Portfolio Architecture Selection
Good morning. Welcome back to the discussion on product architecture. So in the last class, we
discussed various portfolio architecture. We found that there are different types of architectures.
Basically, we have unshared platforms, modular platforms. In modular itself we have modular
generation, modular family, etc. So, the question is how do we actually decide the portfolio for a
particular product? And then in the last class we stopped here.
We mentioned that it basically depends on the market dynamics, what the customer population is
looking for and how their preferences change and what are the variations of their preferences
within the group itself. So, looking at all those aspects and of course, the cost and revenue, we
need to decide the architecture. To do that, we use an architectural selection flowchart and
looking at the flowchart and then deciding, what would be the best architecture for a particular
product can be decided. So, that is what we are going to discuss in this class. What are the ways
in which we can decide the product portfolio architecture?
So, the flowchart is given here, we need to identify the customer needs. As we know, first we
look at the customer needs and prepare a list of customer needs with their importance and then
look at those important needs and then find out what is the mean and standard deviation of these
customer needs in terms of their objective values within the customer population.
So, the total population we take, suppose there are the 1000 customers you identified, for this
thousand customers, probably there may be three or four needs that you can take initially and
then find out, for example, cost, the size and capabilities or processing capability like that you
can identify those needs and then find out what is the mean and standard deviation of these
values, these values of this customer needs, that is the first step, so get a customer population
Normally when you take a distribution you always will be following a normal distribution, so we
will be getting something like this kind of a distribution. Here we will be having a mean value,
the mean of the population. And then we can have a sigma that is the standard deviation of that
particular need. So, for every need you will be able to find out a mean value and a standard
deviation that is for the old population.
Suppose there are 1000 customers, you want to survey for all the 1000 customers, you try to find
out what is the mean and standard deviation for each need. And then choose typical customers
from different segments. So, this is another important point, we need to identify segments within
the population. If there are a 1000 people in that group, it will not belong to the same kind of
segment. For any product you will be able to segment the group, segment the total population
based on some characteristics.
So, the characteristic would be the age or the gender or their buying capacity or their occupation.
There can be many ways you can actually segment the population. We will be able to identify the
low income group and high income group or the student population and the employed population
or male and female or all the elderly people and young people so this way you will be able to
identify segments. You need to identify the segments based on how much affinity they have
within that group and how much they differ from the other groups also.
If there is a lot of difference between these two groups, then that makes them 2 segments or 3
segments, depending on the number of segments that you can identify. Try to identify segments
and then choose typical customers from these segments and trace these values over a period of
time. This is a bit tricky because you want to know how these customers within the segment
change their preferences over a period of time.
A young person who is looking for a phone today will not be having the same kind of preference
after 1 year or 2 year, so we need to have some idea about his preferences and how his
preferences are changing over a period of time. We will get this only through some kind of
observation of the market or looking at the previous generation product in the current generation
and how people change their preferences or asking some questions to the customers.
So, doing this kind of an analysis, you will be getting some information about how these values
change over a period of time. If sigma and μ have a large variation, then you need to think of a
different architecture. If there is no variation over a period of time, then the architecture will be
different. That is the next step to identify the values or a period of time.
And then calculate the mean and standard deviation across users for each need. So, for each need
you need to identify the customers mean and standard deviation, and the last one is, so for each
need, create tables of means and deviation for segment usage, so this one was the population
usage. Now, within the population we identify 3 segments, segment 1 with maybe low income
people, segment 2 can be medium income and segment 3 can be a high income group.
So, we can actually divide the whole population into 3 segments. And then within the segment,
you try to find out what their mean and standard deviation are. So, a low income group will be
having some values for the cost and the size, high income group will be having a different value
for the cost. So, we need to find out what is the mean for each segment, μ1, μ 2, μ3 etc.
For each segment we try to find out the mean and standard deviation for each need. So, we will
be getting a table like this. You will be having the population mean and population sigma and in
addition to that, we will be getting the mean and standard deviation for different segments also.
Now we have a lot of information collected through customer surveys. You have the population
mean and a population standard deviation for all the needs.
Then we have the variation of these values of μ and Sigma over a period of time for each need
and for segments also, we have this μ and sigma for each need. So, with all this information we
can actually choose an architecture based on this information. We use an algorithm to do that.
