Peter Odum

My most recent recent responsibility at Weber was designing the Producer software application for Mac.

This application serves as the primary front end UI for the back end repository containing all smart/connected recipe content for Weber.

Producer is also used to manage all culinary content appearing in the newly released Weber Connect mobile app, and will soon be used for other web-based content as well.

Based on the smart recipe data programmed by culinary staff, a recipe is rendered flexibly into any supported medium, such as the Weber Connect App for iOS and Android. 

Modern smart cooking redefines how people cook their food. It borrows liberally from ideas used in test kitchens and in molecular gastronomy by scientist chefs. In smart kitchen appliances, it uses sensors, AI, multimedia-rich instruction and other technologies to ensure optimal cooking results. 

Weber, the world’s leading barbecue grill brand, and June, famous for pioneering smart cooking technology, together introduced the concept of smart grilling.

I had the primary task at Weber of remaking the back end application and UX/UI enabling culinary authors to create the smart recipes driving Weber and June’s Smart Cooking/Smart Grilling products. This software is called Producer.

Producer enables culinary staff at Weber internationally (and June in the Bay Area) to create and publish smart recipes for the various supported platforms. An important goal was for these recipes to be re-usable and adaptable to many different appliances, across various fuel types (gas, charcoal, pellet, electric) and other variables.

The Producer software was created for Mac and iPad. Screenshots here are from the Mac version, which was the lead platform.

The hardware used with Producer recipes for cooking and grilling has a great deal of automation built in:

The June Oven
A sophisticated piece of tech and an electric kitchen-in-a-box which uses a top-down camera and computer vision to recognize the foods placed in the oven. Advanced Machine Learning algorithms are used to determine number and placement of items, decide how those items will best be cooked, and predict doneness levels and finish times for users. 

Weber Smart Grilling devices
Weber Smart Grilling devices also include a good bit of automation. Built similarly to the June oven but (for now) without cameras for food recognition, these devices are available in standalone configurations (the Weber Connect Smart Grilling Hub) as well as built into various grills (Weber Smokefire, the Genesis and Spirit lines of gas grills). 

A major innovation for Producer was the ability to support ‘Dimensions’ - the variables that can affect how a recipe will be executed on a smart cooking device. And one of the most complex early recipes for June Oven was bacon.

Bacon has multiple variables or ‘dimensions’ that drastically affect how the end product will taste:

- Foil on pan Y/N?
- Thick or thin cut Bacon?
- User’s texture preference: chewy or crispy?
- Number of strips to be cooked, from 1 up to 8

The combinatorics just from these dimensions result in a total of 96 different instruction sets, one for each possible value of each dimension.

For grilling, we needed to support dimension sets like those for bacon, but add dimensions for the various appliances that would do the grilling, since Weber supports many different grills and fuel types.

In the image at left, the platforms of several grill products (identified by internal names) are added to the mix for a New York Strip Steak. 

The benefit of all this complexity is that one recipe record can be used to house information for all the ways Weber appliances can cook a substance. This makes content creation and editing easier, and simplifies the process of adding new appliance models over time. It also means a recipe can have one unified set of string translations, easing localization. 

It also turns out Australia uses a different size for their tablespoons from the US, UK and the rest of the world. So our system needed to support variability in measurements in order to accurately reflect the intention of the recipe author.

Other variations exist from one country to another, from different measurement standards to different ingredients being available, or the same word being used for a physically different ingredient. 

Overall, there was a distinct need to be able to publish the same recipe in different countries at different times, or not to publish certain recipes in certain places. Producer allows the author to publish a recipe for very specific given markets, without publishing it elsewhere.

In creating this version of Producer software, there was a lengthy process of requirements gathering, planning, designing, revising, presenting solutions and then providing test criteria to ensure those solutions were workable for the final products.

I interviewed Culinary stakeholders to identify their needs and gain understanding of what goals must be served.

Then I wrote specification documents detailing what I had learned from Culinary staff (double-checking with them to ensure I had accurately understood their needs).

Finally I double-checked these specs with program managers and developers to understand technical solutions, reflecting this information back to culinary staff so they knew what to expect.

I created and presented user flows demonstrating how recipe content might appear in the final products, to ensure we knew how the back-end recipe work would affect the front end applications.

I created wireframes and visual designs for the internal tool, taking advantage of native elements and structures that could easily be rendered within our SwiftUI framework, coordinating with developers both before and during the design process, adapting to developer requests to provide the best and most efficiently deliverable result.

I wrote Gherkin-formatted test cases to ensure that resulting designs accurately met original design goals, and could be efficiently tested by QA staff.

I also trained and worked with culinary staff as they adopted the new tool, observing their struggles and adjusting the design where possible to make the tool easier to use.

In addition I wrote documentation to help others learn to use the tool as it evolved, providing tips and best practices to aid culinary authors in smoothly creating new recipe content.

When the new Weber Connect software launched in April of 2023, it was entirely dependent on recipe structures imported to and built natively within this back end tool.

The launch of the software was smooth and stable, and it has garnered glowing reviews (average 4.5 stars in the Apple store, 3.7 stars on Google Play where reviews of the old app still factor in).

Producer’s interface and recipes enabled this simultaneous international launch, and allowed us to substantially increase the number of recipes available in the app vs. what had been in the previous Weber Connect app.

I and my team succeeded in taking in a very complex set of requirements, importing a large quantity of existing data, localizing it for 18 different languages, structuring it successfully for use in the new app, adding additional data and systematization, and delivering the resulting recipe database and management tool to support this groundbreaking software release within about 18 months time, all while maintaining live recipe services for previous products.