All businesses collect data — from their CRM systems, logistics, operations, marketing and plenty more. And everybody knows their data is valuable — but what’s lacking for most is a clear vision (or even a fuzzy one) of the path they need to travel in order to start extracting value from their data before it goes mouldy (that’s right, data has a shelf life). To that end, taking the following three steps will ensure your data journey gets started right.
Step 1: Create a Data Strategy
Without an overall data strategy, different parts of an organisation tend to view data in a way that’s only relevant to their specific needs — applying different formats and protocols to its collection, security and storage. In such a fragmented state the potential to leverage this data in analytics projects across a business is lost. What a data strategy does, however, is formalise data governance and collection in a way that best aligns with the business’s objectives. With a data strategy in place the following benefits accrue:
- Data is formatted uniformly — meaning that the foundational data required for any initiative across the business is clean and will ‘fit’ the format required.
- Similarly, when any data is shared between departments, all participants will be speaking the same data language.
- The data component of operational and organisational metrics & KPIs will be common across the business.
- Shortfalls and overlaps in infrastructure and technology are identified — enabling more efficient investments in data assets.
Tony Mitchell speaking on February 15. Credit: Courtney Devereux.
Step 2: Identify Your High Value Analytics Projects
Here’s what to do to ensure you’re not wasting time on projects that don’t matter.
- List your big problems
Whatever they are you’re already aware of them and it shouldn’t take long to come up with a list. What is important, though, is to not overthink them at this stage, just get them written down — the analysis will come later.
- Remove non-starters
Now, filter your list by asking which problems have an analytics component? The fact is for some problems analytics won’t be a factor — for example, a policy change by an international parent company. So shorten your list to only those problems where analytics will assist. A tip? If there is a spreadsheet or a database involved in the problem, then analytics probably has a role to play in the solution.
- Establish where analytics will fit into the process
Identify the stages of the business process that underlies these problems where analytics could have an impact on the quality of decision making — and ask two questions.
1. What would be the likely value of improving those decisions with analytics?
2. What would be the likely cost of undertaking the required analysis to improve those decisions?
- Decision time
You’ve identified your big problems with an analytics component and you’ve pinpointed where data and decision making intersect with the business processes underlying those issues. You’ve also quantified the potential value of improving those decisions and the analytics cost of doing so. Now do the maths and scrap anything with a mediocre or minimal return. Congratulations, what’s left are your high-value analytics projects.
Step 3: Get Started – Fast
Historically one of the main reasons data projects fail is that they simply aren’t completed quickly enough. It’s an easy trap to fall into — after all, ‘big data’ requires big solutions right? Wrong. Don’t try and build a mega-solution. Instead, take your list of high-value projects and refine one down to a single clearly defined use case. Then scope out how you’re going to complete that in 90 days or less (call me if you need a hand). Now repeat steps 2 & 3 ten times and your Big Data problems won’t seem anywhere near as big.
