Menu
Back
B2B
Mobile App
User Testing
Team Collab
Designing a smarter verification flow for informal African businesses
59
out of 64 successfully onboarded
92%
of testers loved the onboardingflow
83%
submitted docs on the first try
Role
Lead Product Designer
Duration
4 weeks
Tools
Figma | Figjam | Maze
Team
Product Designer - 1 | Compliance - 3 | Product Manager - 1 | Head of Product
The beginning of it all
Twolly is an African marketplace platform helping informal local businesses sell smarter, reach further, and get paid faster across borders.
Working with diverse small African businesses, Twolly helps them sell and deliver their merchandise to customers across Ghana, Nigeria and Kenya. This primarily means ensuring each business on the platform goes through the legitimate compliance process to verify themselves as well as their businesses to ensure we’re able to support the right businesses.
What problem needs fixing? Why??
It currently takes 7 days for a business to get verified on Twolly before they are able to start selling on the platform. This is due to the manual processes the Compliance Team goes through to verify the documents businesses provide.
Current implications
- Interest in businesseses to join the platform drops due to how long verification takes
- Business is asked to provide some documents based on their location (i.e Ghana / Nigeria / Kenya)
- Documents are submitted to the Compliance Team for due diligence
- Compliance team manually verifies documents based on factors like country, nature of business, legitimacy of documents etc
- Approval/Denial based on findings
The goals at the end of the day
- Handle different business risk levels (low, medium, high) with the right level of friction
- Stay compliant with local regulations in Nigeria, Ghana, and Kenya while minimizing verification fraud
- Increase the number of businesses that complete the verification process
- An easy to understand and intuitive digital verification process to serve informal, low-tech business users even on patchy connections and low-end devices
The final solution
You’ll see these again as you go along
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
before getting my hands dirty
The team-agreed assumption
We’re betting that if we can:
- Adapt verification flows based on business risk profile
- Speak to users in a way they actually understand
- Guide them step-by-step without assuming tech literacy
- And smooth out country-level edge cases ...
…then businesses will get verified within 24 hours with little Compliance Team interference.
How might we...
How might we craft a seamless digital business verification experience optimized to fully verify businesses at the first try while behaving smartly according to the risk level detected without scaring them off or slowing them down while providing a seamless experience?
My biggest constraint - Time
With just 4 weeks to design & 4 weeks to implement the verification flow, there wasn’t enough runway for deep competitor analysis, so I prioritized speed without compromising context by leaning on:
- Insights from the compliance team to understand their current processes and regulatory must-haves
- Existing research already done across Nigeria, Ghana, and Kenya
- Prioritizing prototype testing to quickly capture early reactions and gut sentiments from real users
research & collaboration
Who is user going to experience this?
I worked with Compliance to better understand who our sellers across Nigeria, Ghana, and Kenya.
What I did
- Created a Seller Risk Persona Matrix mapping sellers into Low, Medium, High risk tiers
- Captured traits, behaviors, documentation levels, and how each group should be onboarded
The early beginnings of trying to map out the Persona Matrix in Figjam
Why it mattered
- Helped us design tailored onboarding flows (e.g. soft nudges for informal sellers, strict gating for high-risk actors)
- Replaced blanket rules with conditional logic based on seller profile
- Improved trust, reduced drop-off, and sped up onboarding for legit businesses
Snapshots of first 3 steps of the verification flow
1
This is one major step to segment users into risk tiers. Based on what they sell (e.g. peer-to-peer loans = high risk), how they run the business (solo vs with a partner), and what country they’re from, my matrix kicks in to determine what flow they see next.
2
High-risk users are given a stricter review process without being punished for it
I softened the experience with:
- A “Save & continue later” escape point
- A “Need help?” section tailored for those who get stuck
- A callback prep screen to guide them before talking to support
3
Instead of showing the same requirements to everyone, I tailor document requests (e.g. asking for 1 document from low-risk businesses.
This shift from blanket rules to conditional requirements is a direct output of the persona matrix.
