Of Analytics Governance – III

On from Analytics Culture

(First published in the WAO/FACTOR newlsetter in April 2013)

Application Governance – This is quite certainly where most companies spend most their mental efforts, and, well, most their money! My personal theory is that it is so much easier to buy a piece of software than a strategy, buttons on which to click more than concepts, while thinking the software will magically take care of everything. Yes, don’t deny it; that mindset exists. I often see it.

No doubt tools are very important in analytics. I don’t think they are the most important component of sound Analytics Governance though, far from it, and I will come back to that in the next issue, but I would be crazy not to acknowledge how essential they are. I also believe that digital analytics was from the beginning, and for a long time, driven by application vendors. Not surprisingly, we ended up measuring what those tools told us to measure, from the data they found when they first opened log files, instead of us imposing our analytical needs. To be fair, we didn’t even know then what we wanted to measure, so kudos to vendors for blazing the trail.

So, what do tools have to do with governance? Well, since investments are so important in this area, there are several aspects that require attention pre and post-purchase.

Pre-purchase

What?

In an ideal world, companies shouldn’t purchase an analytics solution without raising the issue of what it is exactly that they want to measure. How can you possibly choose a tool if you do not know what you want to do with it? In an ideal world again, I believe that this reflection should precede any decision which is to be made on choosing a solution. In principle you should already have established your different Key Performance Indicators (more on that in the next instalment of this series on Governance). These KPIs, coupled with the analytics solution’s capability of efficiently and effectively delivering the essential measurements identified during the reflection process, will form the basis of the selection criteria to be used for choosing your measurement tool.

I am well aware of the fact that companies purchase solutions before identifying what they need to measure, and in most cases this is how companies have entered analytics. Whatever the situation, if your company has not identified what needs to be measured, I strongly recommend that you do so, either to better exploit your current tool or review your initial choice!

Where?

Should tools be implemented within the company’s system, or should they be on the “cloud”, i.e. what used to be referred to as SaaS (Software as a Service)? This is particularly a question of maintenance, more than privacy or protection of data, at least these days I believe. The cloud is getting more and more normal to many companies, and IT managers are getting a less sheepish about it. Then, financial circumstances of the supplier become more important, since you will want them to stick around for some time!

Who?

Beyond implementation, which should always be performed by external experts of the chosen solution (of course, your internal IT “can do it”!), there is the question of staffing the tools, and especially the degree to which you will have to internalize skills and competency. Is it necessary that you have people who are expert in the solutions, or will an intermediate level suffice? I think training will be necessary, and make good room in your budget for it. Choose early where the tool people will reside. One of the worst places I have seen is in IT, in the application support department. Those guys have to maintain and support dozens, if not hundreds, of solutions. To them, if something is up and running then there is no problem.

Most analytics solutions will require deep levels of understanding and skills, even at the intermediate level. And yes, even the free ones need a good amount of training; don’t let the fact that they are free make you lazy, and neglect developing competency. You can do a lot of damage with incompetent use of an analytics solution!

Post-Purchase

Since the skill factor is so important, I encourage you to keep the vendor(s) around for some time, especially for technical support. Fortunately, most solutions come with support included in the initial price, and annual license renewal (or ongoing costs in the cloud version). Use it as much as you can as a way to be educated. Naturally, you will have decided that your internal experts will belong to the same team as those who use analytics (or in a service center outside IT, staffed with people who are highly sensitive to business problems).

To be honest, I think the best balance is to reach strong intermediate skill levels, and get consultants for the occasional configurations that require advanced skills. I am not sure it is worth investing a lot, so that your people will become experts to the point where even your vendors will want to hire them! Not that I think there is a risk it would happen; I just think that with expensively acquired tool-specific skills come more dependency on the said tools.

It is very important that you keep a close eye on the tools integrity. Errors can seep in without warning, and it is always easy to take displayed numbers at face value. Make sure you regularly test your current settings to detect potential errors, so that you can catch them on time. Never forget that it is easier to delay a report than to call it back!

Finally, time is of the essence here; long time that is. Consider that you should use a solution for several years, because regularly changing is costly, particularly in lost productivity. However, don’t see your technology choices as permanent marriage (although I am aware of this metaphor weakness); there will be a time when something much better, worth investing in, will come around.

Make sure that saying goodbye then will not be too dramatic.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.