June 2025. We launched voice payments.
Target user: 25-35 year olds. Tech-savvy. Bilingual. Hate switching keyboards.
We built onboarding tutorials showing young professionals using voice in coffee shops.
"Pay faster. Skip the typing."
Three weeks later, our analytics:
Age 18-34: 8% adoption
Age 55+: 63% adoption
We built for the wrong people.
The Assumptions We Made
Our user research said:
Who wants voice payments?
Why they'd use it:
We interviewed 80 people. Ages 22-38.
94% said they'd use voice payments.
We designed for them.
What We Built
Voice payments optimized for millennials:
1. Fast activation - Say "pay" and go
2. No tutorials - Assumed users know how voice works
3. Code-switching support - Mix Arabic and English freely
4. Slang recognition - "ابعت فلوس" ("send money") worked
5. Background use - Works while using other apps
Launched in a banking app. Riyadh. 40,000 users.
Week 1: 12% tried voice.
Week 4: 19% active users.
We thought we nailed it.
Then We Looked at the Data
Voice adoption by age:
| Age Group | % Who Tried | % Active After 30 Days |
|---|---|---|
| 18-24 | 18% | 6% |
| 25-34 | 14% | 8% |
| 35-44 | 21% | 19% |
| 45-54 | 34% | 38% |
| 55-64 | 58% | 63% |
| 65+ | 41% | 67% |
The older the user, the more they adopted voice.
This was the opposite of our hypothesis.
Why Older Users Loved It
We called 40 users over 50 who used voice daily.
Expected answers:
"It's faster." "I like trying new tech."
Actual answers:
*"I can't see the numbers on the keyboard. This is easier."* — User, 58
*"My fingers don't work as well anymore. Voice is more reliable."* — User, 61
*"I don't have to put on my reading glasses."* — User, 53
*"Text is so small. I always mistype. Voice works the first time."* — User, 66
They weren't using voice because it was faster.
They were using it because typing was harder.
The Real Pain Point
Vision Problems
68% of users over 50 mentioned vision as a factor.
Common issues:
One user: "I always mistype my PIN because the numbers are tiny. With voice I just say the amount and recipient. No mistakes."
Motor Control
Arthritis. Shaky hands. Slower finger movement.
22% mentioned physical difficulty typing.
One user: "My daughter set up the app for me. But I can send money myself now because I don't need to type."
Unfamiliarity with Dual Keyboards
Younger users switch between Arabic and English keyboards constantly.
Older users often stick to one keyboard. Switching is friction.
One user: "I type in Arabic. But amounts are in English numbers. Voice lets me say everything in Arabic. The app understands '500' whether I say 'five hundred' or 'خمسمائة'."
Trust in Confirmation Screens
This one surprised us.
Younger users type fast and submit. Older users double-check.
Voice + visual confirmation = perfect UX.
User says: "حول 200 لأحمد" ("Transfer 200 to Ahmed")
App shows: "Transfer 200 SAR to Ahmed - Savings Account. Confirm?"
User sees exactly what will happen. Taps confirm.
One user: "I trust it more because I can see what I said. When I type I worry I made a mistake."
Where Young Users Dropped Off
We interviewed 30 users ages 22-35 who tried voice once and quit.
Why they stopped:
*"I'm faster at typing."* (41%)
*"I use it in public. Can't speak out loud."* (38%)
*"I tried it but forgot it was there."* (29%)
*"I don't want people hearing my transactions."* (24%)
*"It's easier to just tap."* (18%)
Younger users:
Voice didn't solve a problem for them.
It solved a HUGE problem for users over 50.
The Pivot
We redesigned the feature. August 2025.
Old messaging: "Pay faster. Skip the typing."
New messaging: "Pay with your voice. Clear, simple, accurate."
Old onboarding: Young professional in a coffee shop
New onboarding: User at home, saying "Transfer 500 to Fatima" and seeing the confirmation screen
Key changes:
1. Larger confirmation screens - Show exactly what voice captured
2. Read-aloud confirmation - App speaks back: "Transfer 500 SAR to Fatima. Confirm?"
3. Tutorial for first-time users - Not everyone knows Siri/Alexa
4. Slower pacing - Don't rush to the next screen
5. Repeat option - Easy to say the command again if it wasn't right
8 Weeks After the Pivot
| Metric | Before | After |
|---|---|---|
| Age 55+ adoption | 63% | 71% |
| Age 55+ retention (30 days) | 67% | 79% |
| Age 18-34 adoption | 8% | 11% |
| Age 18-34 retention | 6% | 9% |
We optimized for older users. They loved it more.
Young users didn't care much either way.
The Daughter Effect
This was the unexpected growth driver.
We started seeing spikes in new voice users. All from the same source: family referrals.
Pattern:
One user testimonial:
*"My mom never used the banking app. I had to do her transfers for her. She said the buttons were too small and she always made mistakes. I set up voice. Now she does her own payments. She's so proud."*
We tracked referrals:
Before pivot: 4% of new voice users came from referrals
After pivot: 31% of new voice users came from referrals
Older users became our best marketers.
What We Got Wrong About Age
We assumed:
Young = Tech-savvy = Wants voice
Old = Tech-resistant = Won't use voice
Actual:
Young = Fast at typing = Voice is optional
Old = Typing is hard = Voice is essential
Voice isn't a "cool tech feature."
It's an accessibility feature that happens to be fast.
The Business Impact
Month 3 metrics:
| Metric | Result |
|---|---|
| Total voice users | 9,400 (23.5% of user base) |
| Age 55+ voice users | 6,298 (67% of voice users) |
| Daily voice transactions | 2,800 |
| Voice users who increased transaction frequency | 58% |
| Support tickets from voice users | -41% (vs non-voice users) |
Key insight: Users who adopted voice did MORE transactions.
Why? Because it removed the barrier.
Before: User avoids the app. Asks family to do transfers.
After: User does it themselves. Instantly.
One user: "I used to wait until I saw my son to send money. Now I do it right away. Sometimes 3 or 4 times a day."
The Lesson for Product Teams
Your intuition about your users is probably wrong.
We thought:
Reality:
If we'd launched with our original assumptions, we'd have hit 11% adoption.
By following the data and pivoting to our actual users, we hit 71% adoption in the segment that mattered.
What This Means for Voice Products
Stop designing voice for young professionals in coffee shops.
Start designing for:
1. Users with vision impairment - Large text isn't enough. Voice removes the problem.
2. Users with motor difficulties - Typing is physically hard. Voice works.
3. Users who want accuracy over speed - They don't care about saving 10 seconds. They care about not making mistakes.
4. Users who avoid apps because of UI complexity - Voice removes navigation friction.
Voice isn't a productivity tool.
It's an accessibility tool that makes products usable for people who struggle with traditional UIs.
The Final Numbers
What we projected:
What we got:
Revenue impact:
We built for millennials who wanted speed.
We found older users who needed access.
That's the difference between a feature and a product.