Getting Started with Analytics: What’s the Upfront Investment?

Everyone wants to leverage analytics, but should everyone dive into the deep end right away? Heed some sensible advice on getting started with analytics, and assessing the true upfront investment.

By Christina Prevalsky, FI Consulting.

Do we need large systems and expensive BI tools to use analytics? Maybe. But, not to start.

Analytics wordcloud

Depending on the scale and complexity of your organization, your end state might require a larger IT implementation. Your evaluation and requirements phases don’t have to involve that level of investment.

Start Small. Start Free. Start Open Source.

It’s tempting to start with gathering up all of the data available. Organizations are full of data, systems we use generate data we don’t even keep. Focus on the main questions you’re trying to answer and start with a subset of data. Start small enough that the data is easy to manage and manipulate, yet large enough that you might spot a trend or two. In designing your data model, remember that in the long run you want to set up your data to be used more broadly than one specific case.

Use the resources you already have available. You don’t need to install the latest and greatest BI tool or build out a full system until you know where you’re trying to go. If you have these tools at hand, go for it, but standard desktop applications have come a long way from your standard bar and line charts. You might be surprised what’s possible with just Excel, VBA, and a little imagination. For example, this portfolio analysis tool:

Portfolio analysis

If you want to take it up a notch, try out PowerPivot or Microsoft’s Power BI. Here’s a nifty tool I came across a few weeks back for building simple models to estimate uncertainty around your forecasts: Guesstimate. For the slightly more technically savvy crowd, you might try some of the open source analytical tools. I’m partial to R, but there’s plenty of options out there and they’re adding new analytics capabilities almost daily. For a sample, check out what New Zealand’s Ministry of Business has put together using R and Shiny.

Take a few lessons from methodologies like Lean Startup and Design Thinking. Take an iterative approach and gather feedback early and often. You’re organization’s strategy will undoubtedly change over time; expect changes. Be creative and make sure your end product is providing something actionable that provides a story rather than statistics for the sake of statistics.

What should you invest in upfront?

  • Understanding what kinds of analytics you really need.
  • Getting buy in from the right stakeholders. All of them.
  • Making sure you are using the right data, in the right way.
  • Developing an iterative implementation plan.
  • Building analytics into the process.

Bio: Christina Prevalsky leads teams that perform custom analytics and develop applications to support data-driven decision making for government and private sector clients. She specializes in implementing process improvements and analytical solutions.

Original. Reposted with permission.