosh platform as evidenced by the ROI investigation. This assumes business levels of downtime of hours at most.

COMMENTARY
It must be underscored that the single most effective influence on TCO and tech support costs after having a dedicated technical support dept. is training the user in literate use of the computer. User skills are often the stuff of jokes and outlandish stories of users being told to return the computer to the store they bought it at because they were too stupid to use a computer if they thought they could use one when there was a blackout not to mention the movies circulating the internet from security cameras show users hitting and throwing computers around. The laughter that sometimes happens serves to underscore the actual state of affairs when computers are dropped on people's desks and they are left to try and use it. Add the fact that teachers are expected not only to use the computer themselves but teach others how to use it not only in general but specifically in relation to developing and using higher order thinking skills and you can create what we have: an education system already in crisis in some places.

However no where is there data to be found on how far the ratio can go before hurting effective use of the technology and the moral of the technicians and users. By compiling references the author can support a target ratio of 1 technician for every 160 computers given details mentioned above and having a downtime of hours rather than days or weeks. From the author's personal experience and some anecdotal accounts from others there is a correlation between downtime and overall satisfaction and ratio -

If the ratio is 1:100 tech support and teachers can be proactive and creative implementing technology and training of staff - so called "just in time tech support" is normal. Turn around on problems is often less than a day. Computers are the primary tool of delivering curriculum and teachers can often fix minor problems quickly on their own.

At 1:300 tech support only deals with problems and new projects require cooperation of people not normally assigned a site. Problems often require upto a week though emergencies can be fixed with a day or two plus shipping time. Teachers grumble. While used regularly, computers are a supplemental delivery of curriculum.

At 1:500 tech support only fixes emergencies (within a week) and only rarely puts new projects into action and out sourcing is more common. Principals grumble and teachers sometimes find "someone" to fix their problem. Problem reporting procedures are circumvented sometimes. Computers are used tangentially.

At 1:700 tech support only fixes emergencies a month late and almost always outsources new projects/installations. Tech support isn't trusted, and computers are not used much if this level is maintained for over a year.

The whole subject matter is being examined by a group of people working with Director of Technology Planning and Support, Janice Johnson, of NC DPI. We are surveying NC school systems for effects of inadequate technical support ratios. The Southern Technology Council , mentioned above, reviewed a number of practices at various levels - a summary follows:

REVISION HISTORY

Rev 2.99 added some interesting quotes and began to examine the question of quality of service vs ratio. Also qualified references to dates rather than saying this is the way things are whenever this page is viewed. I honed the ratio argument to this point - a ratio of 1:160 can be supported with a downtime of less than a day or so. Factors that relate to this ratio (in order of importance):
• a dedicated tech support dept is in place overseeing a largely stable installation of computers (other than simply upgrading),
• users are well trained and can solve many small problems and communicate effectively with techs,
• technicians can use network resources like Network Assistant to fix many problems, and
• Mac's are the dominant (probably at least 80% Mac) platform.

Rev 2.9 added some major citations (Milken, CEO Forum, and IDC studies) which mostly supported the arguments already made but also underscored the influence of downtime.

Rev 2.8 added several references which lend support to the general argument and cleaned up a few typos and some structural issues.

Rev 2.5 converted it to a web page. Also Apple re-engineered their web page so that the 1995 Gartner study was no longer directly available though cited in several places. I eventually found it again at http://www.apple.com.au/documents/whymac/ggstudy.html at least on 1/13/99. Numerous Web Mac advocate websites pointed to the former Apple web location (search the internet for "ggstudy").... So in an effort to preserve the raw information (well it's just a summary - btw the phone number to order a copy of the paper doesn't work anymore) I saved it for use here. Hope Apple doesn't mind. Here it is.

Rev 2 fixed a few minor problems but also caught a problem with reference to an economy of scale / formal technical support system. The two were confused and the wrong value was attached anyway. However the outcome of the ratios combining these effects about balanced out. Additional footnotes were added on some points as well.

The kernel of this research began with an afternoon dogged search across the internet for references on appropriate ratios in tech support work. This evolved over a few emails. However the works was crystallized into the first paper in early December, 1998 in a file called Ratio Paper 1.0 or v1.

 References not folded into the above yet -

Staff development training: http://www.aces.k12.ct.us/www/ceca/sdsurveya.html Majority of school systems provide less than 5 days of training for both literacy and professional foci and there is a need for more.

a list of state technology plans: http://mdek12.state.ms.us/plan.htm

http://members.unlimited.net/~kumbach/K12/comptech.html a school's plan which while substantial does not specify what technical support ratio is needed.