It’s been lost to the vagaries of crowd-sourced curation, but for a while my favourite definition of the word “capability” came from Wikipedia, where it was succinctly described as “the sum of expertise and capacity.”

Your data analytics technology stack is an enabler of your analytical capability – it alone is not your analytics capability. I’m not just saying this for philosophical reasons.

The fact is Big Data technology and software solutions are evolving so quickly that analytics leaders (regardless of  industry) simply cannot depend on any past or current standard as a means to define their capability.   

It brings to mind one of the most commonly asked questions I get these days: What technology are you using, or what analytics toolset do you recommend for doing the hardcore stuff? 

Obviously, it’s far too circumstantial a question to answer intelligently without a much deeper conversation. But the first thing I usually ask when the question comes up is, “Tell me about your analytics team.”  

More often than not, what I hear in response leads me to suggest deeper and more strategic consideration for analytical talent before sweating a technology decision much further. Call me a purist, but I would always argue for the budget to recruit the best talent before ever worrying about which technology or analytics tool to invest in. 

In fact, I’m still amazed how often I hear these fundamentally important questions coming from people who are quite frankly unqualified to be making the decision.

I know one of the reasons for it is how adamant vendors in this space can be about how easy their solutions are to integrate, how anyone can use them. 

Don’t get caught up in it. If you’ve hired the right talent, I assure you, they would already be telling you what the stack will need to look like, probably demanding it.

Another way to look at it: If you’re driving a capability (or, worse, a business need) based on the promise of a tool or technology, then you most likely are not capable of leveraging it in the first place.

I spent most of my career in data-mining analytics in the telecommunications and high-tech industries. I was a pretty good analyst, I think. Things were also a lot more straightforward back then. 

A good size relational database, a fairly standard data model, a data dictionary, if you were lucky, and typically statistical analytics software (SAS) as the tool of choice for deep data mining and analysis. 

We’ve all talked before about the eventual path to Big Data; the journey has been well-documented. But the essence of what it is we were trying to do along the way has remained true. There is value in the data; analytics can exploit it; and if you do it well, hopefully it’s good for business and good for the consumer.   

In the meantime, though, this was also good for the analysts. As if they weren’t already the most sought-after positions in the market, today’s data scientists are becoming a far more integral part of the business.

They not only analyse behaviour, they’re also at the forefront of product and experience development.  

If I just look at my own team as an example, they are simply better at doing what they do than I was when I was doing the same thing for a living – period. 

They code in more languages; they understand and leverage data processing, in database and in memory, like I never did; and they are far closer to the resulting customer experience than we old database marketers were. 

One of our senior data scientists walked me through a serious piece of work he recently completed, and I was not only amazed at how creative his thinking and design was, but how intelligently made his choices were along the way. By choices, I mean how he optimised between programming language, data platform, and methodology across each step.

Although I’ve spent a career in this area and I’m still a reasonably intelligent guy, I find I am now clearly out of my league when it comes to the real hands-on analytical work. So exactly how much sole authority should I have when it comes to the right or best choice of technology and software on behalf of these real experts?

Some food for thought if this sort of issue resonates with you:

Hire the expertise first – and don’t over-hire: If you’re a senior analytics leader struggling with decisions on the right technology or software, then you need to hire or replace your talent, because they should be telling you. 

I’ve had the privilege of leading what I consider to be some of the most advanced, high-performance teams in any industry. I never once had to ask their opinion on these types of choices, and I certainly never thought to prescribe analytical solutions on their behalf. 

In my opinion, most advanced data/analytics programmes fail because people who do not have the “expertise” or “capacity” are driving them.

Already planning to hire/build the team? Of course, it’s always ideal to hire the most senior role first and allow him or her the opportunity to build the team. But if headcount is short and/or your data programme is under pressure to deliver tangible results quickly, then I always recommend hiring the absolute best analyst you can afford first. 

I don’t care if you have the most senior title in your organisation, if you’re driving for a results-oriented data strategy, then bring in the person/people who will want to get dirty quickly. I mean hands-on-keyboards type talent.

I’ve never worked in a business that had a shortage of bright ideas people – you’re probably covered there. It’s the “doers” that turn road maps and visions into tangible results.

Data in the cloud: OK, I hear you. You are going to hire the talent – you get that part. But in the meantime you need to advance some of the more technical pieces of the plan now, maybe in preparation for their arrival.  

My advice: Look to the cloud.

It took a while for my team to eventually persuade me that we didn’t always need to build and maintain our data analytics environment in-house. It just seemed so strange to me that our data wouldn’t sit on a big computer I could walk up to and touch.

But I’ll cut to the chase: They were right and they knew best. 

What used to take a considerable investment in time and money can now be done overnight, literally. I won’t get into all the options here, but the larger players offer many profile options and flexible pricing and multiple packages for storage and processing, etc. 

Simply put, what I have seen accomplished in mere weeks leveraging the flexibility of cloud storage and computing simply cannot be accomplished through more traditional in-house solutions. 

Our first piece of serious Big Data analysis was executing three weeks after making the decision to give the cloud solution a shot, and our total bill for that first month was less than C$2,000. 

Again, there are far too many circumstances to consider to offer any specific practical advice here. But for this portion of your data-analytics stack, there is very little risk and nothing but upside to go cloud-based … until your real experts can come in and judge for themselves, that is.

Remember, if you hire well, these things are always their call to make.

Consider and document your objectives and intentions: If you’re actively recruiting talent, then you probably know by now, they will have questions. The best data miners I’ve hired typically ask two types of questions. 

They certainly will ask about the infrastructure and resources they have at their disposal – another good reason for taking a cloud-based approach that offers multiple platform options, so you’re not locked into an infrastructure a great candidate doesn’t like.

The other sorts of questions they will ask usually involve the business objectives the analytics team is meant to help achieve: How will what they do add value and is the business ready for it? 

Not only a fair question, but a fundamental one for everyone to truly understand and answer. Simply put, you need to know what exactly you want to achieve before you can hire the best people for the job, let alone think about building a technical environment capable of supporting them as they do it. 

Perhaps the best way to summarise here is that your organisation’s strategy, its intention, for data analytics technology and tools should be driven as an enabler that supports the actual analytical capability you have – your analytical talent – and not the other way around. 

If, for whatever reason, you cannot hire what you consider to be the best, most qualified, people for the job, then re-assess the purpose of the investment and be honest and transparent about it. It may be gut-check time.

There is little sense in building or over-engineering your race car if you don’t have champion-calibre drivers. On the other hand, the good news is championship-calibre drivers are rarely shy about telling people around them exactly what they need to win.  

At the end of the day, your job, like mine, might simply be to set the vision, establish the objectives, help when asked, and get out of their way.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .