Despite much talk about the "information age" or the "knowledge-based enterprise" there are no generally accepted principles to guide executives in reconciling the euphoric promises of the computer advocates and their staff's ability to prove that information technology investments are profitable. This article will present a number of perspectives from which to judge the value of information-related expenses.
The randomness of the 1994 data should not be surprising. While I was the Chief Information Officer at Xerox in the 1970's I was able to get reliable data about computer budgets, business indicators and financial results for more than fifty worldwide operating divisions. Although the financial results for each operating unit varied over a wide range, their computer budgets and computer technologies were similar. Information technology did not appear to be a key to profitability as was claimed by many.
By 1985 I started circulating a scatter diagram that displayed profit performance for eighty four public corporations, as compared with their computer expenses. No apparent connection between computer spending and financial results were found. In 1990 I finally published computer budget and profitability numbers for 292 companies. The results were again a random scatter.
The 1994 data shows that the relationship between corporate profitability and computer spending has not changed in more than twenty years. It is unlikely that any direct relationship between computerization and profitability can be ever demonstrated. Computers are only a catalysts. Business value is created by well organized, well motivated and knowledgeable people who understand what to do with all of the information that shows up on the computer screens. Different people will get different results from an examination of similar information obtained from identical information technologies.
There is no basis whatsoever for proposing that a particular industry, such as food packaging or appliance manufacturing, may need a particular percentage of revenue devoted to information technology. Although the 1994 median value for all information systems-related spending (inclusive of software, telecommunications, supplies, training, etc.) as a percentage of revenue is only 1.8%, that number varies over a wide range, even among companies in the same SIC (Standard Industrial Code) grouping.
Within each industry corporations employ varying approaches for deciding how much of the required components or distribution services they make or buy. Two corporations, competing in the same industry, reporting identical revenues, may nevertheless operate with a completely different "value chain". Therefore, the amount of computing power they will need to support their production, purchasing, administrative and selling efforts would also vary a great deal per dollar of revenue.
It is primarily the overhead workers who are the major participants in all information management processes and who consume a disproportionate amount of computer services. That becomes apparent by making a simple comparison between the expenditures for Sales, General & Administrative costs and the spending for information technologies for 468 corporations. That sample includes a wide diversity of enterprises, ranging from mining, food processing to finance, services and utilities:
The cost drivers that determine computer spending are more complex than should be inferred from examining only one variable. The simple relationship for Sales, General & Administrative expenses and information technology budgets must be tempered to reflect the influence of additional indicators, such as the differences in employment patterns, the degree to which a business may be computation-intensive and the extent to which some industries have been more inclined to adopt computerization.
For 143 corporations included in the above diagram I have also collected data about employment structure (e.g. managers vs. clerical vs. factory operatives), the extent of automation (e.g. number of personal computers per employee) and affordability (e.g. profitability of the firm). My analyses shows that a small number of selected variables will predict, with better than 90% reliability, the expected level of computer spending for most firms. A computer budget will always grow in a direct proportion to the amounts reported as Sales, General & Administrative expenses. The computer budget will always increase if a company is profitable and be cut if the company loses money. The computer budget will expand if a company employs a large number of professional workers and if it has installed a disproportionate number of personal computers relative to total employment. Amazingly, the computer budget will take a negative adjustment for each employee classified as executive.
I have used this relationship to assess whether computer budgets were greater or lesser than could be expected. I found that corporations with superior profit performance kick in an extra amount of money for computer spending because they can afford it. Since there is no evidence that computer expenses and profitability are directly related, such generosity is most likely misplaced. Any surplus funds should go for more demonstrable means for increasing shareholder value.
The opposite argument could be made against budget cuts for companies in trouble. Corporations that are in a profit crunch always seem to place high on their priority list reductions in computer expenditures. The use of computers in simplifying business processes has been demonstrated in a sufficient number of cases to make that a credible investment. If management possesses a demonstrated capacity how to deploy information technologies for making business process improvements, then information technologies could offer attractive opportunities for relief from competitive or pricing problems which may be less amenable to direct management intervention.
The number of professional employees on the payroll, such as accountants, financial analysts, market researchers and engineers explains about twenty percent of the total spending levels for computers. Of these, the controller's department is always one the largest consumer of computing power. Some parts of finance departments spend more on computers than on employee compensation. Insofar as the computer budget may be an expense for which there is no activity-based accountability, a corporation may find itself spending the expected amount for computers, but still misallocate its resources on activities that do not generate profits. Therefore, the allocation of computer costs wherever they may do the most good is at least as important as making sure that the total amount is not excessive.
