Merchandising optimization: Predict, sense, and shape demand

Retail optimization challenges and opportunities

Inventory availability and promotions continue to influence customer decisions. As new customer journeys and shopping experiences emerge, execution is everything. Consumers won’t tolerate out-of-stock inventory—63% of consumers confirmed they will try another brand rather than wait for a restock. Retailers with more accurate forecasts and insight into their inventories will be able to quickly pivot to meet customer demand and shopping preferences, whether they’re shopping in store, buying online and picking up in store, opting for curbside pickup, or using other channels.

Consumers shop differently depending on what they’re buying. For example, they have different priorities when shopping for gifts than when shopping for seasonal items: Sometimes price is everything; sometimes it’s no object. Retailers can now understand how shoppers respond to every item at a localized level, enabling them to factor the effects of promotions, seasonality, and weather into their forecasts.

Additionally, retailers can increase personalization by targeting customers based on their real-time needs. Customers are more likely to accept offers that have been targeted to them based on an understanding of their behavior, how they shop, and the services they subscribe to, increasing offer acceptance rates and, ultimately, revenue. Conversely, pushing untargeted cross-sell or upsell offers can lead to customer fatigue and ignored or disabled notifications.

Simplify retail planning with advanced analytics and machine learning

Having the right inventory available in the right place to quickly fulfill customer demand, no matter when and where a purchase is made, will help retailers overcome the challenges and exploit the opportunities described above.

Retailers can use advanced analytics and machine learning to identify products and services that are relevant to a given customer. They can then push these products and services to the customer through the appropriate channel at the appropriate time.

Let’s take a closer look at how Oracle Data Platform can help retailers improve forecast accuracy, simplify planning, and optimize inventory.

This image shows how Oracle Data Platform for retail can be used to support merchandising optimization and help retailers predict, sense, and shape demand. The platform includes the following five pillars:

• Data Sources, Discovery
• Ingest, Transform
• Persist, Curate, Create
• Analyze, Learn, Predict
• Measure, Act

The Data Sources, Discovery pillar includes four categories of data.

Business record data comprises sales transactions, customer data, product data, returns transactions, suppliers, inventory, POS system data, revenue, and margin data.

Application data comes from ERP, SCM, CX, and WMS, Fusion SaaS, NetSuite, E-Business Suite, PeopleSoft, JD Edwards, SAP, Salesforce, and Workday applications.

Third-party data includes data from Oracle Data Cloud and social data.

Technical input data includes logs, web clicks, event streams, and beacons.

The Ingest, Transform pillar comprises four capabilities.

Batch ingestion uses OCI Data Integration, Oracle Data Integrator, and DB tools.

Bulk transfer uses OCI FastConnect, OCI Data Transfer, MFT, and OCI CLI.

Change data capture uses OCI GoldenGate.

Streaming ingest uses Kafka Connect.

All four capabilities connect unidirectionally into the serving data store and cloud storage within the Persist, Curate, Create pillar.

Additionally, streaming ingest is connected to stream processing within the Analyze, Learn, Predict pillar.

The Persist, Curate, Create pillar comprises five capabilities.

The serving data store uses Autonomous Data Warehouse and Exadata Cloud Service.

Managed Hadoop uses Oracle Big Data Service.

Cloud storage uses OCI Object Storage.

Batch processing uses OCI Data Flow.

Governance uses OCI Data Catalog.

These capabilities are connected within the pillar. Cloud storage is unidirectionally connected to the serving data store and managed Hadoop; it is also bidirectionally connected to batch processing.

Managed Hadoop is unidirectionally connected to the serving data store.

Two capabilities connect into the Analyze, Learn, Predict pillar: The serving data store connects to both the analytics and visualization capability and the data products, APIs capability, and the cloud storage capability connects to the machine learning capability.

The Analyze, Learn, Predict pillar comprises four capabilities.

Analytics and visualization uses Oracle Analytics Cloud, GraphStudio, and ISVs.

Data products, APIs uses OCI API Gateway and OCI Functions.

Machine learning uses OCI Data Science, Oracle ML, and Oracle ML Notebooks.