Crafting a story for this user(s) - Personas
I translated each risk group into goal-driven, behavior-rooted personas
What I did
- For Low-risk, the matrix showed readiness and legitimacy → I designed a persona that valued speed, trust, and minimal disruption
- For Medium-risk, the confusion and partial legitimacy stood out → the persona reflected a motivated but unsure seller who needs guidance, not friction
- For High-risk, the matrix flagged red zones → the persona became a cautionary archetype to design guardrails around
Why it mattered
- It bridged the gap between compliance risk logic and UX empathy
- It gave teams a way to design per risk, not despite it
- It helped cross-functional teams speak the same language about who we were optimizing for
Snapshot of the Persona of the owner of a medium-risk business
What does the journey of this user look like?
The next logical step was to map the actual journey each seller would face when joining the platform
What I did
- Low-Risk Sellers saw a clean, fast track — their journey reflected confidence, low friction, and automation opportunities
- Medium-Risk Sellers had uncertainty and doc confusion — their map surfaced education gaps and needed reassurance touch-points
- High-Risk Sellers showed mistrust, and rejections and their path reflected empathy, guidance, and escalation lanes
Why it mattered
- It gave product, compliance, and ops teams clear targets for intervention
- It directly fed into how we designed onboarding flows, doc validation logic, and support pathways
- It became a shared blueprint that connected business risk with product experience
Rough user journey of a Low-Risk Business
Rough user journey of a Medium-Risk Business
Rough user journey of a High-Risk Business
experience solutions
The final solution
...again...like I said
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
Looking beyond the happy path - Edge Cases
- Auto-image quality check e.g. blur detection using frame colors as visual cues
- Prevents preventable delays from low-effort mistakes.
- Reinforces quality expectations without friction.
- Managing long waits during verification step using engaging content
- Designing for lost internet connection
- Added a recovery screen with clear messaging and retry option
- Built in progress preservation so users don’t lose work mid-verification
- Used empathetic copy to reduce stress and build trust
- Designed with low-bandwidth realities in mind, especially in Nigeria, Ghana, and Kenya
testing & iteration
Testing as early as possible - Prototype Usability
To observe how real business owners from across Africa experience the onboarding flow especially across risk levels and uncover usability issues, unclear copy, and emotional friction
Participants
- 2 Nigerian sellers (1 formal, 1 informal)
- 2 Ghanaian sellers (both informal)
- 1 Kenyan seller (formal but solo operator)
Outcome Highlights
“I didn’t realize the icon meant I had to take a selfie”
Ghanaian seller
Users misunderstood the abstract icon, leading to confusion about the selfie step and delays in continuing.
“The camera just snapped by itself and I didn’t know if it worked or not.”
Kenyan seller
Auto-capture felt foreign and unreliable to users who were used to tapping the shutter themselves.
“There was so much writing and I still didn’t know what documents to upload.”
Nigerian seller
Overwhelming text and no visual guidance made sellers second-guess what was required for verification.
Problems Identified
Users didn’t recognize the selfie icon as an actionable step. It failed to communicate purpose, causing hesitation and drop-off at the selfie capture step.
With all the information present, sellers couldn’t confidently identify which documents to upload
The automatic selfie capture broke familiar mental models. Instead of feeling seamless, it made users anxious about whether the photo was taken correctly or taken at all.
Iterating/Improvements after testing
1
Added sample selfie with verification badge. Sets a visual standard and subtly reassures users (“this is what success looks like”)
2
Added extra text to further guide users on exactly what to do to be successful in this step
1
Rewrote the guidance beneath the frame to be shorter and more direct; telling users exactly what to do, in the tone they expect from a native camera app.
2
Added “Why we need this” link
Introducing this link gave users helpful context, reducing skepticism and making the request feel more legitimate especially for high-risk users who are more cautious about privacy
3
Switched from automatic to manual capture.
