3 Reasons to Practice Agile Analytics
Are you using an agile approach to analytics?
If not, here are 3 reasons to practice agile analytics.
I’ve worked in analytics for 15 years and now run my own consulting company. I use an agile analytics framework for MOST of the work that I do.
Agile is a development process focused on flexibility and adapting to changes constantly. From a development standpoint, it defines the end goal, but doesn’t try to fully plan
When you think of a development process - let’s say developing and manufacturing a new car model - you’re likely more familiar with the waterfall methodology.
What is the waterfall method?
With waterfall, you define your goal and timeline and then fill in all the details before you start. You ID milestones, what work will take place, how much it will cost, who will be involved.
Waterfall focuses on thorough planning to avoid changes along the way.
This work happens sequentially - one stage gets done, then you move onto the next.
When executed well, the waterfall approach is ideal for development projects where the production costs are high and methods inflexible.
For instance - that car manufacturer where parts need to be sourced well in advance, assembly lines switched to new products often with some adaptations required, testing needs to take place to certify all regulations are met.
It can be VERY expensive to make changes at the end of the process.
What is agile?
The popular method for software development - and increasingly other development projects - is agile.
Agile still sets an end goal, total timeframe, and assigns a budget.
But instead of trying to outline everything at once before any work can start, agile ASSUMES there are going to be changes.
Agile focuses on many small deliveries giving the user the opportunity to test, provide feedback, and then for development to be adjusted along the way.
Agile looks to create a Minimum Viable Product as fast as possible so it can be tested and feedback can start that will inform the rest of the process.
Both approaches show up in the analytics process and each can have their place.
3 reasons to practice agile analytics:
1 - Agile focuses on value.
Agile analytics looks for what the data can tell us rather than focusing on proving a hypothesis.
There’s a significant amount of data analytics and data science that’s focused on proving hypotheses.
However, if you think about the buzzwords and direction of the field, it’s not about creating the best static reports that are backwards looking.
It’s topics like machine learning and AI that dominate.
These are areas that are agile. They constantly build on the information they’re fed to create a new and updated state.
2 - Agile assumes the future may be different.
Agile is all about learning and adapting quickly.
It assumes that just because you did something one way in the past - wanted a specific piece of data in a specific context - doesn’t mean that’s how you’ll always want it.
This doesn’t mean everything is constantly new and different. It DOES mean that you plan for things to be iteratively improved and adapted to the most current needs over time.
Think about the disruption caused in the past year with once-in-a-lifetime events. The result for many businesses has been throwing out their predictions.
The models that worked in the past aren’t necessarily reliable right now.
Agile analytics adapts to that, looking for solutions and ways to change the approach to compensate.
3 - Agile works with users.
For analytics, the users of your work may be people using queries or reports you’ve built for their day to day work or the users may be executives making decisions on the direction of the company.
Agile focuses on user satisfaction by involving them through development.
You can start with building a basic analysis framework and getting feedback from those who will use the information.
Does it contain everything they need?
Are there other pieces of data that would be helpful?
Is that interactive dashboard you created easy to use correctly?
Are there ways for your users to explore the data themselves in a controlled environment?
I don’t think everyone having full access to all the data is ideal.
There are so many things that can go wrong when people don’t understand how datasets interact with each other.
However, when you create a framework that lets others explore the data on their own and minimize the opportunity to come to bad conclusions, you’re creating an environment that encourages turning to the data for answers.
If you want to implement agile analytics, but you’re not quite sure how to use it in your circumstances, book a coaching call with me and I can help you with an implementation plan.