Wishing you a lovely weekend. Today´s subject is about the comparative example of using the BCG, the GE-McKinsey, the Shell Directional Policy, and the ADL Matrixes.

I would like you to please print the set of 14 slides. We will proceed to make the inferences (to infer means to imply, to deduct by observation and thoughtful analysis) of the comparison, (slide by slide) as of slide number 4. Watch each slide, and read with us, s’il vous plait.

Slide 4: Multiple definitions of variables. When building the BCG Growth/Share Matrix, this example teaches us the ambiguity and diversity of results caused by the multiple definitions of the factors that are taken into account when we are analyzing what to include in the X-axis and the Y-axis. The researchers of this example were explicitly unblemished by showing us 4 definitions of growth, and 4 definitions of market share. Each factor that we pick, has multiple definitions. In addition, if we consider a different duration of years (instead of 4 years, we pick 6 or 10 years), then the numbers are also different. The researchers also calculated the mean, the standard deviation, and the range of min-max for each of the 4 market share definitions and for each of 4 market growth definitions per business (SBU), by searching correlations. And the results were 240 answers, each with a different classification per quadrant. So, for SB1 number one, we have 16 different plotting points for that SBU and so on.
What can we infer from this slide? Definitions matter. We have repeated this idea, on and on and on. Business intelligence software sticks to one definition when it comes to variables. But many times, there is not only “one” definition for a variable, but several of them.

Slide 5. We can change the scale. Several matrixes, including the BCG growth/share matrix; are built using certain scales in each respective axis. If the scale is changed (let´s say if the framework by the authors is placed from low to high, researchers can switch it from high to low), but the quadrants are adapted too, and in consequence, the inferences have to be appropriately modified too.

Slide 6. For composite scales, the number of factors considered is variable and numerous. In this case, the Shell Directional matrix was built with a composite measure of the Y-axis taken from the PIMS PAR ROI report. When using databases, each of the 19 factors was related to data extracted from the Profit and Loss and Balance Sheet accounts. So here we are landing into accounting. With annual or quarterly accounting reports usually, all of us end up being lost because accounting needs cleaning and filtering before making it compatible for our portfolio analysis.

Slide 7. Oversimplification. A matrix of 9×9 cells (as Shell Directional Policy matrix) can be simplified to a 2×2 quadrant. But an extreme oversimplification can be self-defeating. To simplify an analytical tool to a point of causing misrepresentation, misconceptions or even mistakes may cause distortions, and in consequence, wrong decision making about the status of SBUs.

Slide 8. To weight differentially by a factor or to unweight? The GE McKinsey Matrix was calculated with 6 factors for the X-axis (competitive capabilities) and with 8 factors for the Y-axis (prospects for sector profitability). Any measure of axis can be formulated with equally weighted dimensions, but also with empirically derived weights. We have already discussed the subjectivity of weighting which depends very much on the experienced consultant. The position of each SBU depends on the ranking and weight assigned to each factor, so the results can be different, depending on the weight assigned.

Slides 10 and 11. Measure the stage of industry maturity and market position. The ADL matrix shows us the X-Axis with 4 stages of life-cycle stages: embryonic, growth, mature, and aging. But the axis can be modified according to the stage of the SBUs. In this specific case, we see how the researchers consider the growth and mature stages. They also switched the order in the X-axis scale: first, they positioned mature SBUs at our left side, and then the quadrant of the growth SBUs. In addition, the Y-axis (market position) starts with the high level of the market share at the bottom, and the low at the upper part of the scale. The ADL matrix was collapsed into a 2 x 2 matrix corresponding to low and high share and mature vs growth businesses.

