Utility analysis

The utility analysis ( NWA ; also called point value method , point evaluation method or scoring model ) is one of the qualitative, non-monetary analysis methods of decision theory . The utility analysis is a methodology that is intended to rationally support decision-making in complex problems. The utility analysis is a relatively old procedure that has its origins in the economic "utility analysis". It became known in German-speaking countries through Zangemeister (1976). While the cost-benefit analysis only considers various criteria from the point of view of efficiency, the benefit analysis evaluates the effectiveness or the outcome. The utility analysis finds u. a. Application in controlling, in project management, in economics and even in public procurement law, wherever an assessment has to be made on the basis of several quantitative and qualitative criteria, goals or conditions. The NWA is the “ analysis of a set of complex alternative courses of action with the purpose of arranging the elements of this set according to the preferences of the decision maker with regard to a multidimensional target system. The order is mapped by specifying the utility values (total values) of the alternatives. "

An NWA is often created when there are “soft” criteria on the basis of which a decision must be made between different alternatives .

operation area

The utility value analysis is primarily intended to serve the systematic preparation of decisions and the selection of complex alternative courses of action within a decision-making process. It should make it possible to obtain a compact key figure for the benefit without losing transparency. Since the NWA takes into account not only monetary but also “soft” factors, complex issues can also be assessed and the risk of wrong decisions can be reduced.

Methodological basics

The decision-theoretical basis for the utility value analysis is the additive multi-attribute value function. This “assigns a value to each alternative depending on its attribute characteristics” . At the end, a total value for each alternative is calculated from the weighted sum of individual values per attribute. The additive multi-attribute value function for the calculation of the total value of an alternative a is:

{\ displaystyle v (a) = \ sum _ {r = 1} ^ {m} w_ {r} v_ {r} (a_ {r})}

All are included and the condition for the validity of the value function applies: ${\ displaystyle w_ {r}> 0}$

{\ displaystyle \ sum _ {r = 1} ^ {m} w_ {r} = 1}

This means that each weight must be greater than 0 and the sum of all weights is 1 (or 100%). The term is the value (the "evaluation") that is assigned to the characteristic . The following example is intended to illustrate the formulas: Three job offers are compared with one another. Two attributes are used for the evaluation, the working hours and the salary. ${\ displaystyle w_ {r}}$ ${\ displaystyle v_ {r} (a_ {r})}$ ${\ displaystyle a_ {r}}$

alternative	salary ${\ displaystyle x_ {1}}$	Assessment salary ${\ displaystyle v_ {1} (x_ {1})}$	working time ${\ displaystyle x_ {2}}$	Evaluation of working time ${\ displaystyle v_ {2} (x_ {2})}$
Consultant	€ 90,000	1.0	60 h	0.0
professor	€ 55,000	0.6	40 h	0.5
Teacher	€ 35,000	0.0	20 h	1.0

Exemplary calculation of attribute evaluations

If one now assumes a weighting for the salary of and for the working hours of , the following table is obtained: ${\ displaystyle w_ {1} = 0 {,} 6}$ ${\ displaystyle w_ {2} = 0 {,} 4}$

alternative	Assessment salary ${\ displaystyle v_ {1} (x_ {1})}$	Weighted salary ${\ displaystyle w_ {1} v_ {1} (x_ {1})}$	Evaluation of working time ${\ displaystyle v_ {2} (x_ {2})}$	Weighted working hours ${\ displaystyle w_ {2} v_ {2} (x_ {2})}$	Total value ${\ displaystyle w_ {1} v_ {1} (x_ {1}) + w_ {2} v_ {2} (x_ {2})}$
Consultant	1.0	0.6	0.0	0.0	0.60
professor	0.6	0.36	0.50	0.20	0.56
Teacher	0.0	0.0	1.0	0.40	0.40

Exemplary total benefit calculation

In the example above, the position of advisor would be the best because the overall value is the highest. One speaks of an "additive" procedure, since in the last step all partial utility values are added. However, for an additive value function to be valid, it must be independent of preference. This means that reducing or increasing one attribute causes a change in the total utility value that is completely independent of the level of the other attributes. I.e. In a track and field competition, an increase in the throwing distance for the shot put from 20 to 25 meters gives an additional number of points, which are independent of the achievements in sprinting, long jump etc.

