Embed Size (px)
Citation preview
Supercharging the Future of Retail withCommerce Cloud EinsteinRetail Connect | Melbourne
“Artificial intelligence will reach human levels by around 2029. Follow that out further to, say, 2045, we will have multiplied the intelligence, the human biological machine intelligence of our civilization a billion-fold.”Ray KurzweilAmerican Author, computer scientist, inventor and futurist
What is Einstein, and how does it work?Personalised recommendations based on the Shopper’s preferences and onsite behaviour
Product Recommendations for Digital
Leverage Commerce Data• Put the power of retailer’s data in their own hands
Personalise Across Channels
• Seamless shopper experience across mobile, desktop, and store touchpoints
Focus on Your Business
• Simplify merchandising for retailers- no data scientist required
Personalise recommendations across channels
Building Blocks of PersonalisationOne-to-All > One-to-Some > One-to-One
IndividualizationOne-to-One
SegmentationOne-to-Some
Dynamic Merchandising
Static Content
PersonalisationO
ne-to-All
Predictive Recommendations
Dynamic Customer GroupsSource Code Groups
DynamicSorting Rules
Commerce Cloud Einstein Data Sources
Product data• Learns about products, attributes, prices,
inventory
Order data• Learns about product relationships
(i.e. which products are bought together)• Learns about user affinity (i.e. who bought what)
Clickstream data• Learns about session behaviour
(i.e. who looked at what)
How Product Recommendations Work
Shopper comes to site and Commerce Cloud
Engine is called
Engine returns the product IDs
Storefront page displays best product
recommendations
Create & assign recommender
145637
876539
727457
554612
665390
Benefits of Commerce Cloud Einstein
Tracking & data learning already running (automatically activated after release 16.1)
• Recently Viewed Items
• User ID
Content slot integration• Scheduling
• Customer groups
• A/B Test
• Content vs. Products
• Campaigns
Flexible configuration of rules
Built into the platform
Type Home Page Footer
Any other page
(Account, Wishlist)
CategoryLanding
Page
Category Grid Page
Product Detail Page Cart Page
Recently Viewed Items
★ ★ ★ ★ ★ ★ ★
Based on all Categories ★ ★ ★ ★ ★ ★ ★
Based on current
Category★ ★
Based on current
Product(s)★ ★
Currently Supported Types and Locations
Types of Recommenders based on their Location
Type Description Anchor Expected
Typical Placement
Default Strategies
Product to Product Given a product or list of products, recommends similar/related affinity products
Product-id PDP • Customers who viewedalso viewed
• Product Affinity Algorithm
Products in A Category
Given a category, recommends products from within that category
Category-id Category Pages • Real-time personalised• Recent Top Sellers
Products in ALL Categories
Recommends products from across ALL categories
None Home PageAccount PageFooterCartMini-CartWish List
• Real-Time Personalised• Recent top sellers
Recently Viewed Shows products recently viewed by the shopper
None Any Page • Recently Viewed
Step by Step enablementWhat is required?
First Step – Data Enablement
Set up your data feeds• Product catalogue feed
• Order history (or legacy sites, store data)
• Clickstream data
The PI engines “digests” your data and uses machinelearning algorithms to process it:
• Collaborative filtering
• Unsupervised, semi-supervised, supervised learning
• Deep learning
The feeds have to be enabled by the Site Administratoron Production
More details in Commerce Cloud Einstein Help
Optimising Your RecommendationsElaborate a strategy and test, test, test!
Einstein AB Test Use CasesAlternate Product Recommendations on the PDP
Section Settings
Recommender Type Products to Product
Strategy Primary: Customers who viewed also viewedSecondary: Product Affinity Algorithm
Rule Any Product > DEMOTE > product_type = Match Anchor
Hypothesis Updated recommender will produce more revenue specific to recommendations and increase basket size of global experience.
Enabled Yes
Key Metric Average Units Per Order
Participation Trigger Pipeline Call: Pipeline: Product-Show
Control (50%) Existing slot configuration
Test Segment A (50%) New slot configuration containing new recommender with settings/configurations recommended above
Einstein AB Test Use CasesProduct Recommendations on the Basket Page
Section Settings
Recommender Type Products in ALL Categories
Strategy Primary: Real Time PersonalizedSecondary: Recent Top Selling
Hypothesis Including recommendations on the basket page increases AOV, but adversely affects Avg. Revenue per Visit.
Enabled Yes
Key Metric Avg. Revenue per Visit
Participation Trigger Pipeline Call: Pipeline: Cart-Show
Control (50%) No recommendation displayed
Test Segment A (50%) Einstein Slot – Products in ALL Categories
Hypothesis Including recommendations on the cart page increases AOV but adversely affects Avg. Revenue per Visit.
Commerce InsightsCorrelations You Had Not Thought Of
Discover the previously undiscoverable• Learn from your own Commerce data by
uncovering key product purchase correlations
Plan Store & Site Merchandising Smarter• Discern which products should be grouped
together for product bundles, deals and store merchandising
Truly understand your customers• Dig into purchase patterns to gain true awareness
Commerce Insights
The Commerce Insights Dashboard has various views:
• First view (previous slide), allows a retailers to choose a key item and see the items most commonly purchased with it.
