Shikarishambu on Software & Related Services

I have long argued that Apple’s “I am a Mac” ads cause confusion in the minds of viewers. Almost all viewers think Apple is taking a dig at Microsoft. Yes, they are. Not, they are not taking a dig at just Microsoft. User experience is not derived from just the operating system but also from the hardware and the software that the user tends to use the most. The Mac advertisements are about user experience.

An integrated platform like Mac where the hardware, the operating system and most of the software comes from a single company (in this case, Apple) allows Apple to offer a “better” experience. (Note: I have not used a Mac in the last decade so I really cannot say whether it offers a better experience). The PC, the Windows Operating System and the applications that run on it are discrete pieces that are created by different parties, assembled by different and experienced by a single user at a given time. All these companies contribute to the user experience – the good, the bad and the ugly.

I was quite surprised to hear of the “Iam a PC” campaign from Microsoft for many reasons. For one, it looks and feels like a knee jerk reaction from Redmond. Then, I don’t think Microsoft has ever had a memorable product name leave along an ad. campaign. So, why waste money. Last but not the least, when did PC mean Microsoft. Yes, Windows run on about 80-90% of the PCs but there are PCs running Linux, Unix and <<you name it>> operating systems. Does Microsoft make PCs? Last time I checked… No. So, shouldn’t Dell, HP, Lenovo, Acer, Asus along with Microsoft do the “Iam a PC” campaign. I guess they are not as stupid as Redmond. A fool and his money are soon parted. I guess that is true for corporations, too.

A lot of words that get thrown around during discussions between business and technology folks are “overloaded” – used to mean different things for different parties. Often parties involved do not recognize this and think that every one is on the same page until the outcome proves otherwise.

Every Technical architect knows the importance of stuff like – availability, reliability, scalability, flexibility and performance. Together these are often called – *ilities and mastery of these is a sign of a good architect. Now, as you move up from the role of a pure Technical Architect to an Enterprise Architect and beyond you have to take an organizational view rather than a systems centric view. Often these require conversations with business heads on their business processes and pains. Sometimes these conversations will include some or all of the *ilities but they mean different things to the business.

Let us take a look at what they mean to both parties

Availability (Technologist) – System uptime.

Availability (Business) – It includes system uptime. It also includes “accessibility”. Is the system available when I am on the road? Is the data available on my phone?

Reliability (Technologist) – What is the accepted “Mean Time Between Failures” (MTBF)? Reliability combined with Availability gives you the acceptable downtime per day/week/year. To a lesser extent this also answers the question – Is there enough system safeguards to maintain data integrity?

Reliability (Business) – Rarely do the business users link “reliability” with “availability” the way the technologists do. For them reliability is all about data reliability and data freshness. Can I rely on this data? When was this data last updated?

Scalability (Technologist) – Can the system architecture handle the growth in data/ transactions. What is the scale out strategy – vertical or horizontal?

Scalability (Business) – Can the system handle growth in business locations (new offices, new markets, new countries etc…)? When do I need to add people to handle transactions in the system?

Flexibility (Technologist) – Are parts of the system behavior configurable? Can changes be implemented easily without having to rewrite/ recompile large parts of the system?

Flexibility (Business) – Can the system handle changes to “business processes” on the fly?

Performance (Technologist) – What is expected and acceptable response time for the system for a specific action under a specific load condition?

Performance (Business) – I need the system to respond immediately to my action.

Most of the challenges I have encountered  with business users is around – scalability, flexibility and performance. I think the “scalability challenge” can be easily overcome if we understand what the business user is asking and we drop the system centric view to take an organizational view. The “performance challenge” is also relatively easy – it is about getting the users to qualify the actions and quantify the acceptable time to complete the action.

