A Little Pain Never Hurt Anyone

Author: Vesa Muhonen, Senior Data Scientist @Mobiquity.

Breaking the Fall

In my free time I dabble a bit in the martial arts. Someone trying to hit you with a sword is a remarkably efficient way to focus your mind away from everything else. One of the basic skills you need is to do proper breakfalls. When you find yourself airborne, and ignoring gravity is not an option, it’s advisable to know what to do upon your rapid reintroduction with the solid floor. Nevertheless, the breakfalls won’t be the point here, but rather the place you train in.

A Tale of Two Rooms

Roughly speaking there are two different types of training halls. Most common by far is equipped with foam filled mats on a hard floor. The rare case has a hard floor as well, but one that has a bit peculiar construction. At first though this latter case would appear to be a far less comfortable place to train. Indeed, a practice session quickly confirms all such suspicions. Regardless, it is a superior place to train. The reason is, quite literally, hidden below the surface as the whole floor is suspended and because of this it’s able to absorb heavy impacts.

I have been fortunate enough to train in such places when visiting Japan. In fact, I can attest that an enthusiastic full-body contact with such a floor can make it feel as if the whole building trembles. Stories aside, is there actually a point hidden here somewhere? Don’t worry, there is.

The fundamental difference between the two setups is revealed when things go awry. Mats comfortably buffer small hits but a heavy impact reveals the unyielding floor underneath. The converse is true in the other case. The hard floor almost highlights the small bumps, but the suspension will absorb large impacts.This is what makes it such an excellent place to train. It unapologetically reveals the small mistakes and forces you to learn. But, when things go bad, the whole structure is there to reduce the damage and hopefully you’ll walk away with some bruises only.

The Business Connection

Great, now we know a good place to do breakfalls without breaking ourselves. What exactly has that to do with anything in business sense? A lot, actually. Companies need this skill as well. Quite obviously not in the physical sense, but as the ability to navigate the bumps and scrapes of their business environment. They need to put themselves in a place where they are able to walk away from major impacts, slightly bruised perhaps, but still alive.

Creating such an environment is not easy by any stretch of the imagination nor can I claim to have found a simple solution that will fit in the margins of this text. Fortunately, even without writing an exhaustive treatise on the topic, we can explore the most crucial components.

There Will Be Data

Without a doubt, the foundation you build on is data. Perhaps only aside from philosophy and mathematics you can’t have knowledge without data. Or at least I have not heard of a successful axiomatically led enterprise. Before utilising data, it of course needs to be acquired in the first place. While data engineering is a fascinating topic in and of itself, it isn’t our main topic today. For the following discussion, we assume that we have reliable data and it’s available for use.

Let us then concentrate on the usage. Unarguably a topic nigh unbounded in diversity. Nonetheless, a very sensible way for an organisation to get acquainted with using data as a decision making tool is with metrics, KPIs, and dashboards, or business intelligence (BI) for short. With all the buzz around the latest hotness in AI, it is all too easy to overlook this quite useful aspect of data work.

Laying the Foundation

Our humble BI serves as the foundation you can build on. It’s a stepping stone towards more advanced analytics, like those aforementioned cool AI projects. No company can complete a successful AI project unless they are familiar with the data they have, are comfortable in creating data products, and have the infrastructure to do so. Building good dashboards is a fine, and not to mention a cost effective, way to practice all that.

Merely having a foundation is not enough but it also needs to be well made. Circling back to our analogy, a well executed BI strategy is indispensable in creating that safe environment we talked about. The specific requirements and details will of course vary by organisation, but there are some principles that apply universally.

Three Core Aspects

We’ll consider three aspects here, accessibility, ownership, and quality. While some may be simpler in practice than others, none of them is a one-off exercise. But as long as this is recognised from the beginning, the maintenance effort will become just a part of the usual data work.

Accessibility

I believe accessibility is the least complicated of these. Unless there is a very specific reason to keep something hidden, the data products should be public within the organization. This of course requires a solution to be maintained that everyone has easy access to, but there are plenty of options to choose from.

However, accessibility here does not just refer to the technical aspect of it. People need access to relevant information. Dumping raw data in front of everyone and hoping that they’ll figure it out just doesn’t cut it. In order to avoid losing the proverbial forest for the trees, there must be a well thought out selection of the most important metrics that is the default for everyone.