• Second view (here), allows a retailer to click into that key items and discover additional insights (i.e. correlated products baskets and percentage rates)
Commerce Insights
Discover Product Sets You Had Not Thought Of
What are Shoppers buying together?
Use Einstein Ecommerce Insights to provide input on set combinations your merchandising team hasn’t thought of – that customers did!
Create Content to Support Seasonal Trends
Identify Seasonal Trends• Commerce Insights shows a high volume of
baskets with complementary winter camping products
Revisit and Refresh Existing Content• The ”Winter Camping Essentials” story has been
evergreened but obviously people are still purchasing items from it.
Feedback From Our Customers
“If you’re not using Commerce Cloud, you’re missing out on quite an opportunity.”Brian Hoven, Global Head of eCommerce, Icebreaker
Icebreaker Uses Einstein to Power Product Recommendations Outerwear and lifestyle clothing – 5,000 stores across 50 countries.
Web site powered by Commerce Cloud with product recommendations from Einstein.
40% more clicks, 11% higher average order value, 28% more revenue from recommended
products.
Predictive SortPromote the right product, first
Einstein Predictive Sort – Available now!
Create 1:1 Grid Pages• Personalise search and category pages for every
shopper, anonymous or logged in
Show the Best Products, First• Drive conversion by showing shoppers what they
want, especially in micro moments on mobile devices
Eliminate the Sorting Rule Guessing Game• Increase productivity with easy to use tools in
existing user interfaace
Infuse personalised product assortments into the shopper journey
How does Predictive Sort work?
With every click, Einstein collects the shopper’s browsing events and updates this shopper’s predictive model, in real-time, to calculate the most relevant products for each shopper.
Activities tracked:• viewCategory
• clickCategory
• viewProduct
The data is then used to re-order the results of site searches or grid pages.
Predictive Sort also available as dynamic attribute for your Sorting Rules.
Why You Should Use Predictive SortBenefits:
• Personalise search and category page for each shopper (know or unknown)
• Ensures your shoppers see the most relevant products to them, first
• Saves time by enabling sort personalisation within your existing business tools
• Increases revenue by leading your customers down a more direct path to purchase
• No data scientist needed!
• Eliminates time-consuming tasks of merchants determining the right sorting rules for various
customer groups and product categories
Einstein Predictive SortSteps to enable Predictive Sort on your PIG
Request Participation with your CSM
Data Enablement (if not already done)
Product Grid Template Change
Sorting Rule Configuration & Validation
Use Predictive Sort in your Storefront
“Predictive Sort eliminates the guessing. Being able to sort products, automatically per customer is huge.”Director ecommerce, CPO Commerce
Predictive Sort at CPO Commerce
America’s leading tool retailer known for offering customers high quality tools at great prices
Goal: Show each customers the best products for them
Predictive Sort ensures that anonymous and known shoppers see the best products in category and search resultsSimple implementation- “less than 5 minutes of work”
The Future of EinsteinProduct Roadmap
Forward-Looking Statements
Statement under the Private Securities Litigation Reform Act of 1995:
This presentation may contain forward-looking statements that involve risks, uncertainties, and assumptions. If any such uncertainties materialize or if any of the assumptions proves incorrect, the results of salesforce.com, inc. could differ materially from the results expressed or implied by the forward-looking statements we make. All statements other than statements of historical fact could be deemed forward-looking, including any projections of product or service availability, subscriber growth, earnings, revenues, or other financial items and any statements regarding strategies or plans of management for future operations, statements of belief, any statements concerning new, planned, or upgraded services or technology developments and customer contracts or use of our services.
The risks and uncertainties referred to above include – but are not limited to – risks associated with developing and delivering new functionality for our service, new products and services, our new business model, our past operating losses, possible fluctuations in our operating results and rate of growth, interruptions or delays in our Web hosting, breach of our security measures, the outcome of any litigation, risks associated with completed and any possible mergers and acquisitions, the immature market in which we operate, our relatively limited operating history, our ability to expand, retain, and motivate our employees and manage our growth, new releases of our service and successful customer deployment, our limited history reselling non-salesforce.com products, and utilization and selling to larger enterprise customers. Further information on potential factors that could affect the financial results of salesforce.com, inc. is included in our annual report on Form 10-K for the most recent fiscal year and in our quarterly report on Form 10-Q for the most recent fiscal quarter. These documents and others containing important disclosures are available on the SEC Filings section of the Investor Information section of our Web site.
Any unreleased services or features referenced in this or other presentations, press releases or public statements are not currently available and may not be delivered on time or at all. Customers who purchase our services should make the purchase decisions based upon features that are currently available. Salesforce.com, inc. assumes no obligation and does not intend to update these forward-looking statements.
Einstein Search Dictionaries (GA FEB 2018)
Discover Search Gaps Automatically• Uncover gaps between your search settings and
the way customers are searching for products
Seamless and Easy to Use• Fully integrated feature allows you to improve
search results with a few clicks
Never miss a search term again
Einstein Search Suggestions (BETA Q1 2018)
Show the right product, First• Autocomplete search, tailored to the individual
shopper
Promote search discovery• Power recommended, related, popular, and
recent searches
Anticipate shopper search intent before she/he types
