Empowering Analytics through Change



Published on March 27, 2018

So, you’ve been using analytics at your organization for some time now and are comfortable with the value it brings. It’s doing what you set out for it to do and has met your expectations. You really haven’t given it two thoughts since you implemented it. But maybe you should… Have you truly unlocked its potential? Have you challenged yourselves to say, “Is this really all we can do?”

Using analytics in an organization should not simply be a check-the-box kind of exercise. Analytics has gotten a lot of focus, but simply implementing it doesn’t make it effective.

The world of analytics is constantly evolving, not only because of the amazing new technologies that are becoming available every day, but by the new ways data can be captured as well.

The true key to unlocking the potential of analytics is picking the right data fields to use and structuring the relationships of those data fields to deliver real insight. And let’s be honest—we’re not talking about analyzing just one relationship, but getting several different views of your organization using different data fields to layer on in order to provide context and meaning to what you’re seeing. Do you feel you’ve hit this mark yet?

With anything that’s new, most people and organizations by nature tend to take baby steps, testing the waters before they dive in. It’s the same with using analytics. You start off with small, basic analytics that make a superficial yet important impact. Is it time to consider taking a bigger leap? Before you do, here are some things to consider.

When you set out to implement analytics in your organization, it may have been a “let’s see what this can do” type of plan. You focused on one particular area with data you already had and created something that was rather great at the time. But did you really take the time to understand the particular process or reason behind why you wanted to create an analytics solution in the first place? What was the challenge or issue that you faced? Did you take a step back to analyze the situation and rethink how that process could be performed or this issue could be handled? Let’s face it, implementing a new tool to solve a problem is good, but turning the problem on its head and solving for that,  drives change. How deeply did you rethink about how a process is done or, better yet, how a process could be done?

Now I’m sure you’re saying, “Easier said than done.” And, of course, that’s very true. It’s not easy to try to rethink the whole process, not to mention reinventing an approach to something you’re so close to. And forget about finding the time and the budget. But if you did that, could…

When you start thinking about “coulds,” a whole new world is unlocked with endless possibilities.

So let’s think about “coulds” for a minute and explore how some of the other emerging technologies can be used in an organization to help drive change and empower the use of analytics.

You’ve all heard the latest buzz words around exponential technologies[1]—like RPA (Robotic Process Automation) and the IoT (Internet of Things). Let’s explore how the use of RPA and the IoT can begin to change the way things get done, how data gets collected and how that can open up a whole new world of possibilities for analytics and change within your organization.

RPA is a process that focuses only on structured data[2] and represents a way to perform rule-based tasks using a software, mimicking a human action and enabling higher operational efficiencies for organizations. The beauty of RPA is that it sits on top of existing systems, which means that if your organization has, as many have, evolved over time and has acquired many different and separate legacy systems, the RPA can essentially liaise with all of them easily. This makes it appear that the systems are more integrated and can better talk to each other without a massive initiative to align all of the systems onto one platform, saving both significant time and money. The RPA software can work across the various systems, accessing the required information from each one and performing routine tasks.

So how does this help with analytics? Let’s consider a brief scenario:

We have a manufacturing organization that produces machines. It has a different system for its revenue subledger, its general ledger system, shipping system and purchase order system. The volume of shipments every month is substantial and peak sales periods are consistent year to year. Because the systems do not integrate well, Beth, our sales clerk, spends an entire day every month-end pulling the invoice data from the revenue subledger and the shipping documents from the shipping system to manually match the goods shipped to the sales recorded. One reason why Beth has to do this is that inventory discrepancies were noted in the past because shipping documents and invoices are not automatically generated.

When the organization implemented analytics a few years back, their biggest concern was around ensuring that revenue was recognized on all goods shipped. So to address this issue, they developed an analytic based on the revenue data, comparing prior period to current period by both dollar and quantity based on the invoice data. This information provided a high-level analysis of the sales information that the manager used to help determine if there were any major discrepancies, expecting that sales by period would be similar to the prior year’s and therefore at a high level and that all goods shipped were recorded. Updating the analytic was time-consuming, so it was only done on a quarterly basis.

Now, let’s see how we can change this scenario by rethinking the issue and the process at hand. Rather than letting Beth toil away at the matching exercise for a whole day every month, the RPA performs it in a matter of minutes. The RPA software can automatically perform the matching of shipping documents to invoices and, as an additional dimension, it can also match the purchase orders so that there would be a three-way match of electronic purchase orders, shipping documents, and invoices.

