Learnings from more than a decade of experience in this realm by Michiel van Staden, Data Analytics Leader/Speaker/Listener & currently Data Analytics Lead @ Absa banking group in South Africa
Michiel van Staden, Data Analytics Lead, Absa
Michiel will be speaking at the virtual IRM UK Business Intelligence and Analytics & Enterprise Data Conference Europe 2-5 November 2020. The conferences are co-located with the IRM UK Master Data Management Summit & Data Governance Conference.
He will be speaking on the subject, ‘Getting Data Analytics to Have an Impact on Strategy in a Large Organization‘
Why do we need data?
Businesses have business problems.
A vision is devised for what the company stands for and a mission statement is carved out to accompany it, outlining the general course in working towards that goal.
Initially, and then at regular intervals thereafter, as seems fit (say every 3-5 years), a strategy is crafted or revised for achieving the business objectives. All of this boils down to annual target setting, of what various levels, down to those on the ground, need to achieve this year to make results happen.
What informs these strategies and the targets set? Data on historical learnings, current or forecasted trends and the industry environment could definitely play a role. How does the business know whether they have achieved the targets or whether they are on track? Data.
Even when all of the above is done and communicated really well, giving the entire workforce clear direction, it still does not tell them what exactly to do each day when they get into their place of work. Personally I find this liberating and empowering, because it implies a sense of autonomy, but with this freedom comes responsibility. Employees need to figure out an action plan for delivering against the demands set out in front of them.
Data can provide significant support in setting smart goals down to each individual, measuring and forecasting progress against these objectives, and ultimately even in informing the best course of action towards getting there and beyond.
Where will our data come from?
As I’ve alluded to above, there are various touch points where data could either be required, or could play a valuable role. In situations like these, the relevant decision maker would ideally just want to summon data, for all of the answers. But unfortunately it does not work like that. Not yet, in any case.
In reality, a range of functions, each with very specialized skills, is needed in the background for data to deliver on its promise:
Suppose it can be done, but realistic target and even strategy setting seems impossible without being informed of the realities. We also do not know how we are doing against these objectives, unless we’re consistently measuring performance.
This is where standard reporting comes in. With good programming skills and a firm understanding of the relevant data at their disposal, an MI (management information) data analyst should be able to compile the information needed.
Expert knowledge of visualization tools allows them to display this information in an intuitive way and make it easily accessible, whilst the ability to put these reports into automated production frees up analysts to ultimately move into BI (business intelligence), where they get to know the business and can add relevant insights to reports.
Once business knows what is happening today, and what came before that, questions naturally arise around what can be expected tomorrow. Modelling specialists have got very scientific training in giving answers to very specific questions, based on historical information.
With machine learning tools and know how, these predictions can be refined automatically as new information comes in and also extended into the realm of artificial intelligence.
What if business wants to try something different from what they’ve been doing? Historical data can often not tell much about something that has not been done.
Roles such as campaign or marketing data analytics holds expertise on how to design specific tests for optimal learning, whilst also developing experience in implementing these tests practically.
Given these functions work well, business knows what is happening and has the support needed to predict or test. Most often the hardest question however, I find, is what specifically to test or predict, as these need to be focused.
For this exact reason, the business data analyst was born. Somebody with sufficient understanding of all the above, together with a broad and deep understanding of the business.
By truly grasping the business challenges, asking the right questions and spending lots of time listening, this incarnation of data analytics can collaborate with the other data specializations towards finding fit for purpose solutions.
Over time the ability can then be developed and customized to present these in a compelling way, get further business input and reiterate towards successful data driven actions.
Businesses don’t take no for an answer
In reality, the few business data analysts who have gained sufficient business knowledge and trust, cannot get round to all of the challenges. Testing and modelling functions, if they exist, can easily get pulled into business analysis, keeping them from practicing and advancing their craft. MI can end up spending all of their time running the same reports and fielding unrefined business analysis questions.
Alas, business could keep running after the next available data person, asking why they can’t just have one data team to go to for all of the answers.
Centralization of the data function is an option, but by being removed from business units, sufficient business knowledge and trust is even harder to come by.
I believe the answers are straightforward :
Literacy – More often, there is no shortage of people with vast amounts of business experience built up over long careers. Rather than solely expecting data specialist to match this, how can we give all business people sufficient data literacy, to add more data driven business analysts?
Automation – is what allows all technical data functions to move forward. If you’re performing the same menial task you did yesterday, you’re standing still and most possibly drowning. Data work needs to be treated more like software development. Analyze to prioritize, design to put as much as possible into robust production, assess and then move onto the next.
Getting there can however be extremely hard. Many business people still feel overwhelmed by data and very few want to wait for data answers to go into production, but I believe it is the way forward for businesses to be able to compete. Some might already be there.
After finishing his Honors in Mathematical Statistics, Michiel joined the newly formed Customer Value Management team at the Absa Banking Group. Over the following decade, he learnt the practicalities of data analytics across fraud prevention, collections, legal recoveries, credit risk & campaigning, whilst he have also explored the root cause of complex business problems as a strategic data analyst. Venturing into the realm of management, he now leads campaign analytics efforts, with a passion for communicating data effectively, sharing his learnings and a keen ear for listening as an aspiring performance development coach.
I believe in Freedom with Respect – Pushing Boundaries, Asking Questions & Listening – Community in Diversity. Follow me on LinkedIn: https://www.linkedin.com/in/michiel-van-staden-1130a7110
Copyright Michiel van Staden, Data Analytics Lead, Absa