Management consulting

the measurement or capacity that results from multiplying the estimate for a single item by the total number of items. For example, the estimated quantity of material required to manufacture a product is multiplied by actual numbers produced during a recent period. This is compared with the quantity of material actually issued from store to production.

Another means of checking estimates is to compare them with data recorded elsewhere. Such a comparison must be made with care and will only be valid if the data being compared relate to identical circumstances. Data for comparison may be found in trade publications, in the files of the consulting organization or from organizations that have collaborated in benchmarking projects.

In addition to checking the validity of estimates, the consultant needs to consider the degree of error they entail and decide whether this is tolerable. Where there is a strong probability that the error is within the limits of tolerance, the estimate can be used. If the error is too high, the consultant will have to devise ways of obtaining more precise and reliable data instead of using an estimate.

Estimates often concern data on developments and trends that are independent of the enterprise concerned (e.g. market prices, energy prices, transportation tariffs, exchange rates, interest rates, inflation). Many of these estimates can be obtained from competent specialized sources, such as government agencies, banks, business research institutes, or financial and market analysts. The consultant should choose an external source of estimates with extreme caution, bearing in mind that not all sources are equally reliable. It is useful to know how the estimate was made – is it a “best guess”, a common opinion shared by many experts on the topic, or was a forecasting model used? On what concepts was that model built? The consultant should never blindly accept, and provide to the client, estimates on the basis of which the client will have to make important investment and financial decisions. Of course, not all risks can be eliminated, but the use of false information must be avoided.

Cultural issues in gathering data

Sensitivity to cultural factors (see Chapter 5) is very important in data-gathering activities, in which the consultant interacts with many different individuals and groups in the client organization. In this respect the consultant must keep in mind both the country’s and the organization’s culture (box 8.4). Even the particular microculture in different parts of an organization can influence how an interview is conducted or whether a group can be observed during work. It may be difficult and time-consuming for the consultant to determine the cultural norms of different groups, but it is essential.

8.5Data analysis

Data cannot be used and converted into meaningful information without analysis, the purpose of which goes beyond research and appraisal. The ultimate

198

Diagnosis

Box 8.4 Cultural factors in data-gathering – some examples

●In many countries, the interview cannot possibly start until the host (respondent or consultant) has first offered a beverage to the visitor.

●There may be cultural biases that hamper the use of a data-gathering technique. In one country where English was not the first language, a consultant went through the usual steps in preparing a questionnaire to be administered to a large group of people. When the data were reviewed, the unanimity of responses was surprising. As the consultant pursued this with some members of the client system, he discovered that it was the custom in that country for those answering a questionnaire to provide the information that they thought was wanted by those administering the questionnaire. It would have been impolite to do otherwise! The respondents had all determined the kind of answer the consultant would want and had provided it.

●In a Muslim country, a consultant was on an assignment that required data to be gathered from workers, some of whom were female. When the first interview was held, the consultant was surprised to find that the respondent brought along another woman, even though the consultant was herself a woman. Obviously, having another person present during the interview raised a question as to the validity of the data. After several interviews had been conducted, the consultant discussed this with the client. Only then did she learn that in that country (and there are differences among Muslim groups) a woman was not permitted to converse with a stranger, even another female, without an older woman from her own household being present.

aim of the consulting process is to initiate and implement change, and data analysis should help to achieve this.

A correct description of reality, i.e. of conditions and events and their causes, is therefore only one aspect of analysis. The other, more important aspect is to establish what can be done, whether the client has the potential to do it, and what future benefits should be obtained from the envisaged changes.

There is, therefore, no clear-cut distinction between analysis and synthesis. Synthesis, in the sense of building a whole from parts, drawing conclusions and developing action proposals, starts somewhere during data analysis. Thus, data analysis evolves gradually to synthesis. Indeed, analysis and synthesis are two sides of one coin, and consultants apply them simultaneously. A consultant does not have to discover new wholes by combining parts each time he or she undertakes an assignment – theoretical knowledge and practical experience help the consultant to synthesize while analysing.

If the consultant can establish that the problem observed follows a general rule, he or she will apply the deductive method, i.e. assuming that the relationships described by the rule exist in the case being dealt with. The consultant has to avoid the temptation to draw hasty conclusions from superficially analysed facts and allow ideas to become fixed before examining the facts in depth (“This is exactly the same case I have seen many times

199

Management consulting

before!”). Put in other terms, it is not possible to use deduction where induction applies, and vice versa. In practical consulting work, the two methods are combined and complement each other, as analysis and synthesis do.

Editing the data

Before being analysed as described below, data need to be edited and screened. This includes checking their completeness, verifying the clarity of recording and presentation, eliminating or correcting errors, and making sure that uniform criteria were applied in data-gathering.

As an example, consider the recording of a production operation: if 19 recordings show a duration of between four and five minutes, one recording indicating 12 minutes cannot be used in calculating an average figure. Such an extreme recording can happen in quite different contexts – for example in accounting, where overhead costs may be inaccurately distributed among various products, or where one account may include items that should be in a different account.