So, this is the flowchart of selection. Here we use a very simple algorithm to choose the
After once you have all this information, we go to this algorithm and then try to see how to select
the portfolio architecture. The algorithm is like this, you determine the population and segment
statistics, as we mentioned in the previous slide, the statistics, basically the μ and standard
deviation, it is the mean and the standard deviation for all the needs for population, segments and
the time variations, or the variation over a period of time.
Now you will ask the question, is population mean fixed with respect to time? That is, is there a
variation of the population mean over a period of time that is, today the population mean says
that the cost of a mobile phone should be 15000. That is an example you take, that is 15000 is the
cost that people are ready to buy to prepare for a phone. Now if this 15000 remains the same for
next year also or there will be a lot of changes in the next year.
If they say that there is no change in the next year or after 5 years or after 1 year they will be
ready to pay 20000 and next year they will be ready to pay 25000 thousand. In such a case, we
know that the mean actually changes over a period of time. So, if there is a change over a period
of time, that is the mean and the population mean fixed with respect to time. Now, this is the
Suppose there is a change, then mean and the standard deviation are not fixed or the mean alone
is not fixed, i.e., the mean value of that particular need changes over a period of time then we
should go for a modular generation platform.
The mean value of the need is changing over a period of time which means that the customer
preferences changes over a period of time. Then we will go for a modular generation platform.
So, this actually comes from the previous analysis. We had an analysis of the mean value over a
period of time and if there is a change over a period of time, we go for a modular generation
So that you will be able to modify the current product for the next generation without making too
many changes, so only a particular module or 1 or 2 modules need to be changed so that it will
become the next generation product. We are looking at which need is changing and the module
corresponding to that can actually be modified to get the new product i.e., the modular
Now suppose it is fixed, there is no change in the mean or change is very small, not very
significant then we can look at the sigma value. So, you know that the distribution can actually
be represented using a mean and standard deviation. So, this is the μ and this is the sigma. Now
you look at the population sigma, large, is the population sigma very large? So it can actually be
So, the distribution can be like this also, that is, you still have the same mean but the sigma can
be small. In this case the sigma is very small. We see that the sigma is very large. When you say
sigma is very large, that means there is a lot of variation, though there is a mean value, but the
requirement of the customer varies widely. So, that is what is meant by a large standard
deviation. The sigma is large means the population has varied requirements. So, that cannot be
said that, everyone is looking for the same kind of a feature, there is large variation within the
Then if there is no large variation, the variation is very small then we can go for a fixed unshared
portfolio. So, you do not really need to go for a modular architecture because everyone is asking
for the same thing and there is not much variation in the requirements of the customer, then we
do not need to have too many products in the market so we can actually have a fixed unshared
Because the requirements of the customer are almost the same and there is a large market for
that, then we can go for a fixed unshared architecture. Now suppose there is a large variation, the
sigma is very large, and then we look at it because when we say that large variation is there, that
means there are people with various requirements. Then we look at what their segment sigma is
and how each segment is performing.
Now we will be having multiple segments in the group so we can have out of this total
population. We can actually find out different segments. And then we look at each segment, what
is the μ for each segment. So, this may be one segment and this may be another segment and this
may be another segment, so each one will be having a different mean and standard deviation.
And that is why we get total population the sigma is very large and therefore do segments sigma
equal to population sigma.
If segment sigma and population sigma are the same that means, the variation within the
population and within the segment are the same then we will I mean, if that is a large variation, is
there in population segment and segment sigma, that means even if you take a segment within
the segment also there is a large variation so 1000 people you take, there is a large variation, and
100 people also you take people are not happy with the similar thing, they also have large
variation that means there is too much of variation within the whole population.
So, if there is a large variation, then we go for an adjustable portfolio architecture because one
product or a few products will not satisfy the people because each segment has again a large
variation. So, one product or a few products will not satisfy you to go for an adjustable
architecture that means you can adjust the architecture depending on the customer requirement.
Again, we have a large variation possible and that is not there if the segment sigma is not equal
to population sigma, then we go for a modular product family. So, a modular product family will
be preferred when the population sigma is very large, but segment sigma’s are not very large,
segment sigma’s are small that means you can actually identify segments where the variations
are very small.
Then for each segment we will be able to produce products. That is basically we go for a
modular product family. So, we can have few members in the family and each segment, each
product will be designed in such a way that it can actually meet that requirement of a particular
segment of customer. So, that is the way we decide the portfolio architecture of our product, I
hope you got it.
