Insights latest news from the field


Build Analytics In A Day?

Can you really build analytics in a day? With a little creativity, know-how, and a willingness to let your solution evolve over time, the answer is yes.

Many data analytics projects take months, quarters, and some years to start to bring value to the business. In that time, the needs of the business have moved on and all the invested work diminishes in value. That just builds frustration from the business and the IT team. What both need is the ability to move quickly, learn what the data can tell you, allow learning from “failure”, and then run at it again.

A recent project of ours had a client team with very fuzzy requirements of what they wanted. There were a lot of words like “kinda thinking this” and “be nice if I could measure that” and all sort of other fluff words. Words that should scare most delivery teams. Not ours though; we know how to work with ambiguity. To us, that just means we all need to take a journey to learn the story the data has to tell us.

We were fortunate that the client had dug up some old code from reports that pointed to the transaction data they wanted. Right away we saw an issue. One set of data was captured at the day level and the other was at the month. Problem? Not really, when we investigated that data and the measures they could provide, we decided that they were really describing two separate facts.

“But I need a report that gives me the sales transaction from fact A and cost data from fact B”

Problem? Not at all. “What is the common ‘thing’ you want them to be measured by?” We ask. The answer is our widgets. Easy. We build a common dimension (Kimballism calls that the conformed dimension, but clients don’t care who Kimball is so…) for widgets that takes the best of both system’s description of a widget.

From there we start building some rudimentary dimensions to describe the different types of transactions. Nothing fancy, but, fancier than where the client started. Create some date dimensions to describe the various events such as transaction, invoice, order pulled, shipped, and other sundry dates. With that, we now have the foundation for a set of reports that can look at how they are performing today, how they performed last week, month, quarter, and year. We can now calculate how they are performing year to date and year over year to date. The data even allows us to project what seasonal widgets are likely to be selling in the coming weeks to allow the client’s sales team a heads up on what they should be focusing on in front of them.

Could we make all this better? Absolutely we could. But, as analytics consultants, our job is to get data in the hands of our customers early. A customer with data is empowered with the facts they need to make immediate changes to their business. They now see value in their data and have a strong business case to continue to build out their analytics into a more robust tool that becomes the fabric of their success.

It all starts with a focus on get data into the hands of the client quick, see what story that data is telling you, and then start acting on data to drive business. It can be achieved.

Get started with Eris Enterprises today

Contact us

Eris Admin