Streaming processing uses GoldenGate Stream Analytics and stream analytics from third parties.

The Measure, Act pillar comprises three consumers: people and partners, things, and applications.

The three central pillars—Ingest, Transform; Persist, Curate, Create; and Analyze, Learn, Predict—are supported by infrastructure, network, security, and IAM.

There are three main ways to inject data into an architecture to enable retailers to predict, sense, and shape demand.

• To start our process, we need to understand our overall inventory position. To do so, we use Oracle GoldenGate to enable change data capture ingestion of near real-time warehouse inventory data from operational databases for all or a subset of product lines.
• We can now add datasets relevant to retail demand, such as point-of-sale information. In addition to sales information, this data provides us with multiple insights, such as the time and location of the sale or information about refunds and swaps. These datasets often comprise large volumes of often on-premises data, and in most cases, batch ingestion is typically most efficient. For our point-of-sale data, we will use Oracle Data Integrator to ingest the data on a four-hour cycle.
• Streaming ingestion will be used to ingest data read from beacons at in-store locations through IoT, machine-to-machine (M2M) communications, and so on. Video imaging can also be consumed this way. Additionally, in this example, we intend to analyze and rapidly respond to consumer sentiment by analyzing social media messages, responses to first-party posts, and trending messages. Social media (application) messages/events will be ingested with the option to perform some basic transformation/aggregation before storing the data in cloud storage. Additional stream analytics can be used to identify correlating consumer events and behavior, and the identified patterns can be fed back (manually) for OCI Data Science to examine the raw data.

Data persistence and processing is built on three (optionally four) components.

• Ingested raw data is stored in cloud storage. We will use OCI Data Flow for the batch processing of this now persisted streamed data, such as tweets (JSON), location, sensor data from beacons and apps, geo-mapping data, and product reference data. These processed datasets are returned to cloud storage for onward persistence, curation, and analysis and ultimately for loading in optimized form to the serving data store. Alternatively, depending on architectural preference, this can be accomplished with Oracle Big Data Service as a managed Hadoop cluster.
• We have now created processed datasets that are ready to be persisted in optimized relational form for curation and query performance in the serving data store. This will enable us to identify and return the top trending product and consumer hashtags that can be enriched with location, inventory, and product data from enterprise systems.

The analysis part is built on two technologies.

• Analytics and visualization services deliver descriptive analytics (describes current trends with histograms and charts), predictive analytics (predicts future events, identifies trends, and determines the probability of uncertain outcomes), and prescriptive analytics (proposes suitable actions, leading to optimal decision-making), to answer questions such as
  • How do actual sales this period compare to the current plan?
  • What is the retail value of inventory on hand, and how does it compare to the same period last year?
  • What are the best-selling items in a division or department?
  • How effective was the last promotion?
• Alongside the use of advanced analytics, machine learning models are developed, trained, and deployed. These models can be accessed via APIs, deployed within the serving data store, or embedded as part of the OCI GoldenGate streaming analytics pipeline.
• Our curated, tested, and high-quality data and models can have governance rules and policies applied and exposed as a “data product” (API) within a data mesh architecture for distribution across the retail organization.

Increase profitability with a retail data platform

The forecast and corresponding pricing and promotional effects, paired with underlying costs and inventory availability, are the foundation for effective pricing and promotional decisions. Increase profitability and assortment flexibility with decreased inventory levels. Anticipate customer demand by maximizing the value of your data, applying analytics that draw from machine learning, artificial intelligence, and decision science disciplines to achieve the following:

• Increase revenue and expand a loyal customer base with higher in-stock rates.
• Increase brand loyalty by up to 77% with higher in-stock rates. Forecasts powered by automated intelligence can drive replenishment and allocation processes at a massive scale. The forecast and corresponding statistical prediction intervals enable supply chain success—having the right product in the right place at the right time—helping retailers plan forward-looking demand and statistical safety stocks while minimizing overall inventory costs.
• Convert up to 50% of shoppers into buyers with time-sensitive sales and promotions. Drive these decisions using forecasts based on contextual what-ifs and predictive analytics as well as optimized and forward-looking prescriptive analytics.