So, just to explain once again, we look at the population mean, first we get all the information
then see whether the population mean is changing over a period of time. So, if the population
mean is changing over a period of time, it is clear that you need to go for a modular generation
platform because today's product will not satisfy the customer in the next year or after 2 years so
we need to have a modular generation platform.
And there is more variation then of course, you need to think of modular family products or
modular architecture. Then we will see, look at the sigma, population sigma is very large. If the
population sigma is very large, that means there is large variation within the population itself,
then you cannot go for a single product or if the variation is very small, every all the people or
most of the people are looking for a particular feature or a particular value in that feature then we
can actually give that one only one product and then that can be made in large numbers.
Then we will go for a fixed unshared portfolio. The population sigma is not large. They are
actually very small sigma, variation is very small then we go for a fixed unshared portfolio.
Now, if the population sigma is large and the segment sigma is also large, that means there is
large variation within the population and within the segment also, that means no product is good,
no single product or not a few products are going to satisfy the customer.
In that case, we have to go for an adjustable portfolio architecture that means large variation
within the group and within the segments. Therefore, we cannot go for a modular family, we
need to go for an adjustable portfolio. And the segment variations are very small compared to the
population variation, then we can go for a modular product family. So, this is the way we can
actually identify the portfolio architecture for a product based on the customer's requirements. I
hope you understood. We will take 1 or 2 examples and then see how we can apply this logic to
get the portfolio selected.
We will take the example of a very simple toaster, to toast bread especially a lot of use for such
products. You can see that people will be buying a toaster for making breakfast most of the time.
And suppose you are after completing your course, you go for a job and then you are staying
alone or you are staying with your friends in an apartment. You would like to have breakfast in
the morning, prepare breakfast and go to your work.
And the easiest thing is to have bread and I mean toast bread and then have bread and butter or
something and then use it. So, you will be looking for a toaster so that you can actually quickly
get your breakfast done. So, your requirement here is that you are not looking for a very high
quality, good looking toaster and looking for something which will actually meet your
requirement just to meet the requirements. You are not going to use it for a long time and you
may throw it away after 1 or 2 years also.
You do not want to spend too much money for that and you are not worried about the quality of
toasting and that somehow we do have to get it toasted, so that is may be your requirement as a
bachelor and or staying with your friends, but you take the case of a housewife who is actually
having a husband and the 2 kids at home and she wants to make bread at home and for breakfast
and she is also looking for a toaster.
And her requirement will be that she wants it to be a good one, good quality. She does not want
to replace it very often. She does not want to repair it very often. She is looking for a very
reliable, good quality product which will do good toasting also. So, her requirement is that even
if it is slightly costly, she doesn't mind. She is looking for a good quality, good product. Maybe
the number of slices 2 may be sufficient because the requirements are also not very huge.
And you take the same case of a hostel or a restaurant. They also need a toaster because they
want to serve breakfast to the customers. Their requirement is that they would like to have a
large number of slices toasted in a short time. So, the requirement is they want a number of slices
to be toasted more and they do not mind paying money for that because they can always make it
from the business. So, they do not worry about the cost.
They do not worry about the size of it. They can also be big. They would like to have it fast and
good quality and number of slices also more. So, now if you look at the requirements of the
customer population of toasters, you can actually see that there is a large variation. So, bachelor
should be looking for something, Housewife would be looking for something, mess or
restaurants would be looking for something.
Now, we need to see how we actually decide what kind of a toaster can be brought into the
market or a company interested in introducing a toaster in the market. They want to know what
would be the architecture of the product, which will actually meet that customer satisfaction. So,
what they do is prepare a customer needs list. So, what are the important things the customer is
looking for in this toaster?
So, this is from the point of view of the user's point of view, as well as for the designer to
understand the requirements. So one thing, maybe the cost definitely cost maybe one important
criteria. The second maybe the number of slices it can toast, the third maybe the quality of the
toast and the quality of the product or the output coming from the product and then maybe the
size, maybe another criteria.
So, we need to identify all those customers’ needs and then find out the variation of these needs
also. Okay, so now we can take that is a number of slices it can toast is one requirement,
inexpensive, compact and uniformly toasted result that is the quality of the toast. So, this can be
considered as the need that you would like to take into consideration when we decide on the
architecture and the weightage, which one is more important, which one is less important, so 4 is
considered to be more weightage and 3 is less weightage like that.