The depressing effect of the number of executives on computer budgets is not only amusing, but also not surprising. It does not mean that executives do not use computers. It only suggests their share of Sales, General & Administrative costs and of personal computer expenses is already greater than what they can consume in computer services. It confirms that executives do not spend as much on computer for themselves as their subordinates do on their jobs.
The budget for information technology is not determined by its Outputs, such as profitability or revenues, but by its Inputs, as shaped by organizational structure, the size of the corporate overhead, the opinions about the benefits of computerization, the proliferation of personal computers or the number of staff people.
Unless the proposed computer budget is clearly out of line with benchmark spending, management would find it more rewarding to examine the causes that drive information technology expenses instead of speculating about their consequences. The information technology budget, in isolation, contains no meaningful information by which some financial analyst could judge either its utility or its size.
The primary focus for any reviews of information technology spending should be the amount expended on Sales, General & Administrative line items. If a company decides to employ additional accountants or financial analysts, then the budget must provide for their computing sustenance, just as one arranges for secretarial support, personnel services, furniture or telephones. If there is a demand for computing services, there should be an efficient supply to satisfy it. Manage the demand first and only then examine whether the supply is efficient.
During the annual budget reviews the insistence that a chief computer executive proves profit gains from the proposed computer spending is misplaced. Value from information technologies can be extracted only by operating executives who have harmonized their organizations to earn a profit in a competitive marketplace. In that arena computers can certainly have a role to play, sometime a leading one, but certainly not the decisive one, which is well managed and well motivated people. A computer by itself is worth only its resale or scrap price, which is not much.
Asking for a direct tie-in between increased funding for computers and a commitment to deliver provable savings is an invitation to prepare figmental projections that demean both those who produce them as well as those who accept them. The value of an oil furnace is not in reducing fictional medical bills, but the price of delivering reliable heat. During budget reviews management must be first satisfied that operating executives have improved their performance expectations because their plans are hinged on reliable computer services. If such is the case then the examination of the efficiency of delivering computer services becomes much easier. It is not the absolute amount of money spent on the information technology budget that matters. It is the full display of the life-cycle unit costs per seat for computer services that would offer a tangible target for any budget probe.
A cost- and quality-focused review of the efficiency of proposed computer expenditures can be accomplished by applying well-understood value-engineering techniques. To address the broader questions of computer effectiveness requires different metrics. The current approach of treating computer expenses as a hard-to-budget overhead expense does not yield satisfactory answers.
Statements such as "...our employees are our principal assets," "...computers are the company's strategic investments for the future ..."and "...we acquired know-how for a premium price..." are empty gestures if they cannot be enumerated.
Comptrollers require that the financial justifications of major applications software development projects includes the calculation of a Return-on-Investment . It assumes that each prospective gain in knowledge how a company functions is a series of isolated happenings. It treats each software investment as if it was a machine with a defined amortization schedule. Yet, the knowledge and skills of a workforce that has become accustomed to doing business by means of computers is a continuum. The effects of increased computerization can be only observed and measured as gains in overall business performance.
The fault with the current approaches to investment analysis of business automation lies in a two hundred year bias that favors capital as the measure of all performance. According to this deeply embedded view, what matters is the efficiency with which capital is deployed. Annual reports will prominently display trends in Return-on-Equity. Executive bonus plans will be tied to achieving Return-on-Assets targets. Finance committees will require a favorable Return-on-Investment figure for all projects. Insofar as the contributions of people, information and knowledge are concerned, the financial statistics remain silent because none of these contributions to creating a greater economic value are recognized in generally accepted accounting procedures.
The fact is that the costs of information have long ago surpassed the costs of equity capital. Capital, in its various forms, is now a commodity, which can be easily borrowed, leased, subcontracted or outsourced for a price. Except for some firms, such as in steel, mining, transportation or real estate, the scarce commodity is information. It is through effective information management practices that the users of information - the managers of a firm - create all business value. I have studied the costs of management, or costs of overhead, for more that twenty years and have found that what corporations report as Sales, General & Administrative expense is a reliable - if modestly understated - approximation of overhead costs obtained by means of a more exhaustive analysis. The following is a diagram showing the distribution of the ratio of management information costs (e.g. S.G.&A) to the annualized costs of equity capital for 2,959 US corporations.
