How Gap leverages data and AI for retail success
Did you miss a session at the Data Summit? Watch On-Demand Here.
As data becomes the new oil, organizations across sectors are racing to leverage their (often siloed) information management systems and drive value across functions with viable AI applications. The need has surged the demand for competent data scientists and analytics professionals, but even with the proper talent, bringing data-driven products to life can be quite a task. According to two Gartner reports, just 53% of AI and ML projects make it to production and 85% of the lot typically fail to deliver on their intended results.
Despite the challenges, including data quality issues, Gap, which is one of the biggest clothing and accessories retailers in the world, has come out with the successful implementation of data and AI to solve key business problems.
“We have truly accelerated our data journey. In fact, one of the first priorities I was given as I came into this role is how can we actually accelerate our data journey and make AI a part of the DNA for how we do our work within Gap,” Heather Mickman, the interim CIO of the company, said during a panel at Venturebeat’s Data Summit.
The Gap data journey
Initially, Gap was leveraging traditional BI (business intelligence) capabilities and operational reports to take day-to-day decisions. Then, Mickman said, the company started investing in data science and analytics teams, which worked with the IT teams to experiment with data and come out with potential business use cases.
“We worked through slow cycles to create bespoke datasets to enable different hypotheses for different use cases. And, after that, we moved to our next evolution what we have today – the Gap data platform,” she said. The solution is driving the shift from manual data curation and experimentation to self-service (consumption-ready datasets curated in a catalog), allowing teams to perform AI/ML on-demand and truly embed it into data decisioning.
In addition, the CIO emphasized that the company has also exited most of its data centers and shifted to the cloud to leverage cloud-native capabilities while focusing on data and data science projects.
Cleaning data at source remains a top priority
As GAP’s data platform begins to open up within the organization, the company is keeping a close eye on ensuring that the information flowing through the pipeline remains clean and accurate.
“There are many nuances in terms of how you make that happen, but that needs to be intentional across each of the datasets,” Mickman said. “I don’t want to see our teams cleaning data midstream. It’s something that needs to happen at the source. Otherwise, you’re going to have different data going across the organization, creating a lot of confusion.”
She also stressed the importance of creating governance roles such as data owners, data stewards and who, within teams, has the ultimate accountability for datasets. Within Gap, these roles have been set up within product management teams as they have more of that business context than technology teams.
Applications to solve business problems
With this data platform and governance strategy, Gap has built some powerful downstream applications to solve key business challenges and drive efficiencies. The biggest one of the lot is an AI engine that forecasts demand in terms of what products are going to sell, where they’re going to sell, and when they will sell.
“It uses predictive analytics and processes various datasets, including sales data and product features, to influence buying, positioning, pricing for inventory across all of [the] Gap brands,” Mickman said while noting that the engine has significant potential to improve product availability and product margins.
Beyond this, the company has also developed AI-driven solutions to optimize inventory movement between Gap fulfillment centers and stores and help brands with size profiling.
“The Inventory optimization model goes to a fairly granular level…to understand dynamics of the sales and margins and what the potential is to ensure that we’re positioning (inventory) in a very smart way to satisfy our customer demand,” Mickman said.
Meanwhile, the size profile solution leverages ML to enable various brands owned by Gap to produce automated and more accurate size profiles that enhance their product coverage and customer satisfaction.
“We use different attributes that you might imagine like sales and inventory data to determine the size selling for a specific item in a store,” she said.
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More
Source: Read Full Article