I think the greatest challenge is around “flexibility” - often (not always) what business wants is a system with rules where rules can be changed on the fly. This would break even a system that has a user-friendly rules engine. And, just because they ask for it does not make it right. It is important that the business understand the concept of standard process, alternates and exceptions. It is also important for the business to have release cycles around business processes just like IT has release cycles for systems. Changes to business processes involves communication, collaboration and confirmation.

About two weeks back I was at Convergence 2008 – the annual conference for users of Microsoft Business Solutions products – aka Microsoft Dynamics. During the conference I had a chance to have conversations with number of other users. Most of them were taken aback when I said our users find Microsoft Dynamics AX to be cumbersome and unfriendly. The response often was – “Have they tried using SAP or Oracle?.” I told them that our users compare it with Microsoft Excel. They often rolled their eyes.  But, let us face it – people love Microsoft Excel but they hate <<Fill in your product or application>>. Why is that?

I call Excel the “two-click” system – a click to start and a click to close. It is the ultimate in “flexibility” – you can make the data in there represent your version of the truth. Everyone can… it is very democratic. Now, I am not saying all Enterprise Applications should try to emulate Excel. I don’t think they can. The purpose of an Enterprise application is to help business process standardization and business process improvements. However, often times they fail not becuase they do not do that but because the users do not know what the business process is. So, the system does not help them get their job done. They see the system as a hinderance.

Very often businesses rolling out Enterprise Applications use the application to communicate the business process to the end user – “You do not have the approval authority because you cannot do that in the system.” The problem with this approach is the users cannot have a conversation with the system – get their queries answered or understand the logic behind the setup. In the best of the cases people manage to figure out what is their scope of work. However, most often this approach only leads to frustration – people attribute their lack of understanding of the business processes to the feature shortcomings of the system. 

 Here is a simple test – Take away the system, Do people know what are the steps in the process, what are the data elements input/ output at each point, what are the artifacts created/ used and who else is involved at each step? If yes, then you have a business process that can be supported in a system. If the answer is no, no system can help you. You see this often in real life – a system crashes or is not available and no one knows how to handle the business process.

I have had difficulty finding words to describe the messages provided by business dashboards and KPIs. That is until I listened to Kevin Schofield’s (of Microsoft Research) keynote at Convergence 2008. He quoted from T.S. Elliot’s “The Rock” to launch into the great stuff that MSR is doing around data

Where is the wisdom we have lost in knowledge?

          Where is the knowledge we have lost in information?

So, now I have the hierarchy – Data, Information, Knowledge and Wisdom (DIKW). Turns out I am not first to come up with this. Nikhil Sharma of UMich has an article on the origin of this (see here). So, how does this DIKW manifest in Business Intelligence? The data resides in data stores (databases, files, emails etc…). The information is derived from the data and is usually delivered in reports. The knowledge is derived from reports and delivered through dashboards and KPIs (assuming you did PMA and found the right parameters). Wisdom is the actions or the business strategies you create based on the knowledge you have.

I just returned from Convergence 2008 in Orlando i.e. the Microsoft Business Solutions conference. It is a great conference and for once Microsoft seems to have got the name right. A lot of the Convergence 2008 sessions were focused on Business Intelligence (BI). Now, don’t be cynical and call “Business Intelligence” an oxymoron. Microsoft seems to have a lot of work and products in the works in this area. Anyway, I am not here to talk about the products but what a speaker said.

During the BI roadmap session the speaker remarked that good BI requires PMA – Plan, Monitor and Analyze. You first plan on your goals, identify the parameters that will help you and define the metrics. Then, you monitor the parameters. Once you have collected you analyze it.

However, quite often businesses go about this using the MAP process.  They start monitoring things and then analyzing it without a real plan. The result is they either end up monitoring wrong parameters which cannot provide answers to the questions they are looking for. Or, they are monitoring a whole lot of parameters which costs a lot of money and then result in them getting the wrong answers because in large datasets it is possible to infer patterns.

While AMP is a great acronym Business Intelligence can never be AMP – you cannot analyze without having the data. So, M has to come before A.