World of data and analytics dilly-dallies between centralized and decentralized approaches. My belief differs slightly from those who believe in adopting what is new. Our infatuation with the most recent buzzwords in data and analytics reminds me of a scene from The Dark Knight Rises movie.

After “breaking” Bruce Wayne, Bane, the antagonist in the movie, puts him in a dungeon-like prison. The prison is unique in that it has no barriers; you can see the open sky and the sun, and you get an illusion of hope that breaking out from this prison will be easy. Bane states, “Hope is a powerful torture weapon because it beings despair when the hope is not realized. This prison gives you constant hope that you can break out. Because you can’t, it becomes the source of your consistent despair.” In my opinion, that “light-at-the-end-of-the-tunnel” in the movie was akin to seeing the next shiny thing or buzzword and trying to grab that, with the hope that the failure last time was because the tool or buzzword we opted was not optimal.

Returning to the crux, I believe we do not have to impose rigid frameworks and architectures in today’s era of easily customizable technology. What is best for some other organization may not always be best for you. For example, why can’t we have hub and spoke architectures instead of centralization vs decentralization? The challenge is that we do not have enough time to understand our problems in detail, which is a requirement to start thinking about customized solutions. So, we believe that something everyone is discussing will probably work.

Analytics is no different. We have tried many different ways and approaches. None of them are wrong or useless. But that does not mean that a single approach can be universal. It is not difficult today to combine existing capabilities in a process format that can help achieve existing goals. So, if you believe a centralized enterprise analytics platform is necessary, you may be correct. But if you think it will ensure a data-driven culture, it may not. Let us try to understand this via an example.

I once had a chance to work for an organization where we identified that parcel shipment costs were significantly higher than needed. To be “safe,” everyone was shipping products the next day, even though temperature-controlled shipper boxes guaranteed three days of controlled temperature. In some cases, temperature was not even an issue. Yet, those products were being shipped the next day.

The challenge was that these folks did not have something in front of them to think twice about their decision. And this is where I say that digital transformation is more about process and people than technology. If, while shipping a product, the implications of your actions are not “in your face,” then you will not take the pain to get into any “self-service” analytics tool to justify not doing something you want.

We obviously remedied this by implementing a process enabled by technology.

When I think about it now, scenarios like these are why centralized tools and technologies or analytics-focused tools may not deliver the value they can. But building a hub-and-spoke structure around it can. And the purpose is simple. Make evaluating the consequences of your decisions (through analytics) easy and accessible.

Analytics tools rapidly become data repositories as organizations try to pull in or link to all the data sources they need for analytics purposes. However, no matter how easy to use or “self-service” they are, they are not very tactical. The majority of them are still in “Business Intelligence” mode. However, you can leverage these analytics engines’ current data connections to perform tactical analytics. Let’s go back to the parcel shipping example. You can build an app (inexpensive to build these days) that leverages your analytics solution to highlight that you do not need to ship the next day and what percentage savings you can generate by switching to the next service level, just by product ID.

Why not do this in the tool itself? Think about this from the UX perspective. If an app can get you something in jiffy, you do not want to navigate through a full-fledged analytics application. This same approach can actually work wonders by transforming into a hub-and-spoke model where many such light apps linked to the analytics engine provide an extremely easy way for employees to evaluate their tactical decisions in numbers and become truly data-driven.

It is not about centralized, decentralized, hybrid, or any other jargon. It is just about what you need and what will make the task of every layer in your organization easy, productive, and efficient.