Making Data & Analytics work

If you can relate to any of these 5 questions below, then this article may have some content relevant to you.

1)?????Is your organisation struggling to get value from your data?

2)?????Is there always a big backlog of BI and data requests?

3)?????Are the insights provided often too late and the decisions have already been made?

4)?????Is the ROI too low ?

5)?????Are the insights provided irrelevant or not insightful ?

If one or more of these questions resonate, you are not alone. “Actionable Intelligence” or “Information at your finger-tips” or “Data Driven Insights” … the list goes on – have been promised for many years, yet very few organisations are really able to leverage their data to anywhere near its full potential.

Outside limited sparks of brilliance in some parts of an organisation, where there is a “hero” analyst or maybe significant investment has been made, it is hard to find exemplars of organisations extracting meaningful value from their data investments.

I recently attended a data & analytics conference in Auckland where the topic of an interactive workshop was “Democratisation of Analytics”. Almost everyone in the session had realised this was needed to gain better value from data and were adjusting their operating models to enable this.

However, what was clear in those conversations that whilst good progress is being made, not one organisation had a mature data eco-system. ?Most were closer to the start of this journey and all were trying to figure out how to enable this?nirvana.

Why is it not working?

Well firstly, it is actually next to impossible to meet all business demand for data & insights, especially at the center ... our Data & Analytics team. However, this doesn’t mean we don’t have a problem … we do … its just that our mainstream approaches are leaving a lot of value on the table. Why is? Many reasons in my opinion: -

1) The Unknown - Firstly, because it is impossible to cater for every single permutation, combination and aggregation of data needed by the business. Just impossible. So we need to cater for the unknown in our approach to delivering intelligence.

2) Insufficient Capacity - Most business demand is quite variable and often needs immediate response. No matter how large your Data & Analytics functions are, you will never have enough people to meet total demand. Additionally, problem statements keep changing all the time. A business is a living breathing entity composed of its many parts and responding to the eco-system it is operating within.

3) Missing Business Context – it is almost impossible to apply business context unless you are immersed in the business trying to solve their problems. You need to bring both the problem statement (which is often unclear), the data, and iteratively build the intelligence. You need to test and learn with the data and the problem you are solving. Usually, it is only after many iterations do you create the desired "meaningful and actionable insights". Even these, have a limited lifetime unfortunately - as the focus changes.

4) High Cost & Low Speed – Most insights need to be delivered straight away, at zero cost and need to be relevant. How do we achieve that? Especially when it typically takes weeks to deliver the engineering needed to deliver the insights. In fact, it can take weeks to just qualify and define requirements.

So, what to do?

I am going to present my view of all the things I am thinking about when trying to deliver Nirvana or at least get closer to it. In my view, not one technology, or person or even methodology is going to address the problem. We need to turn the dial on many dimensions if we want to lift analytics capabilities and the quality of outcomes delivered.

The list below is a bit of a brain dump, but I have found all these areas need to be addressed in some way or another when responding to data gaps and trying to modernise. So, in no order of importance: -

1) Improve Data Discovery – Teams need the ability to find the data they need regardless of where it sits. They also need to know who to contact to understand what that data means including its quality.

2)?Enable easy access & data exploration – Data needs to be easily sourced by business teams for exploration, and once available navigable. Also data means .. all data .. not just the data sitting in our data platforms nicely cleaned and prepared for consumption. We also need to be able to explore our data without having to align to predefined hypotheses i.e. table structures that imply the problem we are trying to solve.

Another key aspect of data exploration must be recognised, only the business can actually identify real value. In order to enable this, we must not limited ourselves to the bounded context of the presentation layers of your data platform. Instead, you need to enable business users to explore both the curated data as well as raw data from source systems. You must also simplify source system data structures so they are easily navigable and can be quickly used to build new insights.

3)?Enable Data Innovation (Everywhere) – Data innovation happens everywhere. We need to reduce the effort and cost required to innovate with data and enable rapid cheap automation of valuable insights where the data team is not at the centre of delivery.

4) Flexible Data Security – Data needs to be secured, with a classification, with a view to enabling greater faster access ... not less – but safely.

5) Appropriate Software, Tools & Technology – We need to align our tools to our users and select the right tool for the job. There is no right answer, and there is no wrong answer but the right tools need to be available. This applies to all layers of our data platforms, from end users through to the data engineers building new capabilities.

6) Digital Collaboration & Engagement – is a must. It is important to create those data & analytics communities within our organisations who can learn and support each other in their journeys, share insights and teach each other.

7)?Embedded Change Management – at various levels, from executives through to analysts, change is one of the hardest and yet one of the easiest things to deliver – depending on the approach. Great insights, often deliver change management for you.

8)?Capability uplift – This is one of the most challenging problems, a target skill profile for each role and then an optimised approach to delivering this is needed. How to deliver this, is a challenge no one has cracked. However, using tools end users are comfortable with is a good start.

9) A stable product foundation – Solid data foundations will enable reuse and allow you to accelerate insights delivery. Value is delivered fastest at the lowest levels .. e.g. just give me my data.

10)?Automation to reduce friction – the more you can automate the better. Automation reduces the friction in producing new data or change the cadence of existing data. It also goes hand in hand with your tools and your information architecture.

11) Data Literacy - One of the biggest issues is data literacy and using commonly available tools in one way of reducing this barrier. Also important not to confuse ability to use tools with true understanding of the data. The tools bit can easily be lifted through the use of common tools such as Microsoft Excel. The data understanding, including all the subtleties, requires a deep understand of the business and current events.

12) Information Architecture – One of the final pieces of this puzzle is the creation of an information architecture that supports all of the above to take place. Unless our data, data flows, engineering practices are setup in a manner to enable data innovation and rapid productionisation of insights - we will struggle to respond at the required velocity.

13) Customer Architecture – Understanding / profiling your customers to understand how best you can enable each “profile” or “persona” of end or citizen user. In fact, discovering and utilising customer personas you can share the workload with is almost more important than the rest of the your users combined.

14) Governance – To prevent chaos, enabling prioritisation and finally decommissioning no longer required data and insights, governance is critical and requires and engaged audience.

15) Target Operating Model – is required to deliver sustainable value and long term maturity. The key is to enable business teams through better support, access and processes & workflows to deliver data faster. Automation is key to reducing the friction in delivery.

16)?Inquisitive Workforce / Data Culture – Last but not least, having an enabled inquisitive workforce will help (and likely drive) acceleration of insights generation, data innovation and finding the gold nuggets in your data. Building that community of practice within the business, who are motivated to use their data is your key to ultimate success.

Closing Thoughts

The foremost thoughts when reassessing your approach and looking at the above, keep asking yourself how do I deliver value, and who is the customer I am delivering value to. It is all too easy to get lost in the latest and greatest solutions, frameworks and methodologies.

A lot of what we are able to do now with modern capabilities was still possible with older tools – but was just a bit harder to achieve. The key is how we use all the tools and people available to us to achieve nirvana.