Use

Deviating from the prevailing definition of utility via preferences over potential exchange operations, the utility of utility value analysis is to be understood as the extent to which a good is suitable for satisfying a need - or another criterion - of a decision maker. Five factors are decisive for the size of the benefit:

the one who uses the good ,
the purpose for which the property is to be used,
the situation in which the good is to be used,
the time at which the property is to be used,
the good itself.

Advantages and disadvantages of utility analysis

advantages

Flexibility of the target system
Adaptation to a large number of special requirements
direct comparability of the individual alternatives
The incomparable is made comparable by selecting common criteria

disadvantage

Comparability of the alternatives, as it cannot always be guaranteed that two alternatives will be compared in the same way.
Problem of agreement when there are several decision-makers with different preferences
Problem with the selection of criteria / weighting
very subjective in terms of weighting

Common mistakes

In the case of a simple utility function, attention is usually not paid to the fact that the individual criteria must be utility-independent. Example: fuel level, consumption and range with a full tank of a car.
For the sake of simplicity, it is not the consequences that are assessed, but the parameters of the alternatives. I.e. you “save” yourself the step of mapping alternatives to consequences. Example: The trunk volume of a car is evaluated and not the question of whether it is sufficient for the luggage.

To ensure transparency and clarity, only the most important criteria should be included in the benefit analysis.
Exclusion criteria are not included in the NWA.
When creating a benefit analysis, particular attention should be paid to the formulation of the goals or the criteria to be measured. The selection of wrong goals or criteria can result in distortions of the overall picture if, for example, irrelevant goals are set. Hall writes about this:

It is more important than choosing [the right alternative] to first determine the right goals. Because if you choose wrong goals, you solve an irrelevant problem; on the other hand, if one chooses [a wrong alternative] (based on correct goals), one ultimately only chooses [a non-optimal alternative]. It is therefore essential to ensure that the utility value analysis contains goals that are appropriate to the situation, i.e. that it takes into account all essential aspects.

Another problem arises from measuring and estimating the scores for the targets. The scaling is particularly problematic here. As shown below, only the cardinal scale offers the possibility of a relatively objective evaluation. Other ratings, measured or estimated on ordinal scales , always involve some degree of inaccuracy. Ordinally scaled values, which are then converted into degrees of target achievement by means of a transformation, can therefore offer a deceptive sham accuracy if they are displayed together with transformed, precisely measured, cardinally scaled values
Finally, the uncertainty about the future influences the result of the utility value analysis. So far, in the description of the utility analysis, it has been assumed that all measured values or estimates that have been recorded will continue to exist in the future. The possibility of improvement or deterioration was not considered. Especially when evaluating long-term projects, however, the possibility that the evaluations or weightings may change over time should be considered. A relatively simple solution to the problem of uncertainty is to carry out a sensitivity analysis and / or to specify evaluation corridors instead of point evaluations. Example: A benefit analysis is to be created for the selection of a house. The “view” criterion is included in the assessment. Instead of evaluating this for a house with “80 out of 100” points, a rating of “75-80 out of 100” could be made if it is known that the house is close to a new development area and there may be future construction sites to disturb. Consequently, assessment corridors are then also available for the total utility values. Another factor to consider when interpreting the results is the subjectivity of the ratings. Especially with surveys or ordinal-scaled criteria, the evaluator has a great influence on the result. This effect can be reduced, for example, by whole teams doing the weightings or evaluations.

Simple utility analysis

In the literature there are various statements about the course of a utility value analysis. Nagel , Tauberger and Hanusch , for example, speak of a seven-stage process, while Westermann describes an 11-stage model. Furthermore, Nollau speaks of a six-stage approach, Büssow of a 14-stage. Since the individual processes often only differ in the choice of words, a generic process model of the utility value analysis is presented below, which summarizes the common features of the individual variants.

Goal definition

At the beginning of the utility analysis, it should be noted what the goal of the analysis is, i.e. which decision problem is to be solved. The documentation is important to ensure the transparency of the NWA.

