BBC Good Food
BBC Good Food is an institution for many an aspiring chef and even, we daresay, a few seasoned pros.
Anyone who has turned to the Internet in a desperate frenzy for a no-frills, dear-cooking-gods-please-spare-my-blushes dinner idea or even tinkered, however fleetingly, with the idea of pulling off the culinary equivalent of Leicester City winning the Premier League can attest to the above.
Despite having such a commanding online presence, tech moves on and more often than not, consumers follow. With this in mind, the BBC Good Food team were looking to elevate their position further still by understanding what their next play should be.
Our belief was that voice represented this next play.
A confluence of factors seemed to suggest that the time was ripe to make a significant move in the voice space before someone else did:
- Using voice is set to become a more commonplace behaviour - Google reported in 2016 that 20% of all searches were done by voice and ComScore has predicted that this is set to rise to 50% by 2020
- Penetration in UK households was increasing at a steady rate in 2017 as both Amazon and Google launched marketing assaults for their respective devices
- Cooking represents a hands-free interaction that a voice assistant naturally lends itself to
- Data shows that the kitchen is a primary location for smart speakers - approximately 50% of consumers house their devices in this part of their household
There existed a clear opportunity to extend BBC Good Food’s dominance into the voice space.
If we were to state the challenge very simply, it was to port BBC Good Food’s extensive recipe database onto voice.
This was far from simple though and with voice still being an emergent platform, we battled with a number of design challenges, most of which only became apparent through huddling around and testing each incremental iteration.
The devil really was in the detail.
Ultimately, the Skill was intended to be a utility to enable hands-free cooking. Therefore, we gave significant thought to the end to end cooking experience. What are people’s mindsets when it comes to cooking and therefore what use cases should we account for? What are some of the easily overlooked minutiae of how people actually cook and how could the Skill alleviate any pain points? What limitations of human cognition do we need to be cognisant of when it comes to listening to (as opposed to reading) a recipe?
The first major challenge was so big that we even gave it a name - ‘the snowball effect of questions and responses’. BBC Good Food has over 11,000 recipes and the Skill needs to funnel the user down to one recipe in the shortest number of steps possible.
Ask too many questions and we risked creating a frustrating user experience. Ask too few questions and the user would be faced with a barrage of recipes and only their voice to navigate this wilderness.
Another problem that quickly became apparent was that all of the content such as the length of the method steps and the phrasing of the ingredients were designed for web and mobile - i.e. they were ideal for reading.
Listening to Alexa rattle through a dense paragraph for one of the steps in preparing a Pan-fried chicken in mushroom sauce was a harrowing experience for all concerned.
To optimise the listening experience, we introduced relevant pauses whilst Alexa is reading out steps for a recipe. We also introduced pauses to reduce cognitive load and make locating/jotting down ingredients for a future shop easier - this is an insight that we leveraged from how people use BBC Good Food on web and mobile.
We also recognised that some steps in a recipe can be overwhelming. In these instances, users are able to ask Alexa to repeat steps and if the user ever gets interrupted mid-recipe with other goings on in the household, the Good Food Skill is able to remember where the user stopped and pick up exactly where they left off.
These were some of the many challenges we overcame but we have finally arrived at a Skill that we believe is perfectly adapted to the nuances of voices and will hopefully deliver a genuinely useful cooking experience in a way that only BBC Good Food can.
The BBC Good Food Skill has progressed from a prototype to an MVP in the space of 3 months and is set to launch in early August 2018, which is a remarkable achievement considering the complexity of the Skill and the design challenges that we faced. As designers and engineers, we are certainly better for the experience.
We were lucky enough to collaborate extensively with the BBC Good Food team and Amazon for this project. It was this collaboration, testing and such fast lines of communication and feedback that allowed for such a rapid turnaround from inception to delivery.
Once the Skill has been out in the wild for a few months, we will be updating this page with hard data on its performance.