Users were more comfortable taking control of the camera. Manual capture aligned with how they already use their phone cameras, reducing confusion and building trust in the process
the aftermath
Milestones
- 59 out of 64 businesses successfully verified and approved
- 92% of testers rated the onboarding flow as “clear” or “very clear”
- Onboarded all 3 seller types: low (38%), medium (44%), high (18%)
- 83% of users submitted valid docs on the first try
Next steps
- Continue monitoring product progress in beta
- Implement analytics on Mixpanel with PM and devs to monitor user behavior
- Prepare support team with real user edge cases (e.g. poor internet, ID rejections) so they’re ready to guide sellers smoothly during onboarding
What I learnt from this project
- Design for emotional reassurance, not just functional completion
- Treat edge cases as core cases; they reveal where real people struggle
- Ask more than “can they do it?” and instead ask “how does it feel while doing it?”
- Make the invisible visible; feedback, progress, and support need to be loud and clear
Contact Me
I dont bite...
© 2025 Frankie’s Portfolio
About Frankie
Design Gallery
Thoughts to Words
Email me
Back
B2B
Mobile App
User Testing
Team Collab
Designing a smarter verification flow for informal African businesses
59
out of 64 successfully onboarded
92%
of testers loved the onboardingflow
83%
submitted docs on the first try
Role
Lead Product Designer
Duration
4 weeks
Tools
Figma | Figjam | Maze
Team
Product Designer - 1 | Compliance - 3 | Product Manager - 1 | Head of Product
The beginning of it all
Twolly is an African marketplace platform helping informal local businesses sell smarter, reach further, and get paid faster across borders.
Working with diverse small African businesses, Twolly helps them sell and deliver their merchandise to customers across Ghana, Nigeria and Kenya. This primarily means ensuring each business on the platform goes through the legitimate compliance process to verify themselves as well as their businesses to ensure we’re able to support the right businesses.
What problem needs fixing? Why??
It currently takes 7 days for a business to get verified on Twolly before they are able to start selling on the platform. This is due to the manual processes the Compliance Team goes through to verify the documents businesses provide.
Current implications
- Interest in businesseses to join the platform drops due to how long verification takes
- Business is asked to provide some documents based on their location (i.e Ghana / Nigeria / Kenya)
- Documents are submitted to the Compliance Team for due diligence
- Compliance team manually verifies documents based on factors like country, nature of business, legitimacy of documents etc
- Approval/Denial based on findings
The goals at the end of the day
- Handle different business risk levels (low, medium, high) with the right level of friction
- Stay compliant with local regulations in Nigeria, Ghana, and Kenya while minimizing verification fraud
- Increase the number of businesses that complete the verification process
- An easy to understand and intuitive digital verification process to serve informal, low-tech business users even on patchy connections and low-end devices
The final solution
You’ll see these again as you go along
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
before getting my hands dirty
The team-agreed assumption
We’re betting that if we can:
- Adapt verification flows based on business risk profile
- Speak to users in a way they actually understand
- Guide them step-by-step without assuming tech literacy
- And smooth out country-level edge cases ...
…then businesses will get verified within 24 hours with little Compliance Team interference.
How might we...
How might we craft a seamless digital business verification experience optimized to fully verify businesses at the first try while behaving smartly according to the risk level detected without scaring them off or slowing them down while providing a seamless experience?
My biggest constraint - Time
With just 4 weeks to design & 4 weeks to implement the verification flow, there wasn’t enough runway for deep competitor analysis, so I prioritized speed without compromising context by leaning on:
- Insights from the compliance team to understand their current processes and regulatory must-haves
- Existing research already done across Nigeria, Ghana, and Kenya
- Prioritizing prototype testing to quickly capture early reactions and gut sentiments from real users
research & collaboration
Who is user going to experience this?
I worked with Compliance to better understand who our sellers across Nigeria, Ghana, and Kenya.
What I did
- Created a Seller Risk Persona Matrix mapping sellers into Low, Medium, High risk tiers
- Captured traits, behaviors, documentation levels, and how each group should be onboarded
The early beginnings of trying to map out the Persona Matrix in Figjam
Why it mattered
- Helped us design tailored onboarding flows (e.g. soft nudges for informal sellers, strict gating for high-risk actors)
- Replaced blanket rules with conditional logic based on seller profile
- Improved trust, reduced drop-off, and sped up onboarding for legit businesses
Snapshots of first 3 steps of the verification flow
1
This is one major step to segment users into risk tiers. Based on what they sell (e.g. peer-to-peer loans = high risk), how they run the business (solo vs with a partner), and what country they’re from, my matrix kicks in to determine what flow they see next.