It is a surprising diagram. Only 9.4% of the corporations are capital-intensive. The remaining corporations are information-intensive, with the median value of 472% and ranging as high as 3,200% for a consulting firm that has practically no assets, but an enormous overhead. There are many ways of interpreting the significance of this distribution, but what stands out is the confirmation that management - the processors of all information - now have a much greater stake in a business than shareholders. It also demonstrates that the capital-based industrial economy has been superseded by the managerial-dominant information economy.
Is the information-intensity of the US corporations an aberration? Here are summary statistics for other economically advanced countries:
# of Median % of Firms Companies Country Info/Equity with Info > Equity Ratio 123 Germany 1,046% 94% 162 Italy 528% 91% 2959 USA 472% 91% 1175 U.K. 465% 86% 1768 Japan 407% 95% 120 Switzerland 379% 87% 308 Canada 158% 60%
With the exception of Canada - a country that depends on natural resources for sources of much of its wealth - the information over equity ratios are not dramatically different for other countries as well.
The lack of relevance of equity-based ratios is also revealed in the large differences between the accounting definition of shareholder equity and its stock-market valuation. Investors often pay large multiples of the accounting valuation to acquire a controlling shareholder position. By paying a premium over "book" value of shareholder equity, the investors recognize that it is the productivity of information that shapes the value-creation capacity of an information-rich enterprise. Investors understand that what matters is the productive capacity of the accumulated information resources, except that the accounting reports remain silent about that.
All measures of productivity are defined as the ratio of Output/Input. The fault with the existing measures is that they determine Output in terms of accounting profits, as defined by generally accepted Financial Statements. Input is defined by what is reported on the Balance Sheet .
For example, the Return-on-Equity would define the Output of a firm as Profit after Taxes, even though this does not include any compensation to the shareholder's for their equity. The Profit after Taxes would be then divide that by Shareholder Equity as the only critical Input. As illustrated above, the Shareholder Equity is not a critical resource. The Return-on-Equity ratio cannot be a good indicator of performance since the numerator that does not fully account for all of the costs and the denominator includes less than 10% of what makes a company profitable.
Accounting rules do not accept that an information-productive corporation can convert a stream of information expenses into earnings or an asset. Information is treated as an instantly perishable commodity. Lenders like this approach because it tells them what is the likely salvage value of a firm if it fails and the employees abandon it. In contrast to this view of a firm as a corpse, information productivity metrics will suggest an examination of the living worth of organizations who have learned how to function within their competitive arena.
Management Value-Added is what is left over after absolutely all costs are fully accounted for. This calls for subtracting from the Profits after Tax an allowance for the costs of shareholder equity as well as other adjustments to correct for accounting peculiarities largely influenced by the tax code. The costs of shareholder equity is obtained by multiplying the Shareholder Equity shown on the Balance Sheet by the costs of shareholder capital.
Rigorous cost analysis is necessary to isolate all expenses that cannot be directly attributed to the delivery of goods or services to customers and then designate them as the Costs of Management. Another way of obtaining identical results is to first identify all direct costs of goods and services. Whatever is left over can be then assigned as the Costs of Management. Both methods are painfully labor-intensive. Both methods run into difficulties from those who find it expedient to attribute their own overhead expenses to where they would be recorded as an allocated element to the direct costs of good and services.
Nevertheless, several hundred organizations have examined their full costs of information-related activities, including the costs of coordination, meetings, secretaries, consultations, interviews and every other activity that is not directly related to delivering what a customer wishes to purchase.
I have designated the ratio that fully accounts for the Management Value-Added as well as all of the Costs of Management as the Return-on-Management productivity index, or R-O-M.
