If you’re introducing advanced data analytics/AI/ML you may find that the biggest challenge is getting started. How can you prove value before investing large amounts of time and money?
We find that the nine principles below help to show benefits quickly and prioritise efforts. By following them, you can develop and evolve your digital strategy in stages. These will be more manageable and each will deliver measurable value.
1 Identify specific pains or opportunities
Define a problem or opportunity that improved use of data would solve or take advantage of. Predict the financial value of any potential improvements. Then plan a short, focused piece of work, otherwise known as a sprint, to deliver a basic but fully functional system. Use this to prove the concept. Test your prediction of the financial benefit by giving users access to the system and gathering their feedback.
2 Be quick and cost-effective
Be prepared to start small with opportunities that are well understood. Look for the low hanging fruit and quick wins. Work in timescales of weeks and months, not quarters and years. Run multiple sprints in parallel using different teams so that in a short space of time you gain an understanding of wide-scale opportunities.
3 Be flexible
When needed, break more complex problems down into a short series of sprints. Focus on the decision support that is needed first and build a prototype for user feedback. Then add analytics to interpret the data. Only integrate the data you need to prove the value of the concept you are testing. This can be scaled later.
4 Scale your wins
Once you have proven a concept, look to improve your results. Plan extra sprints to extend the data analysis and decision support capabilities and continue to gather feedback. Build your systems in stages to add specific value with each sprint. And make results operational as soon as possible to begin making returns on your investment. New features and capabilities can be added over time whilst systems are in use. If a concept can’t be proven, don’t spend further time and resource on it; try another.
5 Work with users
Your digital strategy is not about the IT department or innovation teams, it is about giving decision makers the ability to improve your bottom line. So your users should not be an afterthought. They must be involved. Don’t be scared to do this from day one. They are the customers for what you are building and their buy-in for change is needed. They also have the domain expertise to explain how data can be interpreted and analytics developed. Do not expect to succeed with big data techniques alone.
6 Deliver in a common way
Don’t overcomplicate things by writing a brand new system for every sprint. Be smart. Use platforms and advances in software technology to minimise the work you do each time. This keeps your speed of delivery up and your costs down. A common platform will be more likely to engage your users since they only need to learn one system. It also makes it easier to combine results later since everything is in one place.
7 Every analytic innovation makes a difference
Make use of expertise from all potential sources of data science – internal teams, third party developers, universities and any others. Make sure to choose a platform that can deliver analytics written in any modern data science language (e.g. MATLAB, Python, R, etc). And preferably choose one where these can be deployed without extensive IS and IT support. This will allow developers to use the tools they feel most comfortable with and ensure your sprints proceed at pace.
8 Partner with experts
You may have the skills and resource to deliver your digital strategy in-house. But if you look for external support make sure to choose an appropriate partner. Look for those with an understanding of industry challenges. They should have expertise in interpreting complex operational data and a track record of delivering useable decision support. You are likely to need a partner, not a consultant or standalone developer. Do not underestimate the importance of your partner’s domain knowledge.
9 Build your corporate digital strategy from the ground up
Before starting this process, set high-level goals for your corporate digital strategy. But do not try to define exactly how these will be achieved from the start. Keep your vision at the centre of each sprint but allow yourself the flexibility to try new ideas and approaches. By doing so, you will be able to adapt quickly to changes in markets, regulation and technology. This will greatly increase your chances of overall success.