Define exclusion and selection criteria

Probably the most important and most difficult step is the determination of the criteria to be measured and evaluated. First of all, the so-called exclusion criteria should be defined. These criteria are not included in the NWA, but lead to the immediate exclusion of an alternative if the criterion is not met ("KO criterion"). If you define the maximum price of € 10,000 as an exclusion criterion when buying a car, every car that costs more is immediately disqualified. Then the selection criteria are defined. These are then weighted and rated with points. It can be helpful to use creative techniques like brainstorming to gather ideas for the selection criteria. Basically, one can differentiate between performance, cost and deadline criteria. It is advisable to create a criteria hierarchy as it may simplify the weighting and clarify the relationship between the criteria. The hierarchy gives you different levels. As the level decreases, the goals lose their importance for the total utility value due to the additive methodology. In this respect, goals of the first or second hierarchical level can become the subject of “political” discussions. All criteria should be ascertainable and measurable either qualitatively or quantitatively. The formulation of the characteristics should be as precise as possible. The criterion “reduce costs by 10%” in relation to target achievement is easier to assess than “reduce costs”. In addition, this type of designation is important for the weighting that follows. It is not expedient to compare two attributes with one another. It always depends on the difference in the characteristics of two variables. For example, there is little differentiation to claim that “vacation” is more important than “weekly working time”. It makes more sense to consider whether z. B. 10 days more vacation are more important than 2 hours less weekly working time.

Simple utility value analyzes assume the existence of a multilinear utility function without proving it. The following explanation describes the common practice that neglects or ignores the theoretical fundamentals.

A simple table is often sufficient for private or manageable economic questions. All you have to do is set the various options on the Y-axis one below the other and set the evaluation criterion on the X-axis. Another column contains the individual weighting factor for the respective criterion, i.e. the question of how high the degree of fulfillment of a possibility is in the overall priority.

Now the individual solution or offer options are processed line by line. Each criterion is assigned its fulfillment and the respective weighting with point values and the entire line is multiplied at the end. The result per line gives the determined attractiveness of a solution. It is thus possible to analyze the utility values of any number of variants table by table.

Weighting of the goals (criteria)

The central analysis step when creating the NWA is the weighting of the previously defined selection criteria. The weighting factors indicate the importance of the individual criteria. The weightings are purely subjective, regardless of the methods presented below. The decision maker determines what is important and what is not. So that the NWA and the decision-making remain transparent, the weightings should be methodical. In the relevant literature, the following methods of weighting are treated, among others:

swing
Trade-off
Pairwise comparison (preference analysis)
SIMOS
AHP
Scoring
Direct Ranking / Direct Ratio

In the following, three methods are presented that follow different approaches. The table below shows the methods and their character.

method	character
Direct ranking	The weight quasi directly associated
Preference analysis	The weight results from an unqualified comparison
AHP	The weight results from a qualified comparison

A current study by Zardari shows that the (scientific) interest in weighting methods and decision-theoretical questions has been growing steadily since 2000. For this purpose, the search queries on various scientific databases were evaluated. The following graphic shows the queries to the scopus database as an example (the scopus database of Elsevier Verlag is the world's largest database for scientific literature according to its own information). The development of the inquiries shown can also be seen when evaluating other databases. The weighting methods considered are therefore also up-to-date and relevant in a scientific context.

Development of search queries for various weighting methods in the scopus database

Direct ranking

The simplest, but at the same time inexact method for determining the weights is the direct ranking. In practice, this method is often used due to the simplicity and easy calculability of the weights. In order to obtain the weights for the individual criteria or goals, the decision maker assigns a rank to the criteria or goals. Whether this is from 0–10, as shown below, or in a different range does not matter, as the values are then normalized to 1. This evaluation method is based on an ordinal scale .

As soon as a rank has been assigned to all criteria, the raw weights r of a criterion j (criterion 1: 9) can be normalized to 1 by dividing each raw weight by the sum of the weights, whereby the normalized weight w is obtained. Formally, this can be expressed for criteria or goals from 1 to n as follows:

{\ displaystyle w_ {j} = {\ frac {r_ {j}} {\ sum _ {j = 1} ^ {n} r_ {j}}}}

The major disadvantage of direct ranking is that each criterion is viewed in isolation. It is therefore not possible to carry out plausibility or consistency checks. In addition, the phenomenon often occurs in practice that decision-makers are indifferent in their assessment, i.e. assign the same relevance to different criteria.