The thing is, that not everyone’s default is the same nor are any the defaults static over time. Certainly there are a few company-wide KPIs that everyone should know, but a product owner needs different things than someone in customer support does. Maintaining a sensible set of core figures for each appropriate part of the organisation should not be a difficult thing to manage, but something that needs to be done. A far greater challenge presents itself when we talk about the details of the KPIs themselves and that leads our way to ownership.

Ownership

Ownership is absolutely crucial and too easily overlooked. Data and analytics do not possess an intrinsic value, but rather the business value manifests through using data as a decision making tool. Every KPI and the related definitions must be owned by someone who is capable and mandated to take action regarding the matters the KPI relates to. A KPI without an owner that exists “just because” should be discarded as a useless item only generating technical debt.

It should be noted that one should not mix ownership with exclusivity. Everyone should be free to use any of the KPIs they find useful. Ownership is important because of clarity and evolution. Especially for KPIs that are widely used it is crucial that they are unambiguously defined. In many cases the definition can be quite trivial, but certainly not always.

A good example is something like “an active user”. I think we can all largely agree on what a user is, but what makes them an active one is entirely dependent on your business. Having several inconsistent definitions of such an important concept floating around your organisation would make talking about them fraught with misunderstanding. Designating an owner makes sure that there is one source of truth regarding this concept.

However, consistency and clarity must not mean stagnation. Consequently, another important aspect of ownership is the responsibility to keep the KPIs up-to-date. Business is not a static entity and neither should be the ways you keep track of it. Because the owner has the responsibility to act on their KPIs, they must be the one to adjust them as the business evolves.

Quality

Finally we come to the matter of quality. None of the considerations above have any real meaning if the numbers have none themselves. Unfortunately, there are some easy attractions to lead people astray.

One such stumbling block is quantity over quality. This often manifests early in a project or early in the journey towards becoming more data driven in general. When you’re not sure what’s important, there is temptation to just report everything for the fear of missing out. This is both not seeing the forest for the trees and everything being important meaning that nothing is. Managing to exemplify two well known aphorisms in one go is not exactly a winning strategy here. One must choose what is important and when. Fortunately, the inevitable discussions exploring this question mandate a deep look and understanding into your business and that has value in and of itself.

There is one crucial point that needs to be highlighted here. Not counting everything absolutely does not mean that you should not be able to. Your analytics platform should allow you to calculate any metric you choose. If that is not the case, you should review your platform strategy to understand the limitations. As different things become important when the business evolves, the analytics platform must be the enabling factor in understanding the prevailing situation.

Another of those analytics sirens to draw you off-course are vanity metrics. These are almost the quintessential data analogue of the soft mats over a concrete floor we started this discussion with. They carry the risk of hiding away the hard reality or just make you feel good about yourself.

One great example in the realm of mobile applications is the number of downloads. It’s guaranteed to never decrease and thus displays in charts as a line from lower left to upper right, which is what everyone loves to see. We can attempt a simple thought experiment to explore this metric. Say you really think that this is “The Problem” to solve. It actually should not be that difficult. Stop all activity except marketing and pay 5€ to anyone to download the app. I would expect a very rapid improvement in the metric in question with quite a high degree of certainty. I would not expect a great business to be built in this manner though.

This is perhaps the most ephemeral of the core aspects presented here. It’s not that hard to think about quality in general terms, but the details are very specific to each use case. A high quality KPI is actually relevant for the owner in the current situation and it doesn’t hide any uncomfortable truths. An effort has to be made to maintain them, but it’s a task that requires you question what is really important for the business and that is not a wasted effort.

The Practice Hall That Data Built

So, how do we create an environment where our enterprise can survive the slings and arrows of outrageous fortune? Fundamentally it’s built from data on a solid foundation utilising that data to truly understand the business. It’s not locked away for a chosen few to enjoy but it fully permeates the organisation. It’s a continuous process because it reflects your business which is not stagnant either. It’s valuable because it is an important decision making tool.

Used effectively, data can make all the friction points painfully visible, forcing us to learn and adjust before too much friction stops us or sets us on fire. All this makes us much more nimble and able to react to surprising events or misfortunes. A bit like a training hall designed against heavy impacts…

We hope you enjoyed reading this, and don’t forget to follow us if you’d like to read more interesting stories.

Originally published at https://www.mobiquity.com.