Improving inventory management

The sudden acceleration of high-intensity e-commerce demand, alongside fracturing supply chains in the wake of COVID-19, has pushed inventory management to the forefront. Reuters Events: Supply Chain asked two innovators how they see the space evolving

Image by falco from Pixabay

It’s no secret that supply chains have struggled in recent months. Time and again companies have found themselves without key components or in-demand products. Factory closures, blank sailings, massive cuts to air freight, reduced warehouse capacity and more have hit business-as-usual and often turned carefully calibrated chains on their head.

What’s more plenty on the retail side have found themselves with held inventory but it suddenly locked away in stores that people could no longer frequent, browse and shop from.

This has pushed inventory management far up the agenda for a host of companies seeking to properly get a handle on what is where in their logistics train and where problems the next blockage might occur.

We spoke to two industry figures looking to help out with these issues in Sanjay Sharma, CEO of on demand goods and asset monitoring as a service company Roambee, and David Glick, CTO of on-demand warehouse company FLEXE, to get their perspective on the evolution of inventory management.

A gap in knowledge and from distribution network to consumer

“What's happening is there is a big disconnect,” says Sharma. That disconnect stretches across supply chains due to a lack of overarching visibility caused by the long tails of those supply chains and the manual nature of keeping track of them.

Sharma gives the example of automotive parts manufacturer “Who is a one billion, two billion-dollar company making all these spare parts and I have absolutely no visibility on the repair shops and what inventory they are carrying. So, the only way I'm operating today is I'm doing checkpoints, which means I'm rolling my truck, repair shop by repair shop, every week, every two weeks, every month in the given zip code carrying my inventory and asking them, how much inventory did you sell and how much is on the shelf? That's extremely ineffective.”

This becomes even more severe at a moment of dislocation or stress on a company or supply chain.

I believe that you're going to see sustained push to e-commerce

This is particularly so in those selling to consumers currently, as social distancing measures have totally up-ended the standard models and shifted demand online, just as the pandemic has diminished many supply chains’ capacity to respond.

“I believe that you're going to see sustained push to e-commerce,” says Glick. The sudden acceleration in market share of e-commerce “might go down” as this crisis subsides, “but still, you jumped ahead five years plus in the way we do the supply chain.”

This has led “Some retailers [to say] we're going to bet on our stores, [which have] inventory close to customers, and that's the most important thing…. You see that from Walmart and Target, but the problem is they've only got 100,000 SKUs or 150,000 SKUs in a store, whereas Amazon has 5 million SKUs right down the road, and so I think they need something more than just their stores,” believes Glick, especially for a “for a one day process in the US.”

The new normal is basically to rethink and redesign the supply chain network

There, “You're talking about … 20 to 25 warehouses,” across the country to give that same day coverage. The alternative is “enabling your stores in some fashion to do warehousing and then you get thousands of them. I think people had been putting this off putting this off and putting this off, and now it's staring them in the face and they're all going to try to figure this out over the next year.” However, this will require a new approach to inventory management from the ground-up to be able to track and operate across these geographic and product ranges cost-effectively.

There is a lot of desire for enterprises to become extremely granular in their visibility data collection

Sharma agrees that “The new normal is basically to rethink and redesign the supply chain network. What used to happen in the pre-COVID days, enterprises had what I call it visibility in instalments. What I mean by that is, you know, some companies would have visibility on the last mile, some will have visibility on orders, some would have visibility at their origins, some would have visibility in the form of fleet tracking.”

However, now “There is a lot of desire for enterprises to become extremely granular in their visibility data collection. And why that is, is because the supply chain has become more unpredictable and skewed in many ways.

“So, there are multiple trends that are happening in the industry and one is they're looking to see how we can redesign the network and redesigning the network means can I find new routes? Can I be effective in looking at new origins and destinations? Can I basically bring on my supplier community into a network of visibility very quickly? Can I deliver the visibility I have to my customers so that I can own trust and customer satisfaction? There are these multiple drivers that are pushing the company to embrace visibility. Now, the biggest challenges is how do you get started?”

Starting smart

Sharma believes that the starting point has to be on the areas of biggest risk to your business, as adding in tracking will help reduce the risk of a business-critical line suddenly going south and taking profit with it.

He advises taking “A sensor-based approach,” on critical supply chain lanes or with key customers. If you are delivering to a Best Buy or to a Walmart but “I am losing their trust because every time I give them a schedule of delivery, I'm not able to keep up on the information I provide,” then he advises that “every shipment that goes to my customer Walmart, I'm going to put a sensor on it."

Another approach would be to try and identify “problem SKUs. And what I mean by that is an example would be my 60-inch TV is getting damaged more often than other SKUs. So, I want to inject sensors in monitoring or all 60-inch SKUs that are going out of my distribution centre.”