Preference analysis

While the direct ranking can be quite suitable for weighting a few criteria, for a large number of criteria, methods should be used that carry out a pairwise comparison. One method for this is preference analysis. Here, all criteria are compared one after the other and it is selected which criterion or goal is more important. The graphic below shows one way of making a pairwise comparison using a spreadsheet . The arrows in the figure show which criteria were compared. The letter of the preferred criterion is entered in the table. Then the ranks and the reverse ranks (ordered the other way around) are determined based on the absolute number of entries ("number" in the table). Then the weight of each criterion is determined by dividing the sum of the weights (100) by the sum of the assigned ranks (34) and then multiplying by the reverse rank. Expressed formally, the calculation is:

{\ displaystyle {\ text {weight}} = {\ frac {\ text {sum of weights}} {\ text {sum of ranks}}} \ cdot {\ text {inverted rank}}}

With this method, the weighting of the criteria is more precise than with the direct ranking method. However, weighting through pairwise comparisons takes more time. Above all, it is important to note the increase in effort with an increasing number of criteria. The number of all comparisons to be carried out is calculated as follows, where N is the number of comparisons and n is the number of criteria:

{\ displaystyle N = {\ frac {n ^ {2} -n} {2}}}

In the figure below, 8 criteria were compared, which made a total of 28 comparisons necessary. If you filled out the table completely (14 criteria), this would lead to 91 comparisons.

Analytical Hierarchy Process (AHP)

The AHP method was developed by Thomas Saaty in 1980. The method is similar to utility analysis and uses paired comparisons. Depending on the application, the AHP can be viewed as a substitute for utility value analysis. At the same time, the AHP also offers the option of weighting, with which a utility value analysis is then carried out. The most important difference to the preference analysis in the assessment is that it not only differentiates which criterion or goal is better or more important, but also how much better or worse it is. A scale from 1 to 9 is suggested for this evaluation (the values 2, 4, 6 and 8 serve as intermediate values).

Degree of importance	definition
1	Equal importance
2	Weak
3	Moderate importance
4th	Moderate plus
5	Strong importance
6th	Strong plus
7th	Very strong
8th	Very very strong
9	Extreme importance

The application of the AHP could proceed as follows: First, as with the utility value analysis, a hierarchy of objectives or criteria is created. This breaks a large decision problem into many smaller decision problems. Then pairwise comparisons are made at each level of the decision hierarchy. The results are entered in a reciprocal (inverse) matrix. A matrix is reciprocal ( reciprocity ) if:

{\ displaystyle a_ {ij} = {\ frac {1} {a_ {ji}}}}

Put simply, this means that if criterion A is twice as important as criterion B, then criterion B must be half (1/2) as important as criterion A. The table below shows an example of what a reciprocal matrix could look like.

	Criterion A	Criterion B	Criterion C
Criterion A	1	3	4th
Criterion B	1/3	1	1/5
Criterion C	1/4	5	1

The target weights are then determined from this matrix by determining the eigenvector for the greatest eigenvalue. In practice, however, it is seldom that perfect reciprocal matrices result after the weighting. In order to evaluate the consistency of a matrix, Saaty has defined a consistency index. This measures how consistent a matrix or a decision is. If the consistency index of a matrix is above 0.1, it is to be regarded as consistent. Saaty calculated the reference value of 0.1 from the consistency index for randomly filled matrices. Although there is criticism of the method, e.g. If, for example, the evaluation scheme is imprecise or inconsistencies can occur in the calculation of the consistency index due to the calculation method that has nothing to do with the consistency of the decisions, the method is widespread and accepted in practice.

Alternative definition

In the next step, various alternatives are defined that can be chosen. It is important that the “zero alternative” is also included in the assessment. The zero alternative describes the current status quo. This is required because it is always possible that none of the new alternatives will be more useful than the current state. An exclusion of the zero alternative in the utility value analysis should only take place if the actual state meets one of the previously defined exclusion criteria.

Evaluate the alternatives

After weights have been assigned to the criteria and various alternatives exist, the criteria of the various alternatives must now be evaluated. In order to be able to present the multi-attribute evaluation of an alternative as a one-dimensional utility value, the evaluation of the criteria must be done on a scale. The type of scale depends on the criterion to be assessed. A distinction is made between the following scales:

Nominal scale
Results that are displayed on a nominal scale can only be assessed in binary form. That means you only differentiate whether a criterion applies or not. Typically, nominal scales are used for classification, for example to summarize according to gender or color. In principle, it is not possible to use a nominal scale for a utility value analysis, since no qualifying statements about the characteristics of the characteristics are possible. However, the exclusion criteria can be displayed on a nominal scale in order to decide whether such a “KO criterion” is applicable.
Ordinal scale
The ordinal scale makes it possible not only to state whether two characteristic values are equal or unequal, but also to determine whether the characteristic is greater or smaller than another. Typically, an ordinal scale is a ranking list. However, the ranks do not reflect how big the difference between two ranks is. You can imagine the result of a Formula 1 race in which you know who is first and who is second, but because of the allocation of places it is not clear how much faster the first-placed driver was compared to the second-placed driver. Because of this fact, the use of an ordinal scale in a utility analysis can lead to a bias of the results. In order to reduce this distortion as much as possible, evaluation schemes can be used that determine in advance which characteristic leads to which value on the scale. The advantage of an ordinal scale is that it is easy to carry out, especially when time pressure makes it impossible to use another measurement method.
Cardinal
Scale Data presented on a cardinal scale is based on measurements or counts. Another name for the cardinal scale is the “metric scale”. The values on the scale can be compared, subtracted and added together, making the cardinal scale the ideal scale for a utility analysis.