Now we look at, how is this varying over for the couple population, customer population? So, we
are supposed to have 1000 people interviewed. You want to take the population and then find out
what is their requirement, each one. So, what is the requirement of the number of slices? What is
the cost they are ready to pay? And what is the size they are looking for? And what is the quality
of the toasting they are looking for?
So, this is what is needed from the market. And in this case, because a toaster is not something
which actually changes its technology for a very short time, so we may not be really looking for
the changes over a period of time. So in this case, it is not really needed because Toasters
technology or the customer's requirement of a toaster is not going to change rapidly. Therefore,
we do not consider the modular generation in this case and therefore we do not really look into
the variation of mean over a period of time.
If you feel that a product requires that to be considered, then we need to definitely take that also.
Now, identify the market segments so that is the next term, so we have the population and we
need to identify the market segments. So, as I mentioned, the market segments can be considered
into 3 or 4 segments, depending on how you want to divide the total population.
So, here I can say that those who are looking for a low quality product or a low income group or
as I mentioned, bachelors and those who are staying in hostels or I mean, rented apartments, they
will be looking for some product and housewives be looking for something and restaurants will
be looking for something so we can actually divide this group into mainly 3. First we call it a low
And then the other one is the high end segment that those who are actually looking for a good
quality product, they are ready to pay a high price for that, that is the high end segment, mainly
the housewives and family people. And then the high capacity segment, high capacity segment is
the restaurants and mess and hostels and places like that. So, we can actually identify there are 3
segments in this case.
Now, what we need to do is to survey the customers for the needed target values. So, for each
customer, we try to find out what their target value is. For example, you look at you ask all the
people, so what is the price you are ready to pay? For low price segment, high end segment, high
capacity segment, then you find out the total population also and find out what is the price they
are ready to pay and what is the number of slices they are looking for and what is the size they
are looking for and what is the quality of toast they are looking for.
And then survey segment for time variation, so if needed you can survey it for time variation.
But in this case, it is not very critical. We do not really need to do but in general case we need to
do the time variation also and calculate the mean and standard deviation across users what each
need for each need you try to find out what is the mean value and standard deviation and then
decide on the portfolio based on the analysis. Then you use the architecture that we discussed
and then decide what should be the portfolio to be used. In this case, let us see how we can do
this?We will see that, this is the need, and then we will see the population, the total population will
try to find out the mean and sigma that is the mean value and standard deviation and then we
look for each segment, so the low price segment, high end segment and high capacity segment.
For each one will try to find out mean and sigma values. And this we need to do for each need so
let us take one need first need and see that number of slices it can toast.
We are saying that we can give an option of 1, 2 or 4. I mean, assume that you already suggested
1, 2, 4 all the customers based on their own requirement, but maybe we can identify the total, we
can get from the total population. And we found that the population is actually looking for 2 and
plus or minus 1. So, that is generally the whole population is asking for slice 2, so there is a large
need for 2 slice toasters and the variation is not that high, but of course it is plus or minus 1.
Somebody says oh one is sufficient for me, somebody says no, I would like to have 3 or 4. So,
average if you take it, you will see that it is 2 ± 1.
If I plot it, I can actually say that it is like this. So, 1, 2, 3, 4 like this, so I can tell that this is the
2, this is the population mean, and standard deviation. Now you take the low price segment. So,
the low price segment also says the same 2 and ± 1, so the low place, low price also is almost the
same 2 ± 1. And the high end segment says, the high end segment is the housewife and the
family people, so they are also telling 2, their mean is also 2.
But then their sigma is very, is 0 that means they are actually saying that this is our requirement,
that is the 2, but the standard deviation is 0. So, it is almost the same as a straight line here. So,
that is the requirement. So, here they have 2 and 0 as the mean and standard deviation. Now,
you come to the high capacity segment, they say that our requirement is 4 here sigma is almost 0.
It is almost straight, so they want only 4. They do not want anything else. So, we can see the high
capacity segments says their requirement is 4 slices and high end segments says that their
requirement is 2 but low price says that we are ready with the 2 ±. We do not really care about
these 3, 4 or 2. And their requirement is, I mean they are worried more about something else.
