Maximizing Supply Chain Agility through the “Last  Mile” Commitment

In my last two blogs (Get to Know Your Retail Customer: Accelerating Customer Insight and Relevance, and Improving your Customer-Centric Merchandising with Location-based in-Store Merchandising) we looked at the benefits to retail in building personalized interactions by accessing both structured and unstructured data from website clicks, email and SMS opens, in-store point sale systems and past purchased behaviors. Once establishing deeper knowledge of the customer, we then discussed building ‘segments of one” through localized assortments, tailored promotions, dynamic pricing, and product development that is reflective of the true needs of the customer. Now that we have established tailored demand, we need to figure out how to fulfill demand though a robust and capable supply chain, let’s drive into building an agile retail supply chain.

In today’s retail environment, retailers realize that building demand forecasts simply based upon historical transaction, promo, and pricing data alone is not good enough. Data today has a shelf life much like produce and needs to be updated in real-time to be relevant. Data 100 days old might as well be 100 years old. Integrating new data sources is now required to improve in-stocks and over-stocks.

Both data-in-motion and data-at-rest are leveraged to drive retail’s supply chain agility. Streaming or real-time data from on-vehicle sensors, shelf, or point of sale are leveraged along with historical archives of consumer purchase behavior or inventory stock levels.  Including new data sources like demand signals (e.g. weather, social commentary, competitor pricing, local event calendars, shipping and returns policies, and demand transfer dynamics) not only improves forecast accuracy, but greatly enhances inventory visibility, reduces out-of-stocks, and improves today’s customer fulfillment expectations. 

To enable these business capabilities requires an enterprise data platform to process streaming data at high volume and high scale, to manage and monitor diverse edge applications, and provide data scientists with tools to build, test, refine and deploy predictive machine learning models.  

Today’s supply chain now includes direct-to-consumer, micro-fulfillment centers, and in-store pickup options. Encompassing internal product flow data (which is controlled), but also influencers (that are semi-controlled) provide new challenges, but also more insight into business capabilities delivered through an enterprise data platform approach.  Business capabilities that can be realized by leveraging enterprise and customer-centric data streams:

  • Inventory Visibility—Actively update sales and inventory to enable near real-time inventory availability across channels
  • Flexible Fulfillment—Provide enhanced convenience and customer service with buy anywhere, return anywhere and fulfill from anywhere capabilities
  • Website Operations—Analyze website operations to improve efficiencies in order fulfillment service levels, optimize delivery options offered
  • Consolidated Inventory & Sales Data— Build an enterprise view of sales and inventory across all channels

Retail supply chains are a recognized and proven source of ROI when data analytics are leveraged to improve forecast accuracy and product availability. Gartner, IDC, and ISM have reported that incorporating big data helps improve demand forecasts, building supply chain agility that can provide a 2% average revenue increase and a 15% average inventory reduction simultaneously.

Cloudera worked with a leading European supermarket data science team to develop an opportunity to reduce ‘last mile’ delivery costs while trying to keep the customer fulfillment promise at all costs. Existing route optimization tools were not reflective of the true drive times for deliveries and pre-built demand forecast models (intra-day orders) were leading to further inefficiencies. Additionally, fleet maintenance costs were excessive due to unplanned downtime and the need for additional vehicles needed to anticipate the unplanned vehicle downtime.

The retailer leveraged Cloudera to build an analytics solution for fulfillment delivery that allowed for advanced analytic modeling, A/B testing, and optimization by improved data access of omnichannel orders, logistics, and delivery capacity. Working with the incumbent point solution provider, this retailer brought the analytic modeling and IP in-house improving delivery order demand forecast accuracy and route optimization. Through this collaborative effort, they also reduced point solution costs, improved analytic agility, and established an approach to use with other legacy business applications. This retailer also implemented a follow-on use case of predictive analytics for maintenance, improving uptime of their delivery fleet.

The following illustrative, measurable business impacts were realized:

  • Improved intra-day online order fulfillment demand forecast accuracy by 3%
  • Improved customer delivery capacity, service with shortened delivery windows
  • Reduced number of vehicles / drivers by 140 (@ $150k cost per) = $21m 

Stay tuned to my next and last blog of this series in the retail space that will consider all possibilities when thinking how to reimagine stores.  Additional retail content can be found at our retail resource kit  

David LeGrand
More by this author

Leave a comment

Your email address will not be published. Links are not permitted in comments.