Do you check your balance before making a financial purchase? 

I’m not necessarily talking about a coffee or lunch but let’s say before spending on something small enough that it doesn’t feel like a major purchase but big enough that it will have an impact on your cashflow. Do you shop around before renewing your insurance? Do you look around for cameras and police putting your foot down to get the next meeting on time? If so, you are making data driven decisions already in your life.

We all do this, every time we shop, drive, walk down the street; and increasingly we are making these decisions with the help of technology.

When was the last time you took a journey and planned the route by taking out a paper map to look at the route you should be driving on the roads? When was the last time you relied on receipts and your last statement to tell you your current financial position, the last time you renewed your car insurance did you call around, or use an online insurance aggregator to find yourself the best deal, or just let the auto-renew roll forwards?

Satnav, banking apps and aggregation sites are all examples of where we have come to embrace having data at our fingertips. The integration of data into our lives has been so universal, ubiquitous and easy that we no longer see it for what it is.

So now the big question, why do we struggle to do this in our work lives?

There are big differences in the use of data in our personal lives and our work lives. First there is the volume and complexity. Using data in our daily lives has been made easy, someone has done all the hard work for us and all we need to do is embrace it. In our work lives, often we must be the driver for change. The work needs to be done by us, or if not then we need to push for it to be done. The data does not always exist in nicely curated collections, and we may need to find ways to capture or collate the data required. There are often multiple systems that need to be brought together to deliver the insight required with no clear way of joining the data.

There are many competing technologies and methodologies, picking a direction can be daunting, especially when the cost of a wrong decision is a high level of technical debt, increasing costs and stretching timelines. A failed project can be reputationally damaging. So, with all of this in mind why is it so important that organisations embrace becoming data driven decision makers?

Because the cost of not leveraging your data so is higher. Without data to drive decisions we are left looking at anecdotal evidence and gut feelings as to which way we go. How many organisations have failed due to overextending? How many have failed because fraud or mistakes were covered up?

Robust data analysis brings problems into the light and identifies models of working that should be praised and expanded across organisations. It drives informed discussion and guides us in what is possible and what is not.

This is important in large organisations, in smaller ones it is critical. In a small organisation the effect of bad decisions is magnified. Tighter margins, less capacity and less access to liquid funds to shore up against bad choices mean that a smaller organisation can fail on a single bad choice. Given that the smaller organisation has less in the way of capacity to fund development of an analytics capacity to support this data driven approach how can this be delivered on a budget, making the best use of the resources available?

Most analytics projects that fail do so because they are not thought through fully at the start and the goal, and roadmap to reach it, is unclear. Many don’t even reach that stage because they are thought through so much, they never actually start. The key to delivering an analytics project is asking the right questions and making the choices that need to be made now, leaving the ones that we don’t need to make now until they are needed to be made, mindful of the end goal.

Thkey choices that need to be explored prior to starting down this path are: 

  • Data Governance
  • Data Modelling Methodology
  • Reference Architecture
  • Data Assets
  • Analytics Opportunities

Understanding these from the outset will inform what you can and should do and how it will benefit you. It will also give you the ability to develop things in a modular fashion. Keeping a continuity of design while only standing environment and data up when there is benefit, avoiding unnecessary cost.

Why are these choices so important?

Each of these choices will guide how you build out your platform and will prevent you from creating technical debt that will spiral future costs.

Data Governance

Focusing on data governance prevents issues with GDPR, master data management, and increasingly Ethics. Making sure that Data Processing Impact Assessments (DPIA) are in place will stop work from halting while these are created. Data governance encompasses a lot of the non-functional requirements that are easy to put in place before you start but are a lot more difficult to add in once the build of the system is underway. Preparation is essential to prevent you starting and then having to abandon a project when the governance presents an insurmountable hurdle to your chosen approach.

Data Modelling Methodology

Deciding upfront which data methodology you will be using to create your staging and presentation layers in your analytics environment means that you can build in a modular fashion while keeping a standard way of design and implementation. This will enable the greatest synergies between data sources brought in.

Reference Architecture

Having a reference architecture gives you the ability to know what infrastructure and software will be required for a given type of analytics work. Building out a roadmap of analytics opportunities will enable you to stand up technical environments only when they are required, saving money on environment and support costs. It will also give a single point of signoff for any enterprise, solution and technical architecture boards; reducing the time taken to navigate these tricky spaces as your platform rolls out.

Data Assets

Identifying the data assets available and required will help to prioritise the opportunities identified. If data is easily available, the cost of an opportunity will be lower and therefore will be easier to recoup a return on the investment. If data assets are missing, then understanding if and how they created or collated and added to the solution, driving the maturity of the data asset estate and facilitating future opportunities, is necessary.

Analytics Opportunities

Identifying business driven opportunities within the analytics space will allow targeted development, standing up the minimum viable dataset to deliver a specific analytics outcome. Tying this to the modular approach means that it is feasible to only bring in exactly what is required to deliver what is needed for delivering the opportunity while leaving all avenues of future expansion open.

This approach to modular, rather than monolithic BI/AI/Analytics platform development is ideal for organisations looking to gain the most benefit possible from smaller analytics budgets.

Local Government, NHS Trusts, Smaller central government organisations and QUANGOs can benefit greatly from this approach. Building only what is required to drive specific analytics requirements; with the forethought to tie disparate opportunities into a modular platform will deliver more benefit for money invested.

Matt Thompson will be exploring this further at Council 4.0 on Thursday 13th February in London. For more information and to register, please visit Council4.0

Council 4.0                  

Please note, this event is for public sector only