Predicting a Better Future for Students
Driving Innovation through Business Intelligence
Cloud-Based Collaboration Is the Key to Driving Efficiency,...
Unleashing the Power of Analytics in Academics
Tools That Allow Us To Turn Data Into Business Intelligence
Philip Garland, CIO, PricewaterhouseCoopers
Data Science Isn't Just Business Intelligence: Here's How to Cross...
Yohan Chin, VP Data Science, Tapjoy
How to Conquer the Last Mile of Analytics
Ritesh Ramesh, Data & Analytics Leader, Consumer Markets Vertical, PricewaterhouseCoopers (PwC)
Business Intelligence a Top Priority in 2016, Again?
By Ben Stewart, CIO, Stant USA Corporation
Why is it that providing good, timely information is so difficult? We have collected mountains of data in our ERP systems, various applications, spreadsheets and email. Today’s BI applications are much simpler to navigate, and load data from a wide variety of sources. They can run “in memory” or in the cloud, or both. So, why is it that so many companies struggle to put useful dashboards, automatic reporting, or actionable information into the hands of their workforce, managers, and executives?
Have you tried asking, “What can I give you that would help you do your job, supporting better insights and decisions?” It’s surprisingly difficult to get people to tell you what they need to see. Then if they do, frustrating to have it only satisfy for a minute until the first follow-up question is asked, which of course the report does not answer. Additionally, the report that is useful for one person is not for their peer, as it doesn’t provide the right nuance or detail.
So what can be done? How can you drive value back into your organization, converting data into information?
Let’s look at 3 different approaches that may help you frame your project.
"Data alignment from the shop floor to the top floor, with common calculations and common source"
1. Build a sample to get people excited about the technology and considering the possibilities; then ask them what metrics they’d like to see. This is a fairly common approach, but often results in the consumer not thinking through what they actually need, with all the dimensions that they will invariably require. The approach is best executed in an iterative manner, where you can quickly model what they ask, then adjust and tailor the results as they begin to digest the information.
2. Take an existing report that is used on a regular basis today. One that has been proven useful, and would save the producer of the report time by generating it automatically. Be sure to fully understand the source data, calculations, and time dimensions. Ask the audience of that report today if it truly provides all the information that they need, or if they need something more. Ensure that your new delivery method (dashboard, email, self-service) will have as good or better ease of use and accessibility.
3. Benchmark best-in-class reporting from leaders in your industry. It could be that your company is going to great lengths to produce some reporting today, that really doesn’t drive any better results. Your BI toolset may offer some standard content that would benefit your company. Be wary of flashy technology driven by marketing rather than well thought, proven reporting. Metrics like inventory turns, AR/AP Aging, profitability, Sales vs. Budget, gross margin, and others are foundational at most companies and make a great starting point. Come to agreement on the specific calculations that your owners, or executives want to see. Be prepared to provide drill down to support investigation of the data. You’ll need to be able to answer the first or second question that the information will be highlighting.
Just as software companies have embraced concepts like Agile Development, be prepared to iterate through this process with your users. The shorter the cycle in which you can get feedback from the user, make changes, and get new feedback will result in less wasted time building something that’s not needed. There’s nothing worse than spending countless hours building the perfect report or product, only to find out that you missed the mark shortly after the unveiling. Fail small, correct small.
What about data integrity? As easy as it is to miss the mark on the right format and dimensions, it’s even easier to have the right report with data that is not believable. Once the credibility of the data is shot, people will not trust the tool and move on to other methods. Whether it’s the calculations performed to aggregate the data, extraction routines, or the source data itself; time spent validating the data early and often will be well spent. That doesn’t mean the source data has to be perfect. If your extraction and aggregation of the data is accurate, you can use your new reporting to draw attention to bad business processes, or daily execution errors and drive improvement at the point of entry.
Metrics Driven Management
You’ve got the data, you’ve transformed it into useful information, and you’ve validated that it’s accurate. Will it sit in someone’s inbox unread? Will it drive decision making and change? To make this effort worthwhile people have to use the reporting. It needs to permeate their daily work. A dashboard that starts up with their computer and draws them into any anomalies or unexpected results. Meetings that are run, not based on PowerPoints and Excel manipulations, but spin free automatic reporting based on the data in the system—data alignment from the shop floor to the top floor, with common calculations and common source. One version of the truth—people held accountable to what the data shows, fixing it if it’s wrong, acting on it if it’s accurate, but unsatisfactory.
Decisions based on facts and analysis can become part of your corporate culture, only if driven by management through regular, repeatable review. Make it the expectation, not the exception. It’s not easy, but it’s worth it. If it was easy, you wouldn’t be trying again in 2016.