Model Driven Analysis – A Way To Work Smarter

I was surfing the internet and stumbled upon a book written by Gabriel A. Moreno and Paulo Merson on “Model-Driven Performance Analysis” (find link here). 

As the author describes the best approach to take towards designing a modal for a better performance tuning in the given product development lifecycle, is creating small sets or even subsets of a modal and ultimately generate an implementation of their results in targeted platforms. This approach is necessary to support the analysis of the design with respect to multiple quality attributes and needed business logic without sacrificing performance on any required platform.

To complement the model-to-implementation path of MDE approaches, an MDE tool infrastructure should provide what we call model-driven analysis. – Gabriel A. MorenoPaulo Merson

Understanding PIM…PIM is a Platform Independent Model, it is suggested while designing MDA (Model driven Architecture) we make sure given modal or set of functionality is platform independent. The base use case for this can be if the particular function works differently on the app and differently on the website, it might confuse end user and create a restrict barrier. World known example to this is most of the people prefer booking uber via an app rather than giving a try to their web-based site, reason different PIM.

PIM can be language independent. Most of the technologies we know today are not based on a particular targeted device, it was not the case earlier. Earlier we were depending on HTML, CSS and PHP to design websites and JAVA to design mobile apps. With emerging technologies like react, react-native, angular these borders have become a blur now. A single app which is designed for iOS can be used across android with small tricks and can be rendered on the web with different CSS setups. It really powers up PIM in true way.

Step process to understand and implement PIM in MDA

MDA tool is used to develop, interpret, compare, align, measure, verify, transform, etc. models. We will have look at it how you can leverage them in your next project.

Creation: Make sure you break down every functionality in as small as pieces as possible. For example, add to cart which is a simple action can be broken down as – Sending product to the database, adding the product to cache in case the user decide to abandon purchase, start rendering related product below shopping cart based on which product is added to cart.

Analysis: The best approach to enhance MDA is keeping on analyzing it for its optimal use. World class known example to this is Amazon added “Add to” cart button on two locations on its landing page, one on top and one on decides bottom which resulted in the increase in their final sales. This knowledge comes from analysis.

Transformation: Next step to analysis is to transform your learnings in a viable design change. i.e. when you decide to launch your website in Middle East countries with an observation you can understand the attention plan on action items is mainly biased towards left so adding the call to action or CTA buttons towards left only for those countries can make a whole lot of difference.

Composition: a better ingredient for making result stand out after the action is composition. Once you implemented the change you reanalyze the change it has triggered and make a wise decision based upon the compositions turns out from it.

Test: let’s face it in BA it is 90% testing. And 10% enhancing, the more you test the better you understand your end customer and take wise calls. Give it some cool of time after you implement something. Let users cross the learning curve and give you real data which can be relied on. Test regouriously. 

Simulation: Simulate as a user in different situations to understand the optimal behaviour of the user. It will give you insights about how an user behaves on your web or app. Hotter is making this way easier. (Find more about hotter here. https://www.hotjar.com)

Reverse engineering: you can not get everything right all the time, we all make mistakes but same time we know other people who got it right. Learn from them sometimes doing reverse engineering on competitor products gives us a better understanding of our shortfalls.

Hope this was a helping knowledge, give me your point of views in comment. 🙂