The cardinal scale can also be further subdivided into the interval scale , ratio scale and absolute scale , but this will not be discussed here.

As soon as a suitable scale has been found for all criteria, the criteria can be rated. After the assessment, all scales must be made comparable with one another. This is done either beforehand by z. B. always awards points from 0-10, or afterwards, z. B. with the help of transformation equations. The weighting depends on the preferences of the decision makers. In practice, the criteria weighting is often assigned directly, i.e. without a pairwise preliminary comparison. This is a great simplification and leads to a rather “general estimate” result, in contrast to an actual criteria analysis as suggested by the method.

Utility calculation

The next logical step in the utility value analysis is to calculate partial and total utility values from the weightings and evaluations. Büssow also calls this step a value synthesis. First, the weights of the criteria are multiplied by their evaluation. Then the sum of the partial utility values calculated in this way is formed in order to obtain the total utility value of an alternative. The following table shows the exemplary calculation of a total utility value.

Exemplary calculation of a utility value

The total utility value of alternative 1 is calculated as follows:

{\ displaystyle {\ text {Total utility value}} = 0 {,} 7 \ ​​times 80 + 0 {,} 3 \ times 70 = 77}

Or in general terms (where is the number of alternatives): ${\ displaystyle r}$

{\ displaystyle {\ text {total utility value}} (A_ {1}) = \ sum _ {r = 1} ^ {2} {\ text {weighting}} _ {r} \ cdot {\ text {evaluation}} _ {r} ({\ text {criterion}})}

As you can see, this representation corresponds to the additive multi-attribute value function. Based on the calculated total utility values, a decision can now be made as to which alternative one should choose. In this case, alternative 1 would be preferable to alternative 2 because 77> 64. In addition to the decision for the alternative with the highest total utility value ("addition rule with absolute scale fixation"), there are other methods of choosing an alternative, including:

The Simon Rule

Here all alternatives are divided into two classes. One class contains all alternatives that have at least a certain utility value, the other class all the others. This method is useful, for example, to make a preselection of alternatives.

The majority rule

If the majority rule is applied, alternative A is preferable to alternative B if alternative A has a better partial utility value than alternative B for at least 50% of the criteria.

The Copeland Rule

When applying the Copeland rule, each alternative receives a plus point if a part-use value is higher than a corresponding part-use value of the alternatives. If it is lower, the alternative receives a minus point. The alternative with the greatest number of points is the better in the end. The problem with this method is that it is only considered whether one criterion is better than another - not how much better.

The ranking sum rule

For each criterion, ranks from 1 to x are assigned based on the partial utility values. The ranks are then added for each alternative. The best alternative is the one with the lowest total from the ranks.

Sensitivity analysis

After the utility value calculation or the value synthesis has been completed and a result is available, the question often arises in practice how resilient or robust the result is. A sensitivity analysis is carried out to clarify this question.

Sensitivity analyzes measure the effect of changing an input variable on the result

Example of a simple utility analysis

A free scaling of the degree of fulfillment and the weighting factors, for example between 0 and 9, is typical for the simple utility value analysis:

for "bad" points 0–2 ,

for “ medium ” points 3–5 and

for “good” points 6–8 and

Point 9 allows for “ very good ” .

An example with any weighting could look like this and first set the pros and cons of each sentence in writing, and then multiply with the weighting to arrive at the result of this option. The same table is created for each additional option. In the end, the highest result is the optimal choice:

criteria	Degree of fulfillment applicant	weighting	Result / value
Expertise	5	× weighting factor 9	45
work experience	7th	× weighting factor 6	42
Willingness to educate	3	× weighting factor 8	24
Spatial mobility	2	× weighting factor 7	14th
Temporal flexibility	3	× weighting factor 5	15th
Relationship network	8th	× weighting factor 9	72
Leadership skills	4th	× weighting factor 4	16
Presentation skills	4th	× weighting factor 7	28
Testimonies	3	× weighting factor 4	12
sympathy	7th	× weighting factor 6	42

The individual value of this applicant for the company amounts to a total of 310 points. In comparison with the other applicants, the personnel decision can be prepared objectively.

