Beer Me!
Personalized beer recommendations on the go
Beer has become bonkers!
I love beer, and with the explosion of craft breweries in the last 10 years came increased flavor complexity and a mind-boggling amount of options to choose from... Personally, I think we’re living in the golden age of craft brew! That being said, unless you’re a beer expert you might find yourself slightly overwhelmed or confused about which specific beer to order, let alone what flavors you might like.
Wouldn’t it be great if you could get personalized beer recommendations anywhere, even if you didn’t know anything about beer?
I targeted two groups I felt could benefit the most from personalized recommendations.
After interviews supported the idea, I found it easier to talk about the groups as personas.
Josh (a beer newbie)
“I don’t know where to start when deciding on a beer since I don’t know much about styles or flavors.”
Amy (drinks one type of beer)
“I feel limited in choice because I don’t want to risk a bad beer taste experience.”
Josh & Amy's shared concerns
How could I reduce complexity of choice and ordering anxiety for Josh & Amy?
First, I needed to gain insight about the psychology of choice.
Notes from Sheena Iyengar's research in The Art of Choosing & Barry Schwartz’s, More Isn’t Always Better.
Excessive choice can cause choice paralysis and reduce people’s satisfaction with their decisions, even if they made good ones.
Sometimes it’s better to have others choose for you, but only if you’re properly informed.
Some choice is better than none, and even the illusion of choice makes us happier.
While addressing Josh & Amy’s concern #1 (overwhelmed with choice) the insights helped inform design solutions including:
Limiting beer recommendations to the top 3 (initial view)
Explaining why specific beers are recommended
Providing features like menu scan, where a user does not even have to read an intimidating menu
Addressing Josh & Amy’s concern #2 (helplessness from lack of knowledge) was less clear.
It seemed to me that offering opportunities to gain knowledge is only part of the answer when decision making is involved.
Providing enough easy-to-understand information to make a quick, informed choice may be even more powerful than knowledge in this context.
Addressing Josh & Amy’s concern #3 (risk of a bad beer taste experience) came down to personalization & trust.
Personalized beer recommendations based on flavors a user already likes lowers risk when ordering. Trust in these recommendations comes from accuracy over time.
So how would personalized flavor recommendations be possible?
We need to know how a user’s taste buds work beyond beer flavors. Generating a non-beer flavor profile is especially important for Josh (the beer newbie) who has a low level of experience with beer.
Involving subject matter experts is critical (e.g. flavor/taste scientists, beer sommeliers, etc.). With their input we could create a definitive series of questions to be used for a science-based onboarding quiz.
Since question generation is a project in itself I have shown question placeholder within the onboarding flow.
For Amy (who drinks one type of beer) there is an opportunity at the end of the onboarding quiz to rate past beers or styles which would further inform her flavor profile.
After onboarding Josh & Amy’s unique flavor profile sets the foundation for all their recommendations.
There are two necessary facets:
1. General flavor preferences (food & drink)
2. Beer flavor zone (defining what recommended beers should taste like).
Since Josh hasn’t rated many beers at this point his general flavor preferences have a heavy weight in the recommendation criteria.
Amy has a more defined beer flavor zone since she has rated more beer. When ratings are paired with the general flavor preferences a more complete flavor profile can inform recommendations.
Once Josh & Amy have a flavor profile, how would they access recommendations?
Location is a critical factor in the beer world: every bar, store, and online presence has a different beer selection. The primary recommendation feature needed to account for where Josh & Amy are grabbing beer.
Version 1 of 3
Location is a critical factor in the beer world: every bar, store, and online presence has a different beer selection. The primary recommendation feature needed to account for where Josh & Amy are grabbing beer.
Version 2 of 3
Since the majority of online beer purchases made through apps like Drizzly or Instacart are still made through local stores it made sense to combine online and physical stores into one path.
Version 3 of 3
The final iteration used a search input field to further streamline the flow. Menu Scan became another prominent feature, providing a secondary path to recommendations.
Josh’s path (location input) was the path utilized most in user testing. Amy’s "scan menu" path was secondary but still necessary.
Josh’s idealized path to recommendations
While at Reuben’s Brewery, Josh uses the search feature to access Reuben’s-specific recommendations, examines one of those recommendations in more detail, then rates the beer he just drank.
Amy’s menu scan
During testing, users thought this sequence seemed a little too “magical” to be believe and were unnerved about the abruptness or lack of control during the transition (from taking a photo directly to top 3 recommendations).
Revised Sequence
By adding multiple cues as to what the system is doing and giving the user a choice before continuing, a friendlier and more appropriate user experience was created.
Engaging users, in the short and long term
After more discussions with potential users I realized that what I had in my first iteration was more of a tool than a rich experience.
Users wanted to be led through the initial process and given multiple ways of interacting with an app. “Give me a compelling reason to keep coming back.”
I addressed the issue by exposing many of the previously hidden features on the home screen.
1. A brief highlight of a user’s flavor profile, adding personal connection.
2. Beer style recommendations, encouraging discovery/learning.
3. Nearby recommendations for destination ideas, creating a potential starting point for a user without a destination in mind.
4. Misc. personalized picks including recommendations based on activity, food pairing, weather/season, etc. to add intrigue with constantly changing content.
Initial home screen
Final home screen
What’s up with that crazy looking bear?
His name is Bruno!
A beer recommendation app was a perfect opportunity for a mascot to set a lighthearted tone, further the brand identity, show empathy during a user’s journey, and relate to the subject matter with a pun.
Fun, funky, & hip UI
Interviews and surveys revealed that users (especially users that influenced the Josh persona) are drawn to novelty. The informal voice and fun/funky/hip UI fit the subject matter and meshed well with Josh & Amy’s expectations.