Cross-checking helps in some instances: for example, information obtained in an interview can be verified by subsequent interviews. In other cases there is no possibility of cross-checking and the consultant must use experience and judgement, together with advice from the client’s staff, to “clean” the data prior to analysing them.

Classification

The classification of data was started before the beginning of fact-finding, when criteria were established for the organization and tabulation of data (section 8.3). Further classification, and adjustments to the classification criteria, are made during and after fact-finding (e.g. the consultant may decide to use a more detailed breakdown of data than originally planned). If facts are recorded in a way that permits multiple classification (e.g. in a computer), the consultant can try several possible classifications before deciding which one is most relevant to the purpose of the assignment.

Both quantified and other information needs to be classified. For example, if complaints about the shortage of training opportunities come only from certain departments, or from people in certain age groups, the classification must reveal this.

The main classification criteria used by consultants are:

●time;

●place (unit);

●responsibility;

●structure;

●influencing factors.

Classification of data by time indicates trends, rates of change, and random and periodic fluctuations.

200

Diagnosis

Classification by place or organizational unit helps in examining problems associated with various parts of the organization and devising solutions related to the specific conditions of each unit.

Classification by responsibility for facts and events is a different matter – in many cases responsibility may lie outside the place (unit) where a fact has been identified. The consultant may need to identify responsible organizational units and/or particular persons in these units.

Classification according to the structure of entities and processes is essential and uses a number of criteria. Employees, materials, products, plant and equipment, customers, and so on, can be classified from many different points of view. An important objective in this case is to define how changes in components affect the whole, and to direct action towards those components that have a major influence on total results. For example, the total lead time of a steam turbine may be determined by the machining and assembly time of one component – the rotor. Classification of customer complaints by product or production unit indicates where to focus quality improvement efforts.

Operations in a production process can be classified according to their sequence in time and presented in a table or diagram, or on the plan of the workshop (which makes it possible to indicate the directions and distances of movements of materials). Organizational relations, formal and informal, can be classified by means of charts, diagrams, matrices, and so on.

Classification by influencing factors is a preparatory step in functional and causal analysis. For example, machine stoppages may be classified by the factors that cause them: lack of raw material, break in energy supply, absence of worker, delay in scheduling, lack of spare parts, and so on.

In many cases simple classification (by one criterion) will not suffice: crossclassification is used, which combines two or more variables (e.g. employees classified by age group, sex, and length of employment with the organization).

Analysing organized data

The data, prepared and organized by classification, are analysed in order to identify relationships, proportions and trends. Depending on the nature of the problem and the purpose of the consulting assignment, a variety of techniques can be used in data analysis. The use of statistical techniques is common (averages, dispersion, frequency distribution, correlation and regression), and various other techniques, including mathematical modelling or graphical techniques, are often used. The reader is referred to the specialized literature for detailed discussion of these.

Statistical and other quantitative analysis is meaningful only where qualitative relations can be identified. For example, an association between two variables can be measured by correlation, but correlation cannot explain the nature and the causes of the relationship.

The main objective of quantitative analysis is to establish whether a specific relationship exists between various factors and events described by the data and,

201

Management consulting

if so, to examine its nature. If possible, the relationship is quantified and defined as a function (in the mathematical sense of the term), describing how one or more dependent variables relate to one or more independent variables. The purpose is to discover and define relationships which are substantive and not just accidental. For example, the consultant may find out from data gathered in various firms that the cost of a major overhaul of machine tools is in some relationship to their purchase price. If such a relationship is defined as a function, the consultant can forecast the cost of overhaul and its influence on production costs in other firms using similar equipment.

Using ratios

A common way of expressing and measuring relationships is as ratios. Ratios can be used to test whether inputs to an activity generate commensurate outputs, examine whether resources and commitments are properly balanced, or express the internal structure of a particular factor or resource.

In detailed analytical work, the ratio of global aggregate data may be broken down into analytical ratios. For example, a series of ratios is often used to measure labour productivity:

	V	=		V		x	DH	x	PW	x	W
	E		DH				PW		W		E

where V = value of production,

E = total number of employees, DH = total direct labour hours,

PW = total number of production workers, W = total number of manual workers.

There are no limits to the construction of detailed analytical ratios in any business and any functional area of management. Here again, working with a quantitative ratio makes sense if there is some qualitative relationship, and if using a ratio makes the analysis more meaningful by measuring this relationship and comparing it to a standard or another known case.

Causal analysis

Causal analysis (see also box 8.5) aims to discover causal relationships between conditions and events. It provides a key to planning change and improvements. If the causes of certain situations, results or problems are known, they can be examined in depth and action can focus on them.

In most cases, the consultant would start the investigation with one or more hypotheses as to what the cause(s) of a problem may be, based on knowledge and experience. To confirm the possible main causes the consultant needs to have a comprehensive, synthetic view of the total process or system he or she is examining, and of the whole organizational context.

202