So, this is the case for the number of slices. And overall population is two plus or minus one. So,
we can see that the low price and high end will be happy with 2, but high capacity will not at all
be happy with the 2 slice one. They don't really want to slice. They want only a 4 slice product
which is very obvious because they are into serving more number of customers and they would
like to have high capacity for the product.
So, now let us look at the other ratings. Now look at the price. So, now the price again, you can
see the variation, so we can, suppose you had suggested you gave for 5000, 2500, 1000. Is that
the range you suggested to the customer or the customer, based on the customers you can get the
value but in this case, assume that the population, we can get this from the service. The
population mean is 2500 with ±500 rupees.
So, suppose this is 1000, 1000, 2000 etc. 3000 to 4000 5000. So, the population mean is around
2500, ±500, so their requirement is something like this. So, this is the population mean, but the
low price segment, the low price segment says that they are looking for only 1000 things. So,
there are 1000 ±200. So it is almost like this. They have a 1000 as the requirement.
So, this is a thousand ±200. That is for the price of the product and here high end say they are
looking for 5000. They do not mind paying a high amount for this. So, they have this around
5000. So, their requirement is like 5000±. So, they do not mind paying large money for this. So,
their requirement is like this 5000 ± that is for the high end segment.
And high capacity also saying the same thing so, the high capacity segment also telling this is the
requirement. Now you can see there is a clear differentiation between these two. So population,
overall population if you take the mean values is 2500, but the low price is asking for a 1000
rupee product, high end is asking for 5000 and high capacity is also 5000 rupees. So, there is a
difference in the requirement of the customer in this case that is the price.
Now look at the other things about the size. So, if you say that they can provide 2 I mean, 3
different sizes, one is a large size 20 by 20 the other one is 6 by 20 and 6 by 10, one is very
compact 6 by 10 very compact is a medium size and is a large size. And you ask for the
customers. So ,population you can see that they are only 6 by 20 is the requirement in the middle
value and low price segment they do not really care what is a size 20 by 20 or 20 by whatever it
So, they are telling us we do not care about this 20 by 20 is also fine just then it becomes more
and more small, the cost may go up so they do not want to increase the cost so they are ready to
go for a non-compact type. But the high end segment, they are very clear they do not want to
spend too much space in the kitchen for this. They want a very compact product, 6 by 10
compact product in the kitchen. So, they are very clear they want a very compact product.
Now high capacity they do not mind going for a big size one because the capacity is more
important for them. So, they would like to have a high capacity one even if the size is big, they
don't really care. So, that is the way we can actually decide the size of the product. Now the last
one we will take about the uniform quality, quality of toasting.
So, the quality of the toasting we say that, low, medium and high. If I say low, medium, high as
the toasting quality. So, the population says that the medium quality is ok and the low price
segment also medium quality is ok but the high end says that it should be high quality and high
capacity also says it should be high quality. You like to have a high quality product to satisfy
these two customers and these customers say, I don't really care even if it is a medium quality
toaster because as I told this customer segment is not really worried about those things.
They are only very worried about the cost of the product that is why they do not really care even
if the quality is slightly less. But the other two segments really ask for a high quality product. I
mean the toast. Now, we have all this information with us now we need to take a call. How do
we actually decide the product, the number of products or if you, again, there are few other
things also need to be taken into account.
How many people are there in each segment or if the segment size is very small, then probably it
is not worth going for a product to satisfy that customer segment. So, looking at all those aspects
again, you can actually take a call on how many products to be there and what should be the
architecture of this product.
Now, if the company decides to have only one product in the market, assume that based on the
financial aspect, they decide to have only one product in the market. What should be the
architecture that they should choose for that product, what should be the slice, what should be the
price, etc.? So, if it is only one suppose they have only one option as a one product, then they
should try to go for a 2 slice product and a high quality one, not that compact, but uniform toasty.
So, this will actually cover these two segments very well, low price and high end segments can
actually do this because they are providing 2 slices and are high quality, so this high end segment
will be happy and not so compact. So, if they make it very compact, the price will go up. So,
they do not want to make it very compact so that the prices can still be suitable for the low price
Log in to save your progress and obtain a certificate in Alison’s free Diploma in Engineering Functional and Conceptual Design online course
Sign up to save your progress and obtain a certificate in Alison’s free Diploma in Engineering Functional and Conceptual Design online course
Please enter you email address and we will mail you a link to reset your password.