2
High-risk users are given a stricter review process without being punished for it
I softened the experience with:
- A “Save & continue later” escape point
- A “Need help?” section tailored for those who get stuck
- A callback prep screen to guide them before talking to support
3
Instead of showing the same requirements to everyone, I tailor document requests (e.g. asking for 1 document from low-risk businesses.
This shift from blanket rules to conditional requirements is a direct output of the persona matrix.
Crafting a story for this user(s) - Personas
I translated each risk group into goal-driven, behavior-rooted personas
What I did
- For Low-risk, the matrix showed readiness and legitimacy → I designed a persona that valued speed, trust, and minimal disruption
- For Medium-risk, the confusion and partial legitimacy stood out → the persona reflected a motivated but unsure seller who needs guidance, not friction
- For High-risk, the matrix flagged red zones → the persona became a cautionary archetype to design guardrails around
Why it mattered
- It bridged the gap between compliance risk logic and UX empathy
- It gave teams a way to design per risk, not despite it
- It helped cross-functional teams speak the same language about who we were optimizing for
Snapshot of the Persona of the owner of a medium-risk business
What does the journey of this user look like?
The next logical step was to map the actual journey each seller would face when joining the platform
What I did
- Low-Risk Sellers saw a clean, fast track — their journey reflected confidence, low friction, and automation opportunities
- Medium-Risk Sellers had uncertainty and doc confusion — their map surfaced education gaps and needed reassurance touch-points
- High-Risk Sellers showed mistrust, and rejections and their path reflected empathy, guidance, and escalation lanes
Why it mattered
- It gave product, compliance, and ops teams clear targets for intervention
- It directly fed into how we designed onboarding flows, doc validation logic, and support pathways
- It became a shared blueprint that connected business risk with product experience
Rough user journey of a Low-Risk Business
Rough user journey of a Medium-Risk Business
Rough user journey of a High-Risk Business
experience solutions
The final solution
...again...like I said
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
Looking beyond the happy path - Edge Cases
- Auto-image quality check e.g. blur detection using frame colors as visual cues
- Prevents preventable delays from low-effort mistakes.
- Reinforces quality expectations without friction.
- Managing long waits during verification step using engaging content
- Designing for lost internet connection
- Added a recovery screen with clear messaging and retry option
- Built in progress preservation so users don’t lose work mid-verification
- Used empathetic copy to reduce stress and build trust
- Designed with low-bandwidth realities in mind, especially in Nigeria, Ghana, and Kenya
testing & iteration
Testing as early as possible - Prototype Usability
To observe how real business owners from across Africa experience the onboarding flow especially across risk levels and uncover usability issues, unclear copy, and emotional friction
Participants
- 2 Nigerian sellers (1 formal, 1 informal)
- 2 Ghanaian sellers (both informal)
- 1 Kenyan seller (formal but solo operator)
Outcome Highlights
“I didn’t realize the icon meant I had to take a selfie”
Ghanaian seller
Users misunderstood the abstract icon, leading to confusion about the selfie step and delays in continuing.
“The camera just snapped by itself and I didn’t know if it worked or not.”
Kenyan seller
Auto-capture felt foreign and unreliable to users who were used to tapping the shutter themselves.
“There was so much writing and I still didn’t know what documents to upload.”
Nigerian seller
Overwhelming text and no visual guidance made sellers second-guess what was required for verification.
Problems Identified
Users didn’t recognize the selfie icon as an actionable step. It failed to communicate purpose, causing hesitation and drop-off at the selfie capture step.
With all the information present, sellers couldn’t confidently identify which documents to upload
The automatic selfie capture broke familiar mental models. Instead of feeling seamless, it made users anxious about whether the photo was taken correctly or taken at all.
Iterating/Improvements after testing
1
Added sample selfie with verification badge. Sets a visual standard and subtly reassures users (“this is what success looks like”)
2
Added extra text to further guide users on exactly what to do to be successful in this step
1
Rewrote the guidance beneath the frame to be shorter and more direct; telling users exactly what to do, in the tone they expect from a native camera app.
