Checking inventory by hand is becoming a thing of the past. The broad adoption of mobile apps is making managing inventory much easier for organizations that rely on warehouses. However, apps can come with a catch. Warehouses are often in remote locations — which can be great for storage space but bad for network connectivity.
Grainger is the largest maintenance, repair, and operations (MRO) distributor in North America. It stocks 30 million products across 35 global distribution centers. With MongoDB Atlas Device Sync and machine learning, Grainger isn’t letting poor connectivity get in the way of great inventory management. Its inventory management app makes life easier for distribution specialists whether they have Wi-Fi or not. That’s thanks to Grainger’s decision to prioritize the employee experience and to use MongoDB to tap into AI and machine learning.
When it comes to app development, Grainger is a “big fish that can move quickly,” said staff software engineer Toheeb Okenla. It’s a large enterprise that has broken free of the pitfalls that slow down development for companies at this scale. Okenla is part of the AI and Vision Group at Grainger. Not only is the company innovating quickly at the edge, he said, but this is often where opportunities to explore AI lie.
“MongoDB is a great platform for us. It can handle large quantities of distributed data efficiently and seamlessly integrate the network edge with our back-end systems,” he explains.
Grainger keeps comprehensive inventories of all the equipment it stocks for customers, often items that are crucial to run their businesses, like safety equipment, power tools, and motors. To that end, every site has on-site specialists managing inventory, and they need to know exactly where every item is, down to the rack and shelf number. Previously, Grainger specialists tracked inventory using stickers and spreadsheets, but that wasn’t the experience the company wanted to give its workers.
“We wanted to build a great inventory management app, but developing at the edge in areas of low connectivity can be a real challenge,” says Okenla. “Costs can escalate because you end up trying to work around it – either with conflict resolution systems or by creating a separate offline experience for your apps.”
These offline experiences often limit user functionality and can cause data syncing issues. This stems from a server-centric architecture that relies on devices to manually connect and upload data that has been modified. When devices are offline for extended periods of time and then finally reconnect, the server struggles to keep data consistent with other changes that have taken place over time.
With Grainger’s old applications, users also had to ensure that they were offline before starting work, and later to manually reconnect once they were in an area with sufficient connectivity. These additional steps led to a poor user experience that was neither sustainable nor scalable.
As a “big fish,” Grainger has the budget to spend on software development. But large enterprises like Grainger can face competition from smaller, more agile organizations, who might not be as limited by legacy technology, lengthy approval processes, and strict procedures to follow. To ensure speed and agility, Grainger needed to give its developers the right platform to quickly design intelligent, next-generation apps.
“MongoDB Atlas removes the barriers that slow down innovation,” says Okenla. “Atlas Device Sync comes with out-of-the-box functionality to get around the complexities of offline development, while bringing advanced machine learning capabilities to the table.”
The team deployed machine learning models at the edge that automatically identify the schema of the item distribution staff are collecting. The app creates a digital twin of the schema in MongoDB Atlas, and Atlas Device Sync automatically updates the database when connectivity returns.
With bidirectional syncing between devices and the server, staff can take high-quality images and capture metadata without being connected to the internet. Put another way, instead of a server being the source of truth, every device is its own source of truth.
MongoDB provides eventual consistency so if many devices are offline at the same time, it can handle concurrent updates when they reconnect. All devices, therefore, convert to an identical state and integrate with downstream systems without a cumbersome development process in the background.
And because MongoDB Atlas is so flexible, the team could use the database schema best suited to devices.
“That might seem like a small thing, but when the backend isn’t optimized for mobile devices, you get latency issues or a bad user experience,” notes Okenla. “To be successful, apps need to help your people to be successful. Machine learning can remove a lot of repetitive, low-value tasks and free your people up to shine.”
Grainger is empowering its distribution specialists to be more productive and deliver value from anywhere. That’s just one example of how MongoDB Atlas is helping the team to develop faster, better applications and stay well ahead of the competition.