Application Portfolio Management for Beginners with SMART

May 19, 2015
Written by
Maher Dara
Maher Dara

Application Portfolio Management for Beginners with SMART

Everybody is familiar with the SMART abbreviation. It stand for Specific, Measurable, Acceptable, Realistic and Time bounded. But this blog is not about that kind of SMART. In this blog I will discuss how to apply another type of SMART within the context of Enterprise portfolio management.

Enterprise portfolio management is a powerful approach with lots of value for different users within the enterprise such as senior management, IT & project portfolio managers, program managers and enterprise architects. Many exciting developments are in progress in this field of which one is application portfolio management (APM). APM is about categorizing and analyzing applications within your enterprise and deciding on recourse allocations within the applications roadmap. Such analyses are supported by powerful Enterprise Architecture (EA) software products such as BiZZdesign Enterprise Studio. However, you might be using other EA software products that do not support such features. Or you might be starting the EA discipline within your organization and therefore decided to use free and basic EA software products, or no EA software products at all (yet). I would say a very rational and good approach – think big and act small – also for your EA software products. Then how can you evaluate applications within your enterprise without state of the art EA software products? The answer is by using SMART.

SMART

SMART stands for the Simple Multi Attribute Rating Technique and can be used to rate the value of things based upon multiple attributes. For example, rating the value of a new passenger car when you need to buy one. Let’s say I want to buy a new car and value several attributes that my ideal car should have. Each attribute can be expressed by a specific value. For example the purchasing costs of my car will be expressed by a monetary value (€) while the fuel consumption of my car is expressed by a ratio (litres per 100 kilometres).

Suppose I’m interested in five passenger car models. In order to select the best option I need to compare all models based on single quantitative value. But how can I define one quantitative value based upon different attributes with different values (monetary vs. fuel consumption)? I cannot simply add the values of purchasing cost and fuel consumption rate because they’re expressed in different units of measurements. Therefore I have to normalize them first. Normalization means mapping the values of an attribute to a normalized scale such as the 5 point scale low to high rating. The following illustration shows the attributes for my ideal car and their normalized values.

Normalization of attribute values

Illustration: Normalization of attribute values

But what if I consider the value of one attribute more important than the other? In order to do so I have to weigh each attribute to indicate its importance in respect to the other attributes. For my ideal car the weights of attributes are illustrated below. As you can see, passenger space is the most important attribute followed by acquisition costs.

Normalization of attribute values

Illustration: Weights of attributes

As indicated below I calculated the single quantitative value for each of the 5 car models. You can see that car model A has the highest score. By using SMART I have quantified the values of attributes for each car model. Calculating these values in a total score allows me to decide objectively which car model best fits my needs.

Simple Multi Attribute Rating Technique

Illustration: Weighted total scores

Applying SMART

Let’s apply SMART for application portfolio management. Many valuation criteria can be used for this. An often used analysis of applications is based upon business value and technical value. Both dimensions can be rated based upon several attributes. SMART can be used to rate those. In the following illustration the applications and the attributes of business value and technical value are highlighted.

Applying Simple Multi Attribute Rating Technique

Illustration: Business value and technical value attribute rating with SMART

After normalization, total scores for each application can be calculated. The total scores for business value and technical value can then be plotted in a bubble chart as highlighted below. Now we are ready to discuss the implications of our findings. Based upon our data visualization we can formulate relevant questions such as:

Why does the AFAS – accounting application have a high business value and low technical value? Is it technically outdated?
What are the reasons for a low business value for SAP BPC while its technical value is high? Is it not fully implemented yet?
How should we handle our applications within our technology roadmap based upon their business and technical values?

Business and technical values bubble chart

Illustration: Business and technical values bubble chart

Conclusions & Recommendations

  • In this blog I have highlighted how SMART can be used for application portfolio management.
  • SMART can also be used within a variety of other situations where weighted multi attribute rating is required. For example SMART can be used in combination with the balanced score card, a SWOT analysis and the calculations of performance indicators.
  • SMART can be used to evaluate applications within your enterprise.
  • SMART is technique that can be used when no dedicated EA software products are available.
  • Any spreadsheet application can be used to apply SMART.

You can apply SMART for APM through the following steps (illustrated below).

teps for application of SMART APM

Illustration: Steps for application of SMART

You might notice that applying SMART with your regular spreadsheet application works just fine as long as the number of the to be rated applications, their attributes and weights are limited. For an efficient way of valuing tens or even hundreds of Enterprise Applications, a dedicated EA software product is necessary.