2
Added “Why we need this” link
Introducing this link gave users helpful context, reducing skepticism and making the request feel more legitimate especially for high-risk users who are more cautious about privacy
3
Switched from automatic to manual capture.
Users were more comfortable taking control of the camera. Manual capture aligned with how they already use their phone cameras, reducing confusion and building trust in the process
the aftermath
Milestones
- 59 out of 64 businesses successfully verified and approved
- 92% of testers rated the onboarding flow as “clear” or “very clear”
- Onboarded all 3 seller types: low (38%), medium (44%), high (18%)
- 83% of users submitted valid docs on the first try
Next steps
- Continue monitoring product progress in beta
- Implement analytics on Mixpanel with PM and devs to monitor user behavior
- Prepare support team with real user edge cases (e.g. poor internet, ID rejections) so they’re ready to guide sellers smoothly during onboarding
What I learnt from this project
- Design for emotional reassurance, not just functional completion
- Treat edge cases as core cases; they reveal where real people struggle
- Ask more than “can they do it?” and instead ask “how does it feel while doing it?”
- Make the invisible visible; feedback, progress, and support need to be loud and clear
Contact Me
I dont bite...
© 2025 Frankie’s Portfolio
About Frankie
Design Gallery
Thoughts to Words
Email me
Back
B2B
Mobile App
User Testing
Team Collab
Designing a smarter verification flow for informal African businesses
59
out of 64 successfully onboarded
92%
of testers loved the onboardingflow
83%
submitted docs on the first try
Role
Lead Product Designer
Duration
4 weeks
Tools
Figma | Figjam | Maze
Team
Product Designer - 1 | Compliance - 3 | Product Manager - 1 | Head of Product
The beginning of it all
Twolly is an African marketplace platform helping informal local businesses sell smarter, reach further, and get paid faster across borders.
Working with diverse small African businesses, Twolly helps them sell and deliver their merchandise to customers across Ghana, Nigeria and Kenya. This primarily means ensuring each business on the platform goes through the legitimate compliance process to verify themselves as well as their businesses to ensure we’re able to support the right businesses.
What problem needs fixing? Why??
It currently takes 7 days for a business to get verified on Twolly before they are able to start selling on the platform. This is due to the manual processes the Compliance Team goes through to verify the documents businesses provide.
Current implications
- Interest in businesses to join the platform drops as a result of how long verification takes
- Twolly gets only 2 out of every 5 businesses to complete the verification process
- The Compliance team usually has several back and forths with businesses till the right documents are provided
The goals at the end of the day
- Increase the number of businesses that complete the verification process
- An easy to understand and intuitive digital verification process to serve informal, low-tech business users even on patchy connections and low-end devices
- Handle different business risk levels (low, medium, high) with the right level of friction
- Stay compliant with local regulations in Nigeria, Ghana, and Kenya while minimizing verification fraud
The final solution
You’ll see these again as you go along
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
before getting my hands dirty
The team-agreed assumption
We’re betting that if we can:
- Adapt verification flows based on business risk profile
- Speak to users in a way they actually understand
- Guide them step-by-step without assuming tech literacy
- And smooth out country-level edge cases ...
…then businesses will get verified within 24 hours with little Compliance Team interference.
How might we...
How might we craft a seamless digital business verification experience optimized to fully verify businesses at the first try while behaving smartly according to the risk level detected without scaring them off or slowing them down while providing a seamless experience?
My biggest constraint - Time
With just 4 weeks to design & 4 weeks to implement the verification flow, there wasn’t enough runway for deep competitor analysis, so I prioritized speed without compromising context by leaning on:
- Insights from the compliance team to understand their current processes and regulatory must-haves
- Existing research already done across Nigeria, Ghana, and Kenya
- Prioritizing prototype testing to quickly capture early reactions and gut sentiments from real users
research & collaboration
Who is user going to experience this?
I worked with Compliance to better understand who our sellers across Nigeria, Ghana, and Kenya.
