Local search for a screen the size of an orange
"Alexa, show me nearby coffee shops." Local search sounds simple until the device answering is an Echo Spot: a small, round screen, usually seen from across the room, backed by a voice interface. I designed local search for the Spot as an adaptive UI — one that reshapes the answer around what the user actually asked, instead of showing the same business page every time.
Problem
Optimize for the Echo Spot form factor.
A phone search result can be dense; a Spot cannot. The design had to be glanceable from a distance, concise and scannable, and it had to end in an action — call them, get directions, book a table, get a ride, or send the details to your phone. Same goals, very different needs depending on the question.
How might we…
Same goals, different needs
How might we…Prioritize info for different use cases to match user needs
Glanceable from a distance
How might we…Play with the density and hierarchy to make it more scannable
Integrate follow-on actions
How might we…Develop clear patterns for primary and secondary actions
Adaptive UI
The key move: the response is assembled from the intent. Every query type gets a primary piece of information, a secondary piece, and the most relevant action:
- General ("tell me about Cafe Solstice") — open/closed status, category, book a table.
- Where is it? — location and distance, get directions.
- Phone number? — the number, open/closed, call them.
- Hours? — open/closed and hours, book a table.
One place, four answers. The same Cafe Solstice, reshaped by the question that was asked — the screen, and the spoken line that goes with it:
You“Alexa, where is Cafe Solstice?”
Alexa“I found a Cafe Solstice nearby. It’s half a mile away on E Thomas St in Seattle. It’s open now until midnight.”
You“Alexa, what’s the phone number of Cafe Solstice?”
Alexa“The phone number of Cafe Solstice, located nearby on Thomas St, is (206) 403-1916.”
You“Alexa, when is Cafe Solstice open till?”
Alexa“Cafe Solstice is open now until midnight.”
You“Alexa, tell me more about Cafe Solstice.”
Alexa“Here’s info about Cafe Solstice.”
The pattern held for any business — the same hours question, asked about Starbucks:
You“Alexa, when is Starbucks open till?”
Alexa“The Starbucks on Maury Ave is open now until 10pm.”
Two page types
A listing page answers category searches ("coffee shops near me") with a tight, scrollable set of results. A detail page answers questions about one place — and it isn't fixed: condensed and scroll variants trade density against the maximum lines of text the little screen can show, and social proof (a review highlight — "Happy Hour, Acme Bread, Kumamoto") can lead when it's the most persuasive thing on the page.
Confidence, handled out loud
Voice results aren't always certain, so the wording carries the confidence. High confidence: "Cafe Solstice is half a mile away on E Thomas St in Seattle. It's open now until midnight." Medium confidence: "I found a Cafe Solstice nearby…" — same facts, honestly hedged, with follow-on actions offered either way.
Outcome
Local search launched on Echo Spot as one of the two features I headlined in my Amazon portfolio — an early proof that a voice answer and a screen answer should be designed as one thing, each doing what it's best at.