There is a way of approximating the Return-on-Management productivity measure by accepting the reported costs of Sales, General & Administrative expenses as a fair estimate of the Costs of Management. That becomes the Information Productivity index. Its overwhelming advantage is that it is usually available from public sources as a certified accounting entry. It makes possible quick and preliminary estimates of productivity. It bypasses, at least initially, disputes about what are the Costs of Management To select the right group for benchmarking comparisons I look for corporations that are of a comparable size, operating under similar competitive circumstances and carry related SIC (Standard Industrial Classification) designations. Such information is available from commercial data bases. To illustrate such a ranking I have chosen the pharmaceutical industry ( all data in local currencies, for 1994):
Company SIC Equity Income SGA IP(TM) MERCK & CO., INC. 2830 11,139,000 2,997,000 4,408,100 0.415 ABBOTT LABORATORIES 2830 4,049,400 1,516,683 3,017,971 0.362 LILLY (ELI) AND COMPANY 2834 5,355,600 1,286,100 2,295,400 0.315 SCHERING-PLOUGH 2830 1,574,400 922,000 2,448,900 0.309 AMERICAN HOME PRODUCT 2830 4,251,910 1,528,178 3,992,774 0.271 BEECHAM GROUP PLC 2834 1,491,500 239,400 388,900 0.213 PFIZER INCORPORATED 2830 4,323,900 1,298,400 4,390,200 0.192 ROCHE HOLDING AG 2834 16,422,000 2,860,000 7,218,000 0.157 WARNER-LAMBERT 2830 1,816,400 694,000 3,250,600 0.155 PROCTER & GAMBLE 2834 10,410,000 2,543,000 9,632,000 0.151 SANDOZ AG 2834 6,887,000 1,734,000 7,229,000 0.140 ZENECA GROUP PLC 2834 1,685,000 443,000 1,951,000 0.136 SMITHKLINE BECKMAN 2830 1,546,700 229,200 2,221,800 0.030 SANKYO COMPANY 2834 341,178,000 39,155,000 190,225,000 0.018 BAYER AG 2834 16,602,000 1,970,000 14,654,000 0.015 SMITHKLINE BEECHAM 2834 549,000 72,000 2,790,000 0.005 SANOFI SA 2834 17,359,000 1,328,000 11,237,000 -0.044 TAKEDA CHEMICAL 2830 645,401,000 51,430,000 257,222,000 -0.064 RHONE-POULENC SA 2834 32,496,000 1,915,000 21,163,000 -0.071It is important for the top management of a firm to understand their relative productivity ranking, because the information management initiatives they may need will differ. For instance, high productivity companies will find it always difficult to sustain their superior standing. They should be willing to accept information-related investments that offer high-payoff opportunities, even though that may involve taking high risks. Companies with a negative Information Productivity rank will have to accept a highly conservative approach in selecting information systems projects because they cannot afford many failures.
An Information Productivity ranking should be seen as a start in search of finding out whether or not information technologies are "aligned" with the business of the firm. In more than half of the low-productivity cases examined in the past eleven years it was not a poor choice of technology that explained why information productivity performance was substandard. Corporations suffering from a case of negative Information Productivity either have "congested arteries" (e.g. excessive S.G.&A) or "excessive weight" ( e.g. excessive assets). A further investigation into the probably causes of low or negative Information Productivity will usually result in a fairly good estimate of the priorities and worth of potential information management projects.
Keeping track of Information Productivity for every operating unit of a firm will offer a better way of measuring the contributions of decentralized units and offer more sensitive indicators that would warn about adventures that trade long term results against current accounting profits.
The following tabulation shows the results extracted from a global data base of audited financial statements for 1994. The companies making up the sample adhered to standards largely conforming to generally accepted US accounting principles:
Companies Total Revenues Weighted Country in Database (US$000) Average IP(TM) SWEDEN 24 75,540,278 0.117 FINLAND 57 58,337,766 0.078 IRELAND 48 20,526,573 0.077 UNITED STATES 2,959 4,839,398,019 0.077 NETHERLANDS 78 321,808,668 0.066 NORWAY 65 28,838,430 0.042 UNITED KINGDOM 1,178 1,032,147,384 -0.032 GERMANY 123 650,231,734 -0.036 AUSTRALIA 20 25,393,032 -0.119 KOREA (SOUTH) 41 113,933,255 -0.120 BELGIUM 23 17,921,628 -0.128 FRANCE 124 511,606,973 -0.145 DENMARK 78 45,369,476 -0.159 JAPAN 1,767 5,233,053,608 -0.170 ITALY 161 346,779,093 -0.214 BRAZIL 33 60,641,605 -0.346 CANADA 311 265,364,969 -0.367 MEXICO 30 28,921,380 -0.619 SWITZERLAND 108 244,267,380 -0.689 ----- -------------- 7,228 13,920,081,251
The 7,228 companies in this tabulation account for $13.9 trillion of revenues, which represents perhaps a fourth of the world's productive capacity. It is surprising that the US ranks high above Japan in Information Productivity. This supports one view that whatever gains the Japanese make in their factories, they may lose in their offices. The low productivity of Switzerland is largely explainable by their reliance on ample equity capital for the generation of revenues.
