Sharing ideas with the world

Wednesday, February 08, 2012

Musings On Analytics

7:20 PM Posted by Deepak Nayal , , , No comments
Saying that analytics is important for organizations is like saying that thinking is good for your mind. While it is a major field of study with volumes of books published on the topic, I believe that when it actually comes to analytics, companies focus much more on external factors than internal ones. I do not have not any data to prove this, but based on what I read, hear and observe on a daily basis, it seems that organizations are much more focused towards user and market analytics than operational or application. 

Not that there is any wrong about analyzing users and markets, it is just that based on my experience working with various organizations, I have noticed that there is generally a huge amount of information hidden inside systems and processes of companies. And tapping into these systems and processes using operational metrics can help organizations improve performance and efficiency. For example, if executives in a company know (and have the required data to support) that its buying process is taking unnecessary hops or that a particular checkpoint in vendor selection process takes unusually long time holding back all other downstream activities, then they can take required steps to improve the operational efficiency of that organization. 

So if internal metrics are that important, why are organizations more outward focused when it comes to analytics? I believe one of the primary reasons it is so is that working with external factors is easier and sexier. Don't agree? Think about it yourself. If you had to pick one, what would you rather do - study the user demographics of your product or study the various checkpoints of project approval process? 

Problems In Leveraging Analytics 
It is not that companies do not understand the importance of analytics or do not want to implement it. There are two major problems that hinder the adoption/leveraging of analytics, particularly for inner-facing factors. First is implementing the analytics system(s) in your organization, and second is embedding analytics in your organization. Let me explain these briefly.

Implementing Analytics In Systems
This is actually the easier of the two problems; though, mistakenly seen as the harder one by many - probably because it is the more tangible one. Before you start measuring things around, you need to setup the system that will capture the required data and then display it visually. With a lot of tools and technologies available in the market for doing so, it has become increasingly easier to implement. Companies with consumer/internet based software offerings have it easier here, given the range of software analytics tools available plus the fact that the information can be [relatively] easily tracked and stored in servers. Enterprise players (especially with desktop and server offerings) have it more difficult, as their customers won't be happy having vendors sniffing around their internal networks. 

While it is easier for smaller companies and startups to implement analytics, it is a much harder job for bigger enterprises. That is because large enterprises already have huge amounts of investments in systems and applications. To implement analytics, they will have to modify their systems and processes, which will require huge amounts of time, effort and resources. This is a huge deterrent and probably the biggest reason for enterprises to dance on the sidelines when it comes to implementing and adopting analytics. This, however, is ironic considering that implementing the system is not really the tougher nut to crack.

Embedding Analytics In Organization 
This is the more difficult and understated of the two problems. Embedding analytics in the organizations requires two things: identifying the right metrics and then getting people to use/refer them. Identifying the metrics is definitely a tricky problem because while little information will not help much, too many metrics can lead to information overload. And then looking at the wrong metrics can give you a totally wrong perception of the performance. Identifying and starting with the exact metrics is a very hard thing to do, but it might be advisable to err on the side of more. The reason for doing so is that by capturing more metrics than you thought you would need, you might notice trends or information that you did not consider earlier. Whereas starting with less metrics and then adding new ones later can lead to leaving out of many important ones because of the lack of interest or the effort required in making changes to systems. 

Once these metrics have been identified, they need to be ingrained in the organizational processes to make best use out of them. This is more of a culture issue - people might avoid using these metrics for various reasons. This is why I believe adopting analytics into an organization should be taken up as a change management program. The whole process of identifying the right tools, implementing the analytics systems, identifying the metrics and incorporating them in the process should be taken up as one multi-phase program, else you face the risk of losing momentum and ending up with part-baked solution. 

Different Kinds Of Metrics 
Identifying the right metrics is one of the trickiest parts of leveraging analytics. There is obviously no one-size-fits-all kind solution, as the metrics will vary from company to company and, even within a single company, from division to division. Metrics for a products company will be different than a services one, for a consumer company will be different from an enterprise one, and for a finance department will be different from those of a technology department. What makes it trickier is that these metrics might keep changing as the company grows and should be driven by the organizational level goals it is trying to achieve. 

While there are many ways of identifying and separating metrics, I believe that these generally fall into four major categories as shown in the image below. 

The market-based metrics are actually the easiest to get hold off. There are thousands of market research companies and software providers that provide the required data and tools to get the job done. It is the operational level metrics that are the most difficult to implement and least understood. There is a wealth of information hidden in systems and processes of companies, which can be tapped into by studying these operational metrics. However, as mentioned earlier, identifying and measuring them will most probably require changing existing systems and applications or implementing new systems to start collecting the required data - a big no for many companies (especially in the current economic climate). 

While organizations would prefer to work on all four categories of metrics, the focus should (and probably will) keep changing based on the stage of growth that the organization is in. For example, while for a startup or SME it might make sense to focus more on the market and user driven metrics, I think enterprises ought to have a better handle on the internal metrics. Knowing thyself is important. With better visibility of these internal metrics, organizations can improve their execution and customer service. 

Gut Trumps Data 
While the importance of data and analytics in organizations cannot be refuted, I believe that when it comes to decision making even all of the data in the world cannot trump gut. Data is important but it is only one factor, an input. If data was all that was required for decision making, robots would have been better off running the world economy. So do track and analyze data, but do not forget to listen to your gut as well.


Post a Comment