Comparative product test

Most testing organizations use the comparative product testing for the description of objective subjective criteria for the evaluation and for the consolidation of individual criteria with average values as well as a clear and fast to be detected representation vote numbers and symbols on a five-fold scale International is based on an integral evaluation number of one to five establishes a continuous scale from 0.5 to 5.5, with a higher rating number representing a better rating.

The following table shows some examples:

Integer valuation	Moving valuation	ICRT		Stiftung Warentest
International				Germany
Rating number	Valuation speed range	symbol	semantics	symbol	semantics	Rating range (grading)
5	4.50-5.50	+ +	very good	+ +	very good	0.5 - 1.5
4th	3.50-4.49	+	good	+	Well	1.6-2.5
3	2.50-3.49	O	sufficient	O	satisfying	2.6-3.5
2	1.50 - 2.49	-	less sufficient	ϴ	sufficient	3.6-4.5
1	0.50-1.49	- -	bath	-	inadequate	4.6-5.5

Hierarchical determination of the goals

A somewhat coarser method works with the specification of the target system in the form of approximate values “better than / worse than” required. The procedure must be strictly hierarchical, otherwise it cannot be calculated. Different criteria can be defined in order to exclude some alternatives in advance.

"KO criteria" (must criteria): minimum / maximum condition, the fulfillment of which is mandatory
Target criteria whose fulfillment as far as possible is desirable

You define evaluation criteria that are to be used for the evaluation. It is only about the most important criteria that should ultimately lead to the decision and not all that are known.

criteria	fulfilled yes / no	weighting	Result / value
		× n
		× n
		× n
		× n
		× n
		× n
		× n
		× n
		× n
		× n

In a more precise breakdown, the criteria are first determined by comparing them in pairs, that is, by considering “Is criterion A more important than criterion B”?

If a criterion is less important, it gets zero points
If one criterion is equal to another, it receives a point
If one criterion is more important than the other, it receives a score of two.

This structure gives a more precise result than the simple question “fulfilled / not fulfilled” and, through the pre-weighting, leads to a mathematically useful result.

The alternative with the highest utility value is therefore ranked 1, so it represents the best selection. However, it should always be noted that the utility value analysis provides a comparative result, i.e. it cannot provide an absolute statement about the benefit.

If the outcome is very tight (e.g. in price), additional criteria can be used, such as B. the time of the last price increase or the advisory service.

Experience shows that the rule “less is more” applies to the number of criteria; it makes sense to concentrate on a few concise points. On the one hand, the workload increases the more criteria are to be compared; on the other hand, the comparison becomes increasingly difficult. Three to five criteria are recommended, more than ten are not recommended in practice.

Furthermore, the following four points must be taken into account when selecting the evaluation criteria:

Operationality: Evaluation criteria must be precisely described and measurable.
Relation to hierarchy: Evaluation criteria that belong to a common category are to be arranged together.
Diversity: Different evaluation criteria must describe different characteristics.
Utilization independence: The fulfillment of one criterion must not presuppose the fulfillment of another.

Conclusion

In summary, the following should be observed when creating the utility value analysis in order to make the decision selection and the decision process transparent:

The goals must be relevant and correctly formulated.
The scaling of the ratings can be imprecise for ordinal scales.
Ratings and weights can change in the future.
Ratings and weightings can be (strongly) subjectively influenced.

The advantage that the benefit analysis offers is not only based on the better transparency and traceability of the decision-making. It also lies in the fact that the criteria and arguments that ultimately determine a decision are carefully examined. This often leads to new insights during the decision-making process.

Concentrating on the really decisive factors creates clarity. Based on the numerical representations, a comparability is also established that is not possible without this method. In this way, “gut decisions” are significantly reduced.

This form of reasoning can also include emotional factors such as well-being, weather in the new location, or even sexual attraction. This utility analysis is the most complex deductive argument .