As knowledge-based organizations increase their share of resources devoted to information management it becomes prudent to adopt new metrics for evaluating operating results. Information-based productivity ratios, such as the Return-on-Management (R-O-A) or Information Productivity (IP) indices are especially sensitive in diagnosing deficiencies that otherwise could remain hidden without a measure that surfaces their effects. The Information Productivity measures of firms such as AT&T, Xerox, IBM and General Motors started into a nose-dive years ahead of their decline as reported by their published financial statements. All of these firms declared profits and paid bonuses even through the Information Productivity index turned negative - a sure sign that corporate management was not delivering a positive Management Value-Added.
There is ample evidence now that corporations employing identical land, identical labor force, identical computer technology or comparable amounts of capital assets deliver vastly different financial results. That largely holds true regardless of the form of government, national culture or taxation. Therefore, the only difference that makes a difference must be the skills and capabilities of management. For this reason all value creation in an information-based enterprise must be attributed to management who coordinate, motivate, lead and organize resources. If the Information Productivity metric is taken up by Boards of Directors as one of the metrics by which to judge managerial performance, many corporations now reporting accounting profits may have stop heaping more rewards on management.
Information-based productivity ratios also offer the potential of giving to the public sector better metrics for tracking their often claimed, but never proven improvements. Information Productivity is applicable to assessing the ratios of Outputs to information Inputs in the public sector where conventional techniques reveal only costs and but do not measure the results in a consistent manner.
For example, I have counted how many people in the Pentagon, contractors, clerks, communications specialists, auditors, congressional committee staffs, inspectors, comptrollers, accountants, quality assurance engineers, public relations officials and experts of all sorts (e.g. Costs of Management) were engaged in minding what a single soldier was doing while in the harms way (e.g. delivering Management Value-Added).
Similar statistics are now available for a number of public institutions, such as the New York City Board of Education. It reveals how many dollars of management are necessary to deliver a single dollar's worth of teaching in front of a classroom of pupils. While everyone is complaining about "bureaucracy" more could be accomplished by asking for evidence of Information Productivity gains every time a public agency requests additional funds.
Information Productivity analyses have opened new avenues for exploring issues that have been debated endlessly without facts that could be checked by everyone. For instance, is there such a thing as the much talked about "productivity paradox" that investments in information technologies have not shown up as yet as gains in the aggregate economic productivity? My Information Productivity studies suggest that there is not a shred of evidence that massive investments in information technologies have increased US productivity performance. The best one can say is that computers may have prevented further deterioration than would have otherwise taken place.
Another question that puzzles legislators concerns the productivity of small businesses. Are small business more productive than large corporations? My Information Productivity studies show that very small businesses show a poor average productivity because of their high failure rates. Once a corporation reaches a critical size, in the $100 million per year range, its productivity becomes extraordinary, it is a source of employment gains and becomes a heavy investor in innovation. As corporations grow from large to very large, their average productivity declines severely. The overall poor productivity performance of the US over the last decade has been largely caused by huge losses from a handful of giant corporations that account for a disproportionate share of employment and economic value-added of this country.
The Information Productivity ratios have the advantage that they are derived from audited public databases, reporting the results produced from individually identified firms. Information Productivity analyses are applied micro-economics using good and verifiable data. In contrast to that, public policy discussions about productivity are based on Government statistics that are admittedly incomplete, of questionable applicability to information-intensive enterprises and subject to incomprehensible "adjustments" to account for changes in prices and policies.
The term "management" is used here as applicable to every information activity that is not directly engaged in the generation of revenues. I define customers as the people from whom you collect cash.
If a newly-hired factory worker spends half a day in general orientation and indoctrination meetings, that makes him partake in a managerial activity. The work of an executive secretary can be also seen as managerial, since this job involves information gathering, storage and dissemination tasks. Meetings, training, consultations, giving advice, accounting, administration, interviewing or correcting quality defects by this definition are all managerial functions because if they would be fully accounted for, they would be charged to "overhead" and not to direct costs of sales.
In a typical company an average employee spends at least one third of their time acquiring intra-company information that is unrelated to the delivery of goods or services. Employees in managerial and staff position expend all of their time on tasks not directly related to the delivery of goods or services. More than 25% of payroll dollars in an information-intensive enterprise, and well over 50% of the payroll dollars in most government agencies, are expended on information activities that should be recorded as managerial overhead.
