Taking spreadsheets out of the supply planning process at Extreme Networks

The use of spreadsheets for ad-hoc data integration is the bane of the modern enterprise. Spreadsheets have their uses as a temporary workaround, but there are multiple downsides to embedding them in routine business processes. They tie people up in unproductive manual processes, which in turn slows down the transmission of data, while errors arise easily and often go undetected. Yet in many organizations they remain an integral part of crucial business operations such as supply chain operations and planning, where data is spread across many disparate applications with no unified data store.

This was the case five years ago at Extreme Networks, a digital network equipment manufacturer and software provider headquartered in San Jose, California, which posted revenues of $1 billion in its last financial year. Guy Currin, RapidResponse Platform Manager at the company, explained how it has removed multiple spreadsheets from its supply chain planning in a talk at the Big Ideas in Supply Chain conference, which was hosted in Ottawa last week by Kinaxis, the maker of RapidResponse supply chain planning software. Speaking about the time before the company adopted the Kinaxis solution, he recalls:

Excel was our integration tool. We had a lot of great tools. Behind the scenes, we had Oracle, we had Salesforce, we had some of the E2open products. But none of that was integrated. If we wanted to be able to integrate that, we had to take data out of one system, manipulate it on the spreadsheet, load it into another system. As a result, there was no single instance of truth. We didn't have any place that you could look to at any one time and see what the total picture looked like.

So we had to do a lot of manual work or workarounds. As my boss calls it, we had to rely a lot on human putty — if you had a gap, then just throw people at it and use them to manipulate more spreadsheets, or do the things that they needed to do. We always made it work. We always made our business at the end of the quarter. We always made it happen somehow. But it was extremely difficult to do that.

Scaling up with Kinaxis

Crunch time came in early 2017, when Extreme Networks had to digest the acquisition of Brocade and an Avaya business unit, doubling volumes. The supply chain team recognized that its current processes and technologies would not be able to scale up to cope with this expansion. They began looking for a solution that could replace this spreadsheet-based integration. At the time, Avaya was already a Kinaxis customer, and it seemed to be working well. So the decision was taken to adopt the Kinaxis software at Extreme Networks at the same time as completing the two acquisitions, alongside a data warehouse implementation that would bring all the data together without the aid of spreadsheets.

The supply chain organization also set up its first Center of Excellence in preparation for the implementation in early 2018, which is where Currin joined the team. The goal was to bring all the data together to enable effective planning in the RapidResponse tool. He explains:

We were going to let Oracle and Kinaxis be the center of the universe in supply chain. That gave us the connectivity that we needed to these other great tools that we had, but they just weren't interconnected. So Kinaxis was going to be the thing that pulled all of that together. Real-time data was going to be one of the key benefits out of that, we were going to be able to see everything, we would then know what the real moment of truth, what the instance of truth was, at the time.

Several different systems each feed data into a core data warehouse in the new setup. Oracle is the transaction system of record. There is a 3PL whose system manages the supply chain, along with Salesforce as the system of record for sales. There are two E2open tools — Zyme to pull data from distributors, while Clari analyzes Salesforce data in the data warehouse to calculate the probability that a Salesforce order is going to crystallize. A Talend system then pulls all the data out of the data warehouse for analysis in the Kinaxis platform. An important new feature of the implentation is a connection back to the Oracle system, through which Kinaxis provides promise dates back to customers. At the same time it sends Vendor Managed Inventory (VMI) triggers back to suppliers so that they can pull inventory and ship it to Extreme's warehouses at just the right time.

Continuous evolution

Having successfully completed the implementation, work has continued to build out the capabilities of the system. Currin comments:

It's not just a one and done. You don't implement, which we did in 2018, and then you're done with it, because our users just have a lot of fantastic ideas — or there are a lot of disruptions that are out there — that cause us to continue to do development on top of the platform.

The nice thing about the Kinaxis platform is that it's very easy to build things on top of it without impacting that base functionality that the system does. And, as new releases of the software come along, we're able to put those new releases in and all of these enhancements that we do continue to work.

One example of those disruptions is that, shortly after implementation, the US government introduced tariffs on equipment coming from China that was sold in the US. This meant adding a new capability to the system that would identify where items were sourced from and ensured that, where possible, inventory from China was sold outside the US, while US orders were fulfilled from other suppliers.

Critical component tracking

As the system continues to evolve, two new features have been added, each of which has allowed even more spreadsheets to be retired. One is critical component tracking, which improves visibility into lead times of critical components that OEMs use in the SKUs that Extreme Networks supplies to its customers. The experience of supply chain disruption during the pandemic illustrated the importance of having this level of visibility into the supply chain. Currin elaborates:

We're not a manufacturer. We buy and sell at the SKU level. But these components are a level down from that. So our initial implementation to RapidResponse did not drive down to component levels — we didn't need to. But here we are now with critical components, as a result of COVID, that is causing us to need to plan better for those components, or at least be aware of those components. We couldn't see the impact to what our revenue was going to be if one of those components didn't come to fruition and wasn't delivered on time. We had a very laborious, manual, Excel-based process that allowed us to track that. We could pull demand out of RapidResponse at the SKU level, but then we didn't know how those components were affected by that.

For example, there are 180 critical components in the Bill of Material (BOM) used by one manufacturer that affect 400 SKUs. Those critical component BOMs are now tracked within Kinaxis, allowing more insightful planning at a SKU level. He adds:

Just normal RapidResponse processing allows us to now see that component level impact and what it does at the upper level of SKUs. We can play all kinds of what-ifs with that — what if this one doesn't happen, what is that going to do to us? That's proved to be very effective. So we've replaced these manual spreadsheets, we've automatically got the data calculating in RapidResponse, just using base functionality to do that. The ability to perform what-ifs is just, I can't overemphasize enough how important that is within RapidResponse.

Planning for opportunities

The other new feature reserves inventory for pipeline opportunities. While there was visibility of opportunities in RapidResponse, this was only as a reference resource rather than something that could be planned against — the system could only plan for confirmed sales orders. This meant that, where it was important to reserve inventory for important opportunities, this was being done manually. Currin explains:

We were reserving this inventory manually when we would buy it. We were managing it on a Smartsheet. When inventory came in, we had to set it aside in a reserve location. We were just constantly adjusting that, as inventory came in, as supply came in as, as opportunities became real orders ...

We needed to be able to eliminate these Smartsheets that we were trying to track all of this information on at the time. We wanted to be able to eliminate this manual inventory that we were having to set aside and reserve.

The new capability means that a planner can set up a separate scenario in RapidResponse, where selected opportunities can be assigned as sales orders. The planner can then see the impact of those opportunities alongside confirmed orders. Currin concludes:

It works great. Mo more manual Smartsheet, no manual inventory reserves, and inventories prioritized to the correct orders or opportunities as we wanted it to be.

My take

It's good to see yet more spreadsheets bite the dust. These workarounds are the consequence of earlier generations of computing, when the priority was to introduce automation to specific functions, without giving much thought to how data and processes would link up across the enterprise. Today, priorities have changed and that automation can no longer operate in isolation. This story at Extreme Networks is just one example of the new wave of investment that is now taking place to connect data and enable more joined-up processes across the enterprise.