Taking an agile approach to analytics for a better, faster outcome

Taking an agile approach to analytics for a better, faster outcome

The key to agile analytics is speed and avoiding the need to always get things perfect, writes Daniel Parr of Data Insight.

Start small, create a basic data mart, then encourage your analyst to deliver speed over perfection.

The world has changed rapidly in the past few years thanks to technology and a new ‘always on’ mentality. As a result, we have seen a dramatic shift in consumer behaviour which has put them more firmly in control.

A recent Forrester report shows that 40 per cent of consumers have a high willingness and ability to shift spend, with an additional 25 per cent building that mind-set.

So how can companies respond quickly to their customer’s needs and create opportunities that result in the right action in that moment of truth?

The answer lies in the beauty of data and how you can manipulate large volumes of it to know your customers better and deliver more relevant and engaging experiences to them so they never want to leave.

But just manipulating data on its own isn’t the answer – the real secret lies in leveraging agile analytics to manipulate data quickly to extract value while it is relevant, and then actioning it.

Put simply, agile analytics delivers insights quicker than traditional analytics.

Agile analytics borrows principles from agile software development that allow for better analytics solutions to be delivered faster. The main principles of agile analytics are collaboration and feedback loops, iteration and cross-functional teams.

Traditional analytics has focused on manipulating the data until it is 100 per cent correct, building the perfect model, making sure the product is complete before delivering. But the focus on accuracy comes at the expense of speed, which in a fast, evolving world means opportunities are missed.

In order for businesses to deliver agile analytics, they need the right infrastructure and the right culture in place, empowering analytics teams to utilise agile processes to deliver insights quickly.

...Then learn from your mistakes, and continually improve

Daniel Parr, Data Insight

The right infrastructure provides the foundation for agile analytics

Often analysts will be required to extract data from multiple systems, merge the data, and define metrics, which are time-consuming and inevitably differ every time they are run. By spending a little on data and infrastructure upfront, you can make big savings later by being able to deliver faster analytics.

The right infrastructure for analytics has the following properties:

  • It doesn’t rely on overhauling existing processes – it creates an analytics layer that sits over top of large BI and source systems to utilise the existing data

  • Queries run quickly

  • It doesn’t rely on unnecessary, expensive, proprietary tools – it is better to use a cheaper, simpler solution such as SQL and R to deliver returns quickly and economically

Empowering analysts to deliver analytics ‘fast’, not ‘perfect’, a cultural paradigm

The key to agile analytics is speed and avoiding the need to always get things perfect. The 80:20 rule applies – it is better to get 80 per cent accuracy in a much smaller timeframe, than to spend months in isolation perfecting the analysis which, once implemented, may not be the right solution.

What works is to foster a culture that encourages the delivery of analytics projects quickly and to incorporate learnings into future projects. We think it’s better to prioritise faster delivery over getting perfect results and strip away unnecessary requirements to keep a tight scope on the project. The benefits of this approach are that more projects are being delivered, and there is a culture of speed and constant improvement.

Agile software development principles ensure better analytics are delivered quicker

Collaboration and feedback loops means you fail fast.

Often your stakeholder doesn’t know what they want until they have something tangible in their hands that they can read and critique. Waiting until the end of a project to show them means that there is no time or budget left to action any requested changes. A good starting platform is collaborating early with the client by providing a polished draft early in the process from which feedback can be delivered, before valuable time has been spent analysing the wrong things.

Iteration allows for constant improvement while diving straight in. By embracing an iterative approach, the upfront planning can be reduced to a minimum, enabling the analysis to begin faster.

Agile analytics has many benefits – increased speed, improved accuracy through feedback and collaboration.

If you’re sold on the idea of agile analytics but wondering where to begin, the best place to start, using the principle of agile analytics, is anywhere. Start small, create a basic data mart, then encourage your analyst to deliver speed over perfection. Then learn from your mistakes, and continually improve. Once you do that, you’re already practicing agile analytics!

Daniel Parr is head of insights at Data Insight. He has over 10 years of experience in data analytics. to joining Data Insight in 2014, He worked at MoneySupermarket and dunnhumby in the UK and then Spark NZ.

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the newsletter!

Error: Please check your email address.

Tags big dataanalyticsData InsightDaniel Parr

More about AgileInsightSpark

Show Comments