All of this learning, talking and listening costs money. If that accumulation is ultimately convertible in greater productivity for the enterprise, then the expense was worth it by earning a return on the Knowledge Capital investment.
Consider the costs of managerial knowledge accumulated by an employee over a ten year period. With full costs of employment at about $60,000 per annum, the decade-long exposure to managerial information would result in knowledge Inputs costing about $150,000. What would be then the measurable Outputs from all of that accumulated knowledge?
The creation of Management Value-Added is something that defies the laws of conservation of energy. These laws state that Output of any system in the Universe can never be greater than its Input. Delivering a positive Management Value-Added must be therefore an act of creativity that springs forth from something that is intangible, as if it were an artistic conception. The source of this creative energy is Knowledge Capital which can be quantified only indirectly by observing how much Management Value-Added it yields.
Another way of looking at the same phenomenon is to infer the value of Knowledge Capital from its periodic yield. If Management Value-Added is the interest earned from an accumulation of knowledge residing with the firm, then the value of this principal can be calculated by dividing the Management Value-Added by the price one pays for such capital.
Mergers and acquisitions of companies have made the pricing of all capital explicit. It is the risk-adjusted interest that an investor is willing to pay. Since investors cannot differentiate between the price of capital for financial or knowledge investments because they are intermingled, I use the identical price for all capital as a first approximation. This yields a simple equation:
Knowledge Capital = Management Value-Added / Price of Capital
This relation makes it possible to prepare a revised Balance Sheet for any firm, by adding a line item Knowledge Capital on the Asset side of the ledger, and by increasing (or decreasing) the reported valuation of Shareholder Equity by the identical amount.
Abbott Laboratories is an example of a company that has successfully kept accumulating Knowledge Capital faster than Shareholder Capital. Its earning capacity and productivity is gaining not because they are hoarding financial assets, but because they are using the capabilities of their people more effectively.
In the chart below I am showing a plot of a three-year moving average of the ratio of Market Value to either Shareholder Equity or to Shareholder Equity plus Knowledge Capital for Abbott Laboratories. According to the traditional Market-to-Book valuation method, an investor may have started worrying in 1991 that the shares of Abbott Laboratories were overpriced. However, if the investor would plot a three-year moving average of Market -to Book plus Knowledge Capital , the stock would be seen as priced fairly:
The sustained stability of the Market-to-Capital ratio which accounts for the steady rise in the Knowledge Capital of Abbott Laboratories confirms that the stock market will recognize the accumulation of knowledge as an asset even though the accountants do not. The stock market will also reward the accumulators of Knowledge Capital because investors recognize that the worth of a corporation is largely its management, not its physical or financial assets.
I have analyzed a number of corporations using this method and find that adding Knowledge Capital to book value Equity Capital shows a good correlation with the prices investors are willing to pay for shares of information-intensive enterprises.
Almost everything that counts as an accumulation of knowledge is usually paid for and written off as an overhead expense and charged against current profits. This can decreases profits, increases expenses and diminish Information Productivity unless management sets out deliberately to treat all overhead expense as a potential investment in Knowledge Capital. Every manager should therefore monitor what portion of their Sales, General & Administrative expense is frittered away as a one-time happening and how much of it can be seen as an asset with a residual value.
In the case of Abbott Laboratories, that is an important question since more than a half of its stock value is derived from its gains in Knowledge Capital. The answer can be found in computing the firm's Overhead-to-Asset Conversion Efficiency.
The ten year sum of all Sales, General & Administrative expenses for Abbott Laboratories adds up to $18.9 billion. During that period Knowledge Capital has grown by $8.6 billion. This the defines the Overhead-to-Asset Conversion Efficiency as 44.3%. It means that slightly less than a half of overhead expenses has been well expended for the benefit long-term utility. A way of displaying this steady trend is shown in the following chart:
The Abbott Laboratories have succeeded in generating Knowledge Capital faster than its S.G.&A. They are highly profitable because their accumulated knowledge can be reapplied without further expense. Their current S.G.&A. is indeed lower than most of their competitors, because they do not have to pay for all of it in every fiscal year. They re-cycle S.G.&A at a very low cost, which saves on expenses and increases the value of each employee. 