I would contend that you're better off spending your time doing trying to figure out how to take $1 out of delivery

Glick also advises focusing on those problematic cost centres first, which in his view is “In the transportation, especially when you go to a one-day or same-day promise. You can take a nickel or a dime or a quarter out of pick, right?” says Glick but that means investing in automation and “trading upfront capital costs for reduced variable costs over the long term. I would contend that you're better off spending your time doing trying to figure out how to take $1 out of delivery.”

In his opinion the way you do that is “By putting inventory close to customers. The way you do that is by having a multi-node network” and “by putting the nodes in the right place”  so “you can do next day for $6 if you have inventory close to the customer, as opposed to it's going to cost three, four times that if you put it on plane and fly it.” This approach, however, will mean a smart approach to understanding demand patterns and inventory placement across supply chains.

Forecasting in the eye of the storm

It is no easy feat to do this, especially in the topsy-turvy environment of 2020.

“We used to have a forecasting system at Amazon that looked at seasonality and sort of traditional volume,” says Glick, “and then if there was a one-time promotion or something,” they would note not to “count this in the future forecast, because it was an exogenous event. So, we were joking that 2020 is one big causal. It's going to screw up everybody's inventory ordering and forecasting for the foreseeable future,” to the extent that “we're seeing these people who have automated systems to buy are now moving back to humans,” notes Glick.

Absence of … long term forecasting, you need to be agile

Glick recommends that in the “Absence of … long term forecasting, you need to be agile,” which means putting inventory in the right locations, distributing that risk through multiple stock locations and the ability to “change your ordering habits as quickly as possible. That means shortening the supply chain, if at all possible.

“If you are agile and able to react quickly, and you've built those processes, that resilience, that anti-fragility … it allows you to be sloppier on your forecast,” says Glick.

Sharma also advises focusing on risk management in these times rather than just trying to rely on a refined forecasting approach. Installing sensors and gaining more granular visibility within the supply chain means you can “Build risk profiles. Risk profiles, not only by suppliers, but the regions in which these suppliers operate. You will be able to build risk profiles around lanes. You will be able to build risk profiles around your own distribution centres and stores.” This means a more targeted approach to solving supply chain blockages and a greater ability to react before the problem is insurmountable.

However, it all starts with a network approach says Sharma. “Imagine now you've collected a year worth of data on the shipments that you're monitoring, whether it's raw material or finished goods. You can then step back and start looking at patterns and giving risk scores to your points within the network, and based on the risk scores when a pandemic like this happens, you can go back tothose risk profiles, identify the heights or hotspots that are going to have faster breakdowns than others and retune or rejig your supply chain much more quickly than what we saw in this pandemic.”

Inventory management, evolved

Critically, the barriers are coming down when it comes to stepping up your inventory management game. The cost of adding capabilities is falling as the agility of software and the shrinking price point of hardware are making it faster and cheaper to have oversight over a greater proportion of the supply chain and the inventory within it.

FLEXE by the nature of their business as an on-demand, felixble warehousing provider have focused on standing up customers quickly and getting a hold on their inventory levels and placement.  

With Bluetooth and cellular technology, the game has changed

“A customer comes to us and says, ‘Hey, we want to be able to do two day delivery to 93% of the customers with X amount of inventory, 99% of the time,” explains Glick. “We will ask them a bunch of questions and make sure we understand the scope of what they're doing. Then we will go out and source warehouses for them…. Then from a setup … standpoint, it's a few configurations in our system. It's all cloud based. So, the warehouse provider doesn't have any software installed locally. They just log on to our portal to see the shipments that need to be done and they do that at their PC. We give them scanners so that their associates have the apps to go pick and pack those products. Then the customer will send us their item master data, their product data, and then they start cutting POs to send to our warehouses. So, while the actual standing up of the warehouse takes a day, that whole process from, from the time we engage with a customer to the time where we're shipping products to them can be less than a month,” indicating how quickly inventory management can be done now through cheap, deployable systems interfacing with the cloud.

Similarly, Sharma believes that “With Bluetooth and cellular technology, the game has changed.”

In the next three to five years, there will be a much higher adoption of automated track and trace solutions

Not only is this in the form of reducing hardware costs for placing trackers onto shipments but in the data management and analysis. Sharma feels that “The new trend is going to be how I can have a one-to-many relationship, meaning one sensor to many shipments? Some shipments will have sensors, some shipments won't have some sensors,” but the big change is being “able to infer the condition or the KPIs on the shipments that don't have the sensor,” from those that do “is going to be basically the new direction where artificial intelligence and machine learning will play a big role.”

He believes that this force multiplier for inventory tracking will mean that “In the next three to five years, there will be a much higher adoption of automated track and trace solutions,” that would expand to becoming a door-to-door visibility platform.”

In his opinion “It's a no brainer – most companies will embrace that and possibly have hundred percent coverage in the next two, to three, to five years. Then I see that the slower part would be how do you take the automated track-and-trace service or solution and translate that into an early warning system. Companies who are at the leading edge who are fighting to retune their KPIs and looking at sensor logistics data are reading risk profile much better and can put a better early warning system [in place] than others.”

comments powered by Disqus