Introduction
Whisk can make food experiences more personal, enjoyable, and intuitive by learning about a user’s food preferences — everything from diets and allergies, favorite cuisines, preferred store brands, nutrition goals, and more.
Whisk’s personalization system builds a deep understanding of a user (their preferences, behavior, and context) by collecting a wide range of data from both explicit and implicit behavior.
Whisk can provide more relevant content (such as recipe recommendations) and make it quicker and easier to use by intuitively acting based on user preferences (such as which store items they want to be based on brand or budget).
How it works
Hard Constraints vs. Soft Influences
Whisk personalization applies two basic principles to the user experience:
Whisk uses a combination of data in three categories of hard constraints and soft influence: preferences, context, and user behavior.
Hard Constraints vs. Soft Influences
Hard Constraints - These are explicit exclusions. Things that don’t adhere to the selected preferences will not appear.
For example, if a user says they have a peanut allergy, Whisk won’t show peanuts to them in any recipe.
Soft Influences - These are loose rules that impact the priority and frequency of recommendations. These appear more often than non-preferences but will not totally exclude non-preferences.
For example, if a user selects Mexican cuisine as a preference Whisk will show more Mexican dishes but continue to show other cuisines.
Whisk uses a combination of data in three categories of hard constraints and soft influence: preferences, context, and user behavior.
1. Preferences
Preferences are data that the user has explicitly and intentionally entered into the system. Preferences can be broken down into three sections:
1.1 Constraints (Hard)
Name |
Description |
Options |
Diet
|
Exclude foods not adhering to a selected diet
|
Pescatarian |
Avoidances
|
Exclude foods containing selected avoidances
|
Wheat |
Dislikes
|
Exclude foods containing disliked ingredients
|
Any ingredient can be
|
1.2 Interests (Soft)
Name |
Description |
Options |
Cuisines
|
Boost foods that meet
|
African |
Health Goals
|
Recommend foods and recipes that help
|
Weight Loss |
Organic
|
Preferences for Organic ingredients
|
Organic |
Budget Preferences
|
Preferences for the budget of the user
|
Budget-Friendly |
1.3 Cooking Preferences (Soft)
Name |
Description |
Options |
Household Size
|
Number of people being cooked for, so
|
1 to 10
|
Cooking Experience
|
Cooking experience to recommend easier
|
Amateur |
Cooking Time
|
Preferences for cook time of recipes
|
15 min (or less) |
1.4 Cooking Preferences (Soft)
Name |
Description |
Options |
Preferred Retailers
|
Preferred places to shop and buy
|
US Acme
US Albertsons
US Food Lion
US GetNow
US Giant Food
US HEB
US Hy-Vee
US Instacart
US Jewel-Osco
US Peapod
US Price Rite
US Publix
US Safeway
US Shop Rite
US Stop and Shop
US Vons
US Walmart
GB Asda
GB Ocado
GB Tesco
GB Waitrose
DE Billa
DE Bringmeister
DE Rewe
AU Woolworths
ALL AmazonFresh
|
Location
|
Automatically collected to help
|
|
2. Context
Contexts are external factors and circumstances surrounding a user’s interactions with the system. These all affect personalizations and recommendations throughout the system.
Name |
Description |
Location
|
Regional variation in culture, cuisine, tastes, and availability of
|
Time
|
A specific time of day can impact what a user wants from the system.
|
Date
|
A user’s food choice can vary by time of year, especially in relation
|
Weather
|
Seasons and weather impacts the availability and preferences of food choices
|
Trends
|
Food can be a very social experience and therefore the tastes, interests,
|
Product Sales & Availability
|
The availability of certain seasonal products
|
3. Behaviour
Whisk collects large amounts of data from a user to algorithmically learn about tastes and preferences. Data is collected automatically from user interactions with the system, including:
Name |
Description |
View Recipe
|
User opens a recipe page
|
Save Recipe
|
User adds a recipe to their cookbook
|
Shop Recipe
|
User adds a recipe to their shopping list and sends it
|
Selects Store Item
|
User picks a store item in the online checkout
|
Shop Store Item
|
User purchases a selected store item
|
Best Practice
You can read more about how best practices in integrating personalization into your app in our separate article here.