I have analyzed the Overhead-to-Asset Conversion Efficiency of hundreds of companies and found that a surprising number of companies suffer from a negative conversion efficiency. As they cut S.G.&A. during re-engineering, their long-term Information Productivity(TM) declines because their attrition of Knowledge Capital proceeds at a faster rate than the savings generated from wholesale dismissals of people. There seems to be a trade-offs between indiscriminate cost-cutting and the demoralization of valuable employees that leads to a suicidal death spiral.
One of the most efficient instances of Overhead-to-Asset Conversion Efficiency is the Microsoft Corporation. In the period from 1986 through 1995 it gained $8.3 billion in Knowledge Capital while expending only $10.5 billion for S.G.&A. To explain Microsoft's extraordinary Overhead-to-Asset Conversion Efficiency of 79% one has to understand that Knowledge Capital does not need to reside exclusively in the heads of employees. It also occupies the mind-share of customers who have expended their own time and money to became habituated to Microsoft products.
An additional 10% of all computer budgets is expended on new projects. A close examination of proposals will show that much of the financial justification for starting anew is to reduce expenditures for maintenance. If someone would try to sell a house that requires an annual upkeep equal to a half of the purchase price, nobody would buy it. A rapidly deteriorating capital asset is not worth much. Yet, the very high ratio of life-cycle maintenance costs to the original acquisition cost demonstrates that today's application software is one of the flimsiest artifacts that management will ever buy.
The idea of constructing software to qualify as a high-residual value, low-maintenance capital asset has never been accepted. Just as in the story of the three little pigs, management does not have the patience to invest in an architecture that is survivable in the long run. The computer people, the vendors and the consultants also prefer whatever is new, fashionable and quick.
The reason for the flimsiness of the application software can be found in the lack of understanding by most executives that software has become an increasingly significant store of a corporation's Knowledge Capital. While a comptroller may question the trade-in value of fork-lift trucks, when it comes to software it will be written off without any examination as to its reuse. Software expenses are now wasted because management uncritically accepts the view that software is largely unrecoverable every time technology, organization or business practices suffer from major changes.
The existing methods of accounting do not recognize that for most corporations the accumulation of expenditures for software over a ten year period will exceed the value of Shareholder Equity in about 30% of cases. As long as software is treated as an expense that must realize short-term returns, corporations will be paying for software that performs similar business functions many times over without the benefit of any reuse.
Software asset management is perhaps one of the most exciting new opportunities for accelerating the accumulation of Knowledge Capital because it represents an encapsulation of accumulated expert knowledge that can be purchased in the open market at a fraction of its original cost.
Software should be seen as one of the best means for accumulating and preserving enormous amount of information about the ways how a corporation functions. It should be recognized as a knowledge asset so that it can be managed as something that keeps accruing value steadily, reliably and safely. It must be designed for evolutionary growth instead of keeping it alive by patching it up until such time when a sudden convulsion makes it necessary to replace it without much delay.
Management must insist that applications software is preserved by means of technical designs that accommodate rapid changes in computer technologies. Management should demand delivery of software applications that take advantage of innovations in operating systems, that adapts to revisions in organization structure and takes advantage of any streamlining of business practices
The existing methods and concepts of accounting, budgeting and planning are biased against anything that is not a tangible asset. No wonder that many prior attempts to calculate the productivity of "information" have foundered on the reluctance of the current stakeholders to be subjected to the sort of measurements that were previously reserved only for the laboring classes.
The reasons for the preservation of accounting methods that were suitable for the industrial era is for students of corporate power politics to debate. It should suffice to remind us that when industrialization induced a shift from the extraction of funds from feudal land possessions to earning profits on invested capital, most of the assumptions how to measure performance had to change. When the expenses for acquiring information capabilities cease to be an arbitrary budget allocation and become the means for gaining Knowledge Capital, much of what is presently accepted as management of information will have to shift from a largely technological view of efficiency to an asset management perspective.
Most of information technology is used in the support of overhead functions. Budgeting of computers should take place only after business processes have been simplified and information management has been made more effective.
Information costs are larger than capital costs. Therefore, conventional measures of capital productivity, such as calculating the ROA or ROI of information systems investments are inadequate in assessing the effects of improvements in information management.
Information productivity can be measured. That requires attributing to management the credit as well as the blame for all results. Information productivity evaluations offer a useful metric for assessing the performance of local units both in the private and public sectors.
Knowledge capital must be counted. Without considering the effects on the knowledge accumulated in the heads of employees, suppliers or customers, all other performance indicators remain incomplete and inadequate for judging the capabilities of an organization.