Similar methods

Analytic Hierarchy Process by Thomas Saaty. The AHP is mathematically more demanding and more precise, requires a pairwise comparison even with the alternatives and also measures the logic and quality of a decision via the inconsistency factor .
Analytic Network Process by Thomas Saaty. The ANP allows the creation of decision networks.
Cost-benefit analysis
Peren-Clement-Index by Franz W. Peren and Reiner Clement. One of two established indices for country risk analysis is used to assess country risks in direct investments.
Price-performance model by Prof. Heinz Lothar Grob eliminates structural defects in the utility value analysis such as B. the constancy of the marginal utility, independence of the part utility, dimensionlessness of the target value and the globalization of the criteria.
QFD Quality Function Deployment
TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) by C.-L. Hwang and K. Yoon. In TOPSIS, an alternative is evaluated by measuring its distance to the best and the worst alternative.
VDI guideline: VDI 2225 sheet 3, construction methodology - technical and economic design - technical and economic evaluation
Value analysis

Comparison of utility analysis and analytic hierarchy process

Pen and paper are sufficient to calculate the utility value analysis (NWA). Therefore, the NWA was already used at times when there was no IT. The method of the Analytic Hierarchy Process (AHP) is mathematically based on an iteration of matrix multiplications (see matrix ). This required computing power, which in practice was only successfully available to the AHP from 1990, with the beginning of the computer age. The NWA, on the other hand, is only an additive approximation method and is satisfied with the basic arithmetic operations .
With the NWA, in contrast to the AHP, the ranking of the criteria is not determined by many users by comparing them in pairs (not “every criterion with every other criterion”). Instead, many users manually enter their estimated percentage directly in the ranking table of the criteria (see above, weighting of the goals). In these cases, the “methodology” of the NWA is essentially reduced to the fact that the sum of all weighting factors must not result in more than 100%.
But even with "correct" application of the NWA, only a very narrow scale with a narrow range from 0 to 2 is available for the paired evaluation of the criteria for the point values, in contrast to the AHP, which has a much larger range (1-2 -3-4-5-6-7-8-9) allows significantly more differentiated evaluations. Evaluations with a larger bandwidth would not even be able to be handled with the NWA due to the simple mathematics (only basic arithmetic operations).
In the NWA, the ranking of the alternatives is basically determined without a pairwise comparison. The Analytic Hierarchy Process, on the other hand, “forces” a pairwise comparison and reflection, even with the alternatives.
In contrast to the AHP, the NWA cannot check the consistency of a decision based on the subjective evaluations. Due to the simple mathematics, there are a relatively large number of mathematical deviations in the NWA in practice, which users can modify depending on their personal taste or specific question.
The NWA requires the conversion of the hard criteria (e.g. euros, km, kg) within an additional auxiliary table for the creation of the “target achievement factors” (see allocation of points for the variants). With the AHP you can enter the ratings directly without this detour.

Individual evidence

↑ G. Westermann, S. Finger: Cost-benefit analysis. Introduction and case studies. (= ESV basics ). E. Schmidt, Berlin 2012.

↑ Christof Zangemeister: Benefit analysis in system technology - a methodology for multidimensional evaluation and selection of project alternatives. Dissertation. Techn. Univ. Berlin, 1970. 4th edition. Wittemann, Munich 1976, ISBN 3-923264-00-3 .

^ H. Nollau: Business process optimization in medium-sized companies. (= economy and labor. Vol. 5). 1st edition. Eul Verlag, Lohmar 2004.

↑ F. Eisenführ, T. Langer, M. Weber: Rational decision making. (= Springer textbook ). 5th, revised. u. exp. Edition. Springer, Berlin 2010.

^ AD Hall: A Methodology for Systems Engineering. Princeton, New York 1962.

↑ ^a ^b Markus Bautsch: Usability and practical value. In: Tilo Pfeifer, Robert Schmitt (Hrsg.): Masing handbook quality management . 6th, revised edition. Carl Hanser Fachbuchverlag, Munich / Vienna 2014, ISBN 978-3-446-43431-8 , chapter 35.

↑ N. Zardari: Weighting methods and their effects on multi-criteria decision making model outcomes in water resources management. (= SpringerBriefs in Water Science and Technology ). 2015, ISBN 978-3-319-12585-5 .

↑ N. Zardari: Weighting methods and their effects on multi-criteria decision making model outcomes in water resources management. (= SpringerBriefs in Water Science and Technology ). 2015.

^ R. Fiedler: Controlling of projects. Project planning, project management and project control. [with online service for the book]. 2., verb. and exp. Edition. Vieweg, Wiesbaden 2003.

