What is analytics?
People really like to sort stuff. This trait is more pronounced in some people but to some extent most of us do it. It could be sorting people by perceived character or re-arranging all your mobile apps to be grouped by colour. Sorting often results in groups of similar things. If any group contains members that can be sorted further, we tend to split the group up and continue. This final result is logic and easy to understand.
The analytics consultant - A data ninja!
Sometimes boxing things in too small groups tend to break the requirements of reality. This often apply to Analytics consultants, since these often straddle the domain of both Business and IT. The Analytics person is the Babel fish of data driven projects; she explains to IT what the Business needs and translates IT-lingo to normal language for the Business to act on. A good Analytics consultant is hence not only IT and not only Business. Splitting the role into several roles will indeed be a logic and easy thing to do, but it will not be simple to achieve the intended result of providing business value.
So, what is an Analytics consultant? It's being a Data Ninja! There, sorted! Not enough detail? Well, keep reading!
The data model
An Analytics project always, always begin with an understanding the customers business. Even if the project was initiated by IT. Being a bit dogmatic, I often find that Analytics projects that is run by IT from start to finish, never actually begun.
The business processes are parsed, sliced, diced, simplified and reassembled to arrive at a business model that is easy for both IT and the Business to understand and relate to.
This simplified business model, the Data Model, is next used to talk to IT.
"Ok, people, we need some Data. Stat!"
This is often a roller-coaster ride; Wee, we got the data! And also Oh noo, this cant be right…. Or maybe a trip to the Haunted House, learning that the most simple requirements can not be met or the Data is of such horrid quality that you just say "nope, wont touch that".
So, where are we now? We got a good grasp of business requirements and we got some data thingies that we assume will meet them. So, done? Not quite.
Even if every requirement would have a solid data point to support it, the Data does not do much good if its not available to the users. At this point, a decision is made of what data to move and also how to move it. This is very IT-heavy stuff and the Analytics consultant need to go techy. Extract, Load, wrinse and repeat. Since Data mostly sit in several different data sources and rarely is easy relatable to eachother, the Analytics consultant must find a way to both merge Data logically and find the common denominator that binds it all together. This glue would be Master Data.
Without a solid understanding of Master Data, Data is doomed to be isolated islands with very angry natives on them. Even if the Analytics project have done great so far, failing in the Master Data step will without doubt cause the whole thing to do its best Titanic impression. Prioritizing what Master Data is needed is not always a given thing. Talking to IT will likely result in a too theoretical Information Model walk-through that results in a Confused, But On A Higher Level-feeling. Talking to Business representatives only is often hard since dependencies is complicated to see and thus prioritization becomes difficult.
Visualization - Let the fun begin
Lets take a step back. Do the Analytics consultant have any tool that might aid this work? Yes, indeed! Looking at the business requirements and the Data Model, the Analytics consultant get an immediate list of Master Data candidates. Business requirements understood. Master Data process in place. Data fetched. Time to go home? No, now the actual fun begins! Visualizations!
Any Data is only as useful as the user deems it. Or understands it. The human brain processes 11 million bits of information every second. 10 million is done via the eyes. In the end, 16 bits of information reach the brain. What is the implication of this? This: A shitty visualization of data will very, very, VERY effectively obfuscate any useful value of that data. Spending weeks and months on getting the Data to the users is of no use if this critical step is ignored or simply left to the users themselves to handle. We can't afford ugly piecharts, now can we?
An Analytics consultant is wicked on some but likely not all the above tasks. But she knows the importance of every piece of the puzzle and make sure that every step is taken care of. She is the Data Whisperer that transcend the border of IT and Business and make Everything Just Work, because she masters both. This is the reason why the sorting of roles into subatomic pieces is counterproductive to achieving business value. Notably, her Excalibur is the Data Model. The Data Model is a great communication tool to translate business requirements to IT and vice versa, it is the go to tool when pinpointing the Master Data and how to structure fetched information. Finally, it is the starting point for visualizating the Data. Foundation for all further work.
What about the other stuff?
Data Whispering and the Data Model is The Magic Sauce of a great Analytics consultant! The image sums it up well. What about AI, Machine Learning and all those things? Well spotted! Why didn't I bring in AI and ML in this picture? Because you need to learn to walk before you run. You could of course start running day one, but missing out on the core concepts I mentioned, will give you severe headache. Or create an isolated data island. With very angry natives. When I said that an Analytics consultant was a Data Ninja, I was actually spot on!
Do you need help staffing your projects? Or need an inspirational seminar to know where to begin? Or would you like training in Visualization or Analytics Architecture? At Enfo we will assist you with the entire range! If you have any questions or thoughts, contact me!
Magnus Hagdahl works as consultant at Enfo Analytics