Kumo is now a Snowflake Native App Learn more

Solutions

Personalization

With the same data integration and just a few predictive queries, ensure that every touchpoint with your customer is as delightful as possible and drives long-term engagement while improving product sales.

Personalization

Related case studies

DRIVING $100M IN GMV LIFT FOR ON-DEMAND DELIVERY

An F500 on-demand food delivery service increased GMV by $100M from better personalized store recommendations

Read case study

POWERING PERSONALIZATION IN GROCERY

A leading grocery chain improved personalized new item recommendations by 24% in 1 week

Read case study

IMPROVING MERCHANT RECOMMENDATIONS

An emerging markets on-demand delivery service improved merchant recommendations by 40% in 4 days

Read case study

Without any ML expertise required

  • Flexibly predict any measure of affinity between user and item:
    • Probability of click, purchase, return, or any other kind of interaction event
    • At the level of an individual item or any higher-level item attribute (category, brand, price band, etc.)
    • Over the timeframe of your choice–from days to quarters
  • Run a wide variety of analyses of your choice:
    • Affinity between any arbitrary user or item
    • Top K users likely to buy a target item or top K items to show a target user
    • Top K Items similar to a target item or top K users similar to a target user
    • …and more!

Downstream workflows turbocharged

  • Curate customer experiences across web/mobile app/etc. that look and feel unique to each user and address their individual needs
  • Anticipate growing trends in the marketplace to establish yourself as a trail-blazer among competitors
  • Leverage user preferences to advise new product designs and make your business more customer-focused
  • Programming chatbots that are capable of accommodating customer needs on an individual basis
  • Enhancing product catalogue or enterprise search systems
  • Enriching the content of any organic outreach (emails, notifications, sales calls, etc.) to drive higher customer engagement
  • Choosing which specific users to target with a new event, major upcoming product launch, thematic promotion, etc.
  • …and more!

Representative workflow

Representative Workflow

Define the ML Problem

Kumo makes it easy to define the ML task using a simple, declarative language, and then uses AutoML to execute a highly optimized and fully tuned model. Build the query and predict the future!

Structure your problem using a simple query to predict affinity to categories, brands, and products.

Predict 7-day category-specific purchase affinity:

PREDICT LIST_DISTINCT(Orders.Store_ID -> Stores.category, 0, 7)
FOR EACH Users.ID

Predict 7-day brand-specific purchase affinity:

PREDICT LIST_DISTINCT(Orders.Store_ID -> Stores.brand, 0, 7)
FOR EACH Users.ID

Predict user will return a product they purchased within 30 days:

PREDICT EXIST(Returns.itemID, 0, 30)
ASSUMING EXIST(Orders.itemID, 0, -30)
FOR EACH Users.ID

Query the future

Unleash the predictive power of your enterprise data

Get Free Trial