How Did we Get Here & Where are we Going? Innovation in Warehousing

I’ll admit, I’m new to warehousing. Or at least, I speak with enough 20-year salty industry veterans every day to make me still feel like a rookie. That said, I’m not new to software. So when I saw “never upgrade again” as the big selling point that Manhattan Associates WMS started pushing last year, I was shocked.

My first reaction was: “Oh my God, did warehouses just discover that the cloud existed?” SaaS platforms automating the upgrade process has been a thing since Salesforce launched in 1999. To think 20 years later that the use of Cloud technologies to make upgrades seamless is just now becoming prevalent in any industry is astounding to me. When was the last time you had to upgrade Facebook, or Amazon, or Netflix? The answer is you don’t — it just happens without you caring, and that’s been the case for a long time for most modern software systems.

To put this in perspective, when SaaS automatic upgrades became the standard for software in other industries, the best commercial laptop was the Dell Latitude Notebook, offering a single CPU with 64 MB of RAM. Clearly, you’re reading this on a device that’s much more powerful than that, so you’ve successfully adopted technology faster than warehousing software groups (or vendors). So my question is this: how did we get here, and where are we trying to go?

In the famous words of Gartner Analyst Dwight Klappich, “warehousing is about 25 years behind manufacturing in adopting constraint-based planning techniques.” I’d argue that you can replace “manufacturing” in that statement with nearly every privatized industry and it would still hold true. So let’s explore what other industries have done and created a practical roadmap for how we can accelerate software technology adoption inside of warehousing. There are going to be terms in here you’ve seen a lot, but my hope is to ground them in reality so that you can identify exactly how to pull all this data into one place.

Cloud Computing
High-Level Concept to Know:
First comes the cloud: this gives you access to all of your data in real-time. Software that lives here doesn’t need to be updated. Compute is infinitely scalable. Your IT team has to deal with significantly less (but still some) headaches on a day-to-day basis. With good design, costs go down.

What You Really Need to Know:
Finding people that are experts in managing the hardware infrastructure needed to run our software world is really challenging. Luckily, our friends at Amazon, Microsoft, Google, and Oracle have made it really easy for us to migrate to the Cloud. Yes, you do need people who are experts in doing this, and you’ll start to hire folks with “DevOps” somewhere in their title. The good news is that they’re way easier to find than database administrators these days.

Descriptive Analytics
High-Level Concept to Know
Once you have your data in the cloud, you can start to understand why the things that happen, happen. You can put together dashboards that show you what is happening in real-time so that you can react accordingly.

What You Really Need to Know
There is a lot of software out there that helps you do this more easily. There is a 100% chance somebody on your team tells you that they can go and build a series of dashboards and applications that will address every one of your needs. This is likely true. However, as a company focused on running a complex supply chain, the last thing you want to do is maintain software releases over time. Buy from a specialized supply chain vendor like Longbow Advantage Rebus, or even use a dashboarding solution like Tableau or PowerBI before deciding to build (and maintain) it yourself.

Predictive Analytics
High-Level Concept to Know
Once you know why things are happening, you want to start to predict when they will happen again. Knowing when something will be a problem is the first step in avoiding it.

What You Really Need to Know
AI! Machine Learning! These are building blocks, not solutions. Use these terms with caution, and mentally interchange them with “math.” If it sounds stupid when replacing “machine learning” with “math,” it probably is. While these technologies can absolutely add value and more “predictivity” into your distribution center, understanding what you really want to accomplish is more critical. When thinking about the predictivity you want for your site, it’s helpful to phrase it this way: “If I knew that X was going to happen Y hours ahead of time, then I would be able to Z.” At the end of the day, it doesn’t matter what X and Y are if Z doesn’t result in time savings, dollar savings, or increased customer fill rates.

Prescriptive Analytics
High-Level Concept to Know
Prescriptive analytics is all about taking actions to traverse the optimal path to the best outcome in the future. If we know where we are and what’s going to happen, we should be able to use mathematics to influence decision-making. With prescriptive analytics, you can consistently review potential outcomes and take the appropriate actions to optimize results.

What You Really Need to Know
Most engineers don’t use the term prescriptive analytics because they’re not quite sure exactly what it means. Marketing teams use it because it sounds cool and it’s easy for decision-makers to understand. You get prescribed medicine, and it makes you feel better. You get prescribed analytics and magically your problems are solved. The reality is, in warehousing, prescriptive analytics can help to drive a few core changes that make distribution centers more efficient, merging highly fragmented decision processes (loading, picking, receiving, cuts management, etc.), increasing capacity per headcount, increasing fill rate, and optimizing work allocation. Even at a mid-sized distribution center, these few opportunities can add up to millions of dollars in capacity growth! That said, this optimized decision-making sits on the shoulders of giants and requires cloud-centric descriptive and predictive analytics to be successful.

What Have You Got, And What Do You Need?
If you consider the quest for prescriptive analytics a journey, there are a few software products that can help you drive towards a more efficient operation. Here’s a table outlining a reasonable trajectory:

Business Need Software Category Key Vendors
Collecting Data Warehouse Management System ● Blue Yonder
● Manhattan Associates
● Korber

Moving Data to the Cloud Data Replica & Data Lake ● Longbow Advantage Rebus
● Microsoft Azure
● Amazon Web Services
● Google Cloud Platform

Descriptive Analytics Dashboarding Platforms ● Longbow Advantage Rebus
● Tableau
● Microsoft PowerBI

Predictive Analytics Predictive Platforms ● Optricity
● AutoScheduler.AI
● Azure ML Studio
● Dataiku

Prescriptive Analytics WMS Accelerators ● AutoScheduler.AI

So, the bottom line is, there are a lot of critical software components required to move a distribution center into the 21st century. Those pieces require us to break the mold of a single monolithic software system doing everything and understand where new technologies can help us deliver more and drive efficiency. Keep your expectations realistic and consider that the movement to dark warehousing is the layering of numerous technologies. Start with small projects and ensure you have a team around you able to support and sustain them. Work with the right vendors, partners, and councilors to make sure you are integrating technology in a way that works for you. In the end, be constantly innovating—your competitors are.

By Keith Moore, Chief Product Officer, AutoScheduler.AI

Show More

Related Articles

Back to top button