The RPA provides the list of exceptions to the matching process so that Beth can now focus her attention on investigating them rather than performing the matching. Now that the RPA is able to pull the required data from the different systems in a consistent format, an analytic can be created that can be automatically updated every period using the same data. This analytic can now provide a more granular level of detail by highlighting the exceptions in matching for Beth, and can also enable her manager to explore the same aggregated view of the sales trend period-over-period.  She can now drill into that data to view it by product and by customer. In addition, the credit risk manager can also use the same analytic to view revenue by product and by customer to identify where there are decreasing sales and increasing receivables, and to identify potential collection issues. The production manager can use the information to understand customer demand by product and by period to better schedule production times and drive operational efficiencies. Sound possible?

The other exponential technology that can be used to change how things are done and help evolve analytics in an organization is IoT. IoT is simply the network of physical objects connected to the internet such as wearable devices, vehicles, equipment, buildings and just about any other thing that can be embedded with electronics, software, sensors and network connectivity. How does IoT work? Well, it simply allows connected devices to talk to each other. For instance, if there were IoT devices in a vending machine that connects to the supplier and purchasing clerk, then if the vending machine gets low on inventory, it can automatically send a purchase order to the supplier and notify the purchasing clerk that an order was placed and to schedule a refill of the machine. IoT has many potential use cases and can potentially do some pretty cool things, but how does that help with analytics? Because IoT devices collect information, it means that there is a whole new data set to consider that can help augment existing analytics or can be used to create some new, powerful ones.

Let’s continue using our previous example. The organization that Beth works for installed IoT devices, RFID (radio-frequency identification) tags and other sensors, on all finished goods. Now when a good is shipped from the warehouse, the IoT sensor tracks the data and the RPA software removes the finished good from inventory and the sale is recorded. The invoice is automatically processed and sent to the customer. Whenever Beth has to investigate any discrepancies in the matching done by the RPA, using RFID tags and other sensors on the finished goods she can determine whether the goods are in the warehouse and if so, where the goods are located.  These RFID tags are also used to significantly streamline their inventory count process.

Once the machine is received and installed by the customer, the IoT devices generate information that can be monitored from Beth’s organization related to the wear and tear of the machine.  It notifies them automatically when parts need to be replaced before the machine breaks down to reduce downtime at the customer site. Now, with the new information that the IoT devices produce, new possibilities open up in terms of their analytics. For instance, based on the diagnostics of the machines at their customers’ sites by the IoT devices, the management team can predict when replacement parts will be required or machines replaced, allowing them to be one step ahead and ensuring they have the right parts and machines on hand just-in-time. It also enables them to predict their sales and production levels so that rather than the manager looking at what had happened in terms of sales, they can focus on what will happen, generating a more efficient, well-run organization.

Now, compare where Beth and her manager started off with analytics, looking at the past rather than the future. What they were doing with analytics was good, but they didn’t have nearly the same potential to help evolve the organization. With the use of technologies like RPA and IoT, there are so many possibilities. Or does it all still seem rather far-fetched? Maybe it’s more realistic than you think, if only you could...

If we continue to challenge ourselves on how we do things and how exponential technologies can be combined with analytics to drive change through an organization and make a substantial impact, maybe those “coulds” don’t seem that far out of reach after all.

Our Insight Driven Organization Labs are a proven way to get your analytics journey off to a positive start. Our session with you will be highly interactive and use the latest design thinking techniques to help evaluate where and how analytics insights can be adopted to solve business challenges, leaving you with a clear direction to pursue analytics. Reach out to any of our Deloitte partners or to Nicole Deschamps to request a session.

[1] “Exponential technologies” is a phrase coined by Ray Kurzweil from Singularity University that refers to technologies where credible evidence exists of exponential growth in computer memory (Kryder’s law), processing power (Moore’s Law) and bandwidth (Nielsen’s law).

[2] Any data that resides in a fixed field in a record or file.  This would include data included in spreadsheets as well as relational databases.




Nicole Deschamps Nicole Deschamps
Senior manager, National Services - Innovation and Analytics and Member of the CPA Canada Audit Data Analytics Committee
Nicole has been in public accounting for 18 years, serving a wide variety of clients including both public and private, in the U.S. and Canada with a majority of clients operating in the manufacturing industry. She is responsible for the development of innovations and analytics to be used by audit practitioners in the performance of audits.



Correction list for hyphenation

These words serve as exceptions. Once entered, they are only hyphenated at the specified hyphenation points. Each word should be on a separate line.