Slides 12 and 13. Final Results. Let´s compare the first 4 matrixes that were adjusted to a 2×2 (four-quadrant categories). We have provided slides 12 and slide 13 with the results. If you skim-examine the matrixes, there are considerable differences in the classification of the 15 businesses. We will leave out the ADL model because no continuous scale is available for the life cycle dimension. Let´s compare the graphs, and look where each of the SBU (the little circles) is positioned:

The last table suggests very low correlations among the various frameworks. Only one business (number 12) appears to be falling into the high/high category and was consistently classified in quadrant IV by the four models. What about the rest of SBUs. The rest of the SBUs are not consistent in the classification by the four models.

Overall, what can we conclude from this example? Basically, each portfolio framework has been designed to emphasize different portfolio objectives.  Each X-axis and Y-axis was defined differently in each matrix in theory. And in this example, it was also defined with its own dimensions, changing scales, and using a PIMS ROI report for the external factors (industry). In consequence, however, the basics were respected from the theory to the practice, the mix of factors that are considered in the Shell Matrix, GE Matrix, and ADL was different in each of the cases. Is it congruent to compare when using different models based on different variables? No, it is not harmonious.

Before filling the models with data, a statistical test of the degree of similarity/differences design between the 4 models would have been desirable. It is crucial to examine the premises of portfolio analysis per each model, before filling it with data. Why? Because we end up with inconsistent results if we don´t use or don´t compare the same external and internal factors in between models. Only one SBU (number 12) felt in the same classified category. This tells us that the other 14 SBUs were sensitive to the specific variables chosen, definitions, cut-off points, weights, and oversimplification of the models which were reduced to a 2×2 matrix. As the researchers remark: “Hence, any strategic generalization concerning them is suspect and at best unstable”.

An integrated multi-model approach: From the latter example, it is desirable to integrate the 4 various models, to take advantage of their unique capabilities. Creating a hybrid model that allows a “tailormade” one that can test the sensitivity of the portfolio classification, will help to land into understanding the definitions of variables that suit each corporation, the ranking, weights of the dimensions, cut-off point rules, determine the variables which will be the base for the composites measures; and finally the classification of the quadrants.

The disadvantages of portfolio analysis will be conducted next Tuesday. Thank you.

Why did we choose Ed Sheran´s theme in our last post, Perfect? This song is pretty. We were captured not only by the lyrics, which are so romantic but also because of the melodic arrangement made by the Brazilian Allegretto Producoes musicians. Love romantic music is always cherished because its composition holds our heart with people to which we can relate and attach our feelings. The purpose of Eleonora Escalante Strategy when supporting music played with acoustic traditional genuine instruments is beyond the music per se. We believe that music played with real instruments is superior to music produced by machines because when humans play the instruments, the feeling and emotion of the musician are transmitted to the one that listens to it.

Song of today. Today I will leave you with a song that I hope could be included in my wedding repertoire with Alejandro Guillermo Lozano Artolachipi. Being heterosexual is a blessing. I cherish and celebrate that blessing daily. Train’s theme “Marry me” is a modern wedding piece that must be included in every nuptial party. We include the original video by Train and the United Guitar Players masterpiece (only guitars). Which do you like the most?

See you next Tuesday, with one of the latest episodes of this saga. We will publish 5 additional posts, and we are done.  As I mentioned, next year we will begin a new saga baptized under the name of “Bees at work”. With “Bees at work”, we will explore the multiple edges of “work”. I am preparing the outline with a heuristic-integral approach. In the past, we already explored that high levels of efficiency-productivity are not always the best output. This new saga goes beyond that. Time to live matters beyond work. More information to come. Stay tuned!.

Sources of reference utilized today:

1. Wind, Yoram; Mahajan, V.; Swire, D.  An empirical Comparison of Standardized Portfolio Models.

Disclaimer: Illustrations in Watercolor are painted by Eleonora Escalante. Other types of illustrations or videos (which are not mine) are used for educational purposes ONLY. Nevertheless, the majority of the pictures, images, or videos shown on this blog are not mine. I do not own any of the lovely photos or images posted unless otherwise stated.

Published by eleonoraescalantestrategy

Corporate Strategy Artist, San Salvador, El Salvador View all posts by eleonoraescalantestrategy