↑ A. Ishizaka, P. Nemery: Multi-criteria decision analysis. Methods and software. 2013.

↑ ^a ^b T. Saaty, L. Vargas: Models, Methods, Concepts & Applications of the Analytic Hierarchy Process. (= International Series in Operations Research & Management Science. 175). 2nd Edition. Springer US, Boston, MA 2012. doi: 10.1007 / 978-1-4614-3597-6 .

↑ K. Nagel: Benefit of information processing. Methods for evaluating strategic competitive advantages, productivity improvements and cost savings. 2., revised. and exp. Edition. Oldenbourg, Munich 1990.

^ W. Hoffmeister: Investment calculation and utility value analysis. A decision-oriented presentation with many examples and exercises. Kohlhammer, Stuttgart / Berlin / Cologne 2000.

↑ Horst Dürr: The overall verdict in the comparative product test - structure and accuracy. Chapter 2: Assessment Scales.

↑ IOCU Testing Committee: Guide to the principles of comparative testing. 1985, Chapter III.5: Ranking scales

^ Hans-Dieter Lösenbeck : Stiftung Warentest. A review. Chapter 6: The changing methodological foundations. Berlin 2003, ISBN 3-931908-76-3 , p. 103.

↑ see also external example

literature

Christof Zangemeister: Benefit analysis in system technology - a method for multidimensional evaluation and selection of project alternatives. Dissertation. Techn. Univ. Berlin, 1970, 4th edition. Wittemann, Munich 1976, ISBN 3-923264-00-3 .

RL Keeney, H. Raiffa: Decisions with Multiple Objectives; Preferences and Value Tradeoffs. John Wiley & Sons, 1976, ISBN 0-471-46510-0 .

Arnim Bechmann: Benefit analysis, evaluation theory and planning. 1st edition. Haupt, 1978, ISBN 3-258-02694-7 .

G. Bamberg, A. Gerhard Coenenberg, M. Krapp: Business decision-making. (= Vahlens short textbooks ). 15th edition. Vahlen, Munich 2012.

N. Bhushan, K. Rai: Strategic decision making. Applying the analytic hierarchy process. (= Decision Engineering ). Springer, London / New York 2004.

C. Büssow, H. Baumgarten: Process evaluation in logistics. Key figure-based analysis methodology to increase logistics competence. (= Gabler Edition Wissenschaft Logistik-Management ). 1st edition. German Univ.-Verlag, Wiesbaden 2004.

F. Eisenführ, T. Langer, M. Weber: Rational decision-making. (= Springer textbook ). 5th, revised. u. exp. Edition. Springer, Berlin a. a 2010.

R. Grünig, R. Kühn: Decision-making procedures for complex problems. A heuristic approach. 4th edition. Springer, Berlin / Heidelberg 2013, ISBN 978-3-642-31459-9 .

P. Guarnieri: Decision Models in Engineering and Management. (= Decision Engineering ). Edition. Springer International Publishing, Cham 2015.

W. Hoffmeister: Investment calculation and utility value analysis. A decision-oriented presentation with many examples and exercises. Kohlhammer, Stuttgart / Berlin / Cologne 2000.

A. Ishizaka, P. Nemery: Multi-criteria decision analysis. Methods and software. 2013.

K. Nagel: Use of information processing. Methods for evaluating strategic competitive advantages, productivity improvements and cost savings. 2., revised. and exp. Edition. Oldenbourg, Munich 1990.

H. Rommelfanger, S. Eickemeier: Decision Theory . Classic concepts and fuzzy extensions; with 109 tables. (= Springer textbook ). Springer, Berlin / Heidelberg a. a 2002.

G. Westermann, S. Finger: Cost-benefit analysis. Introduction and case studies. (= ESV basics ). E. Schmidt, Berlin 2012.

N. Zardari: Weighting methods and their effects on multi-criteria decision making model outcomes in water resources management (SpringerBriefs in Water Science and Technology). 2015.

Web links

Encyclopedia of business informatics; Frank Bensberg: utility analysis

[1] G. Westermann, S. Finger: Cost-benefit analysis. Introduction and case studies. (= ESV basics ). E. Schmidt, Berlin 2012.

[Zangemeister-2] Christof Zangemeister: Benefit analysis in system technology - a methodology for multidimensional evaluation and selection of project alternatives. Dissertation. Techn. Univ. Berlin, 1970. 4th edition. Wittemann, Munich 1976, ISBN 3-923264-00-3 .