When it comes to maximising the true value of an organisation’s data, many business leaders are quick to turn to the IT team to deliver solutions that will turn disparate siloes of Excel spreadsheets into meaningful insights. However, at best, this will result in your data project falling into the dreaded IT implementation queue, at worst, the solution that’s delivered is not fit for purpose.
The journey must begin with detailed, contextual insights and active reporting, explains Jelle de Jong, CEO of Lexia Analytics, who shares the steps a business needs to take before approaching IT to make the journey to data equilibrium much quicker and more straightforward.
Five steps to take before approaching IT
It is counter-productive and counter-intuitive to suggest first setting up the correct system and processes before proper reporting can be introduced, which will often result in delaying the quick progress that can in fact be made early on.
1. Establish the most important questions that will need answering by your data implementation. This is a much more agile approach which will provide the material difference to effective decision making
2. Build a first draft analysis, often in Excel or a simple SQL database. Only once this report or analysis has proven its work and is being used as expected can a system solution be considered. This also means any system requirements and design to structurally deliver such a report or analysis is significantly informed by the work done to date
3. Work with the actual users of the system and see how this helps them. Success breeds success and often to show 1-2 examples of how the data / insight can be converted into actual bottom-line benefit goes a long way
4. Keep working in paper and basic xls templates for at least six months before you start to think about a real hard-coded solution. This will enable you to find and address issues before the IT deployment begins
5. Start small, scale afterwards. It is much better to build the basic functionality first and only once this is up-and-running to move towards a hard-coded solution
With the route to IT now mapped out, it’s important to drill down into each of those steps, starting with identifying the top three specific questions that will help inform the system requirements needed from any IT implementation. But how do you successfully achieve this?
How to identify specific questions to inform IT
IT developments are often like pouring concrete. In the pouring phase, every shape is achievable but once it has set, it is fixed. As such, it is important to identify what the mini-steps are rather than trying to define an overall end-result in advance.
With analytics and overall data solutions, rather than build a major data repository, only to establish how you are going to use it after it has been built, it is much better to start small. For example, build a small ‘data lake’ and work that first.
The key question you should ask therefore is: what are the most obvious, easy-to-action use cases that we can identify with minimal IT involvement in order to prove the relevance to the business?
Analysing and using the first draft analysis
Now that you have your first cleansed and structured database with crucial data to address one of the pre-identified use cases, it is up to the analysis team and the business stakeholders to show that this approach to solving business issues has merit. For example, the first overview of all pricing data and profitability must show the business how it can increase profitability by developing a better pricing structure. Similarly, the first aggregation of all production data for a packaging line in the factory must lead to improved productivity and reduced downtime.
Only once these initial wins become visible and tangible can you move to scaling up and discussing internally how to build a more sustainable solution.
Once you’re clear on the journey you are taking, it’s important to know when the time is right to mobilise the IT department.
How to know when to get IT involved
It always makes sense to engage IT relatively early on as they have a unique perspective on what is easily achievable and what is not. However, moving from ad hoc, fit-for-purpose solutions to a real system implementation should only happen when it is absolutely clear what a successful first implementation would be and what the longer-term roadmap might look like.
Agile, small improvements and builds can then start where, ideally, every step itself is a viable solution that already has value for the business. 2-3 months is the average timescale before fully involving IT.
The age-old phrase ‘fail to prepare, prepare to fail’ is critical in data solution deployments. It’s critical for leaders to start at the beginning, with detailed context that will help inform accurate representations of need. Following the five steps will help ensure your brief to IT will deliver a successful solution that meets the needs of the business in terms of data, analysis and insight-driven decision making.