Seek Quick Wins With High Impact For Successful Data Analytics
The Harvard Business Review (HBR) is well-known for their leading research, presented in a format that is easy to digest. I was therefore pleased to see that they published a guide to data analytics and the five different approaches that most companies take when they build out a data science strategy. What’s really interesting about the article is that of the five common approaches, two almost always fail, two work partially, and one actually works. Here is a short summary of the different approaches outlined in the HBR:
- What problems do you need to solve? This is generally when the CEO hires a data scientist or sets up a team and they go off looking for problems to solve with data, but without any specific guidance from the business.
- Boil the ocean. Spending millions and trying to change the entire business without first fixing all the legacy data issues inside the company, therefore having almost no impact.
- Let a thousand flowers bloom. Embracing data analytics from the top level and encouraging each business unit to use it, but not forcing or guiding them on what they can achieve.
- Three years and $10m from now it will be great. Building out committees and workshops and designing change that may only take effect years in the future can possibly work, but often it will not. People lose interest and move on.
- Start with high-leverage business problems. To build value quickly, identify the quick wins where data analytics can have a direct impact on the business and seek these opportunities across every team.
As you might expect, it’s the first two that don’t work, the next two that can work sometimes, and the final approach that works best. It works because it is pragmatic. First you identify some projects that could really benefit from using data analytics and promote the impact that it has had on these areas. Other teams will naturally want to learn from the initial data analytics projects and will take it further.
This pragmatic approach and focus on quick wins is frequently used by consultants advising on major change programs so it is a surprise that with data analytics it is not often applied. I believe this is because there are so many fascinating use cases for data analytics that it looks like a done deal – it has to be good for the business so there is no need to run pilots or seek quick wins.
This is a dangerous approach for any business and reminds me of when CEOs rushed to buy expensive CRM systems without building any of the processes and people training they would need to make it work. Technology itself, including data analytics, is never a quick fix. You need to identify specific areas of the business where it could make the quickest impact and then communicate the results so people across the business buy into the new ideas and systems.