Walmart’s knowledge graph strategy for connected learning networks at KMWorld 2022

A single source of truth is an aspiration sought by enterprises burdened by ineffective and inefficient content management, only aggravated by the growth of hybrid and remote working environments. Shifting from a disjointed, ad-hoc approach to a connected learning experience can make the difference between a scalable, flexible knowledge foundation and one with perpetual shortcomings.

At KMWorld 2022, Todd Fahlberg, a portfolio manager at Enterprise Knowledge, led a discussion on the crucialness of managing structured and unstructured content throughout a multitude of complex systems. Using Walmart’s solution as the leading example, Fahlberg addressed common challenges and roadblocks in enterprise KM as an enterprises’ success demands scale and generates highly interrelated data in his session, “Building an Innovative Learning Ecosystem at Scale with Graph Technologies.”

The picture of Walmart’s operations is expansive. With the Walmart Academy supporting the core functions of 2.4 million associates at Walmart, Inc.—including more than 3.5M assignments, more than 4k video assets, and more than 4.3k eLearning projects—managing and organizing that knowledge is a wildly large task.

The reality is Walmart associates generate more than 80k SharePoint site visits every 90 days; this sort of large scale data leads to challenges in visibility, search optimization, and incorporating sufficient content data models. Simultaneously, you don’t want to “boil the ocean,” or try to reinvent the wheel of knowledge organization ad-hoc.

“For Walmart, their L&D experts often can’t find content and must recreate new content, or are not clear on the data, or don’t see its value,” said Fahlberg. “Part of our efforts is showing them that it can be findable through collaboration.”

To formulate the building blocks for a collaborative solution, Walmart focused on what it was exactly that they wanted to change, and what the outcome was that they desired. It narrowed down to a “library” approach to organizing information: a card catalog for drilling down interests and a detailed index that helps find information applied through semantic search and auto-tagged, systematically-related metadata.

“Ultimately, we want to make each individual content as unique as possible, so you can empower users to find what they need at the time that they need,” remarked Fahlberg.

This led Walmart to knowledge graph technology as the ideal foundation for organizing its knowledge. They needed tech that met their current knowledge needs and could continue to scale, could manage interrelated data and organize siloed data, as well as remain flexible with new integrations, data models, and content as the company evolves.

So, why a graph database over relational? Its data is represented by entities and their relationships, and can easily express complex relationships between multiple entities. It’s entirely purpose-built to manage relationships and power flexibility, meeting Walmart’s vast data needs.

Underpinned by a knowledge graph database, Walmart developed their connected learning experience solution: a centralized digital library, or learning content database (LCD), acting as a single source of truth for Walmart’s learning portfolio that is widely accessible and streamlined.

“We wanted to keep it simple,” said Fahlberg. “We wanted direct access to content and made a simple search query.”

The solution features quick access to a content details page, connections to “source systems” for quick view, and related learning within search results. Apache Solr, the source enterprise search platform built on Apache Lucene, allows them to navigate auto-tagged data through semantic search, powered by LCD’s knowledge graph and taxonomies.

With Walmart’s foundational graph technology supporting its large knowledge network paired with their connected learning solution, the enterprise is able to visualize its interdependent data while keeping knowledge accessible, reliable, and simple.

KMWorld returned in-person to the J.W. Marriott in Washington D.C. on November 7-10, with pre-conference workshops held on November 7.

KMWorld 2022 is a part of a unique program of five co-located conferences, which also includes Enterprise Search & Discovery, Office 365 Symposium, Taxonomy Boot Camp, and Text Analytics Forum.