What I did
- Created a Seller Risk Persona Matrix mapping sellers into Low, Medium, High risk tiers
- Captured traits, behaviors, documentation levels, and how each group should be onboarded
The early beginnings of trying to map out the Persona Matrix in Figjam
Why it mattered
- Helped us design tailored onboarding flows (e.g. soft nudges for informal sellers, strict gating for high-risk actors)
- Replaced blanket rules with conditional logic based on seller profile
- Improved trust, reduced drop-off, and sped up onboarding for legit businesses
Snapshots of first 3 steps of the verification flow
1
This is one major step to segment users into risk tiers. Based on what they sell (e.g. peer-to-peer loans = high risk), how they run the business (solo vs with a partner), and what country they’re from, my matrix kicks in to determine what flow they see next.
2
High-risk users are given a stricter review process without being punished for it
I softened the experience with:
- A “Save & continue later” escape point
- A “Need help?” section tailored for those who get stuck
- A callback prep screen to guide them before talking to support
3
Instead of showing the same requirements to everyone, I tailor document requests (e.g. asking for 1 document from low-risk businesses.
This shift from blanket rules to conditional requirements is a direct output of the persona matrix.
Crafting a story for this user(s) - Personas
I translated each risk group into goal-driven, behavior-rooted personas
What I did
- For Low-risk, the matrix showed readiness and legitimacy → I designed a persona that valued speed, trust, and minimal disruption
- For Medium-risk, the confusion and partial legitimacy stood out → the persona reflected a motivated but unsure seller who needs guidance, not friction
- For High-risk, the matrix flagged red zones → the persona became a cautionary archetype to design guardrails around
Why it mattered
- It bridged the gap between compliance risk logic and UX empathy
- It gave teams a way to design per risk, not despite it
- It helped cross-functional teams speak the same language about who we were optimizing for
Snapshot of the Persona of the owner of a medium-risk business
What does the journey of this user look like?
The next logical step was to map the actual journey each seller would face when joining the platform
What I did
- Low-Risk Sellers saw a clean, fast track — their journey reflected confidence, low friction, and automation opportunities
- Medium-Risk Sellers had uncertainty and doc confusion — their map surfaced education gaps and needed reassurance touch-points
- High-Risk Sellers showed mistrust, and rejections and their path reflected empathy, guidance, and escalation lanes
Why it mattered
- It gave product, compliance, and ops teams clear targets for intervention
- It directly fed into how we designed onboarding flows, doc validation logic, and support pathways
- It became a shared blueprint that connected business risk with product experience
Rough user journey of a Low-Risk Business
Rough user journey of a Medium-Risk Business
Rough user journey of a High-Risk Business
experience solutions
The final solution
...again...like I said
- Low-risk Nigerian Business selling Physical Goods
- Provides country/business specific document
- Successfully verified
- Medium-risk Ghanaian Business selling Health & Beauty Goods
- Provides country/business specific document
- Saves & continues later
- Resumes from last step
- High-risk Kenyan offering Peer-to-peer loans
- Doesn’t provide documents
- Saves & continues later
- Continues after 24 hours
- Sees “help section”
- Requests call back
Looking beyond the happy path - Edge Cases
- Auto-image quality check e.g. blur detection using frame colors as visual cues
- Prevents preventable delays from low-effort mistakes.
- Reinforces quality expectations without friction.
- Managing long waits during verification step using engaging content
- Designing for lost internet connection
- Added a recovery screen with clear messaging and retry option
- Built in progress preservation so users don’t lose work mid-verification
- Used empathetic copy to reduce stress and build trust
- Designed with low-bandwidth realities in mind, especially in Nigeria, Ghana, and Kenya
testing & iteration
Testing as early as possible - Prototype Usability
To observe how real business owners from across Africa experience the onboarding flow especially across risk levels and uncover usability issues, unclear copy, and emotional friction
Participants
- 2 Nigerian sellers (1 formal, 1 informal)
- 2 Ghanaian sellers (both informal)
- 1 Kenyan seller (formal but solo operator)
Outcome Highlights
“I didn’t realize the icon meant I had to take a selfie”
Ghanaian seller
Users misunderstood the abstract icon, leading to confusion about the selfie step and delays in continuing.
“The camera just snapped by itself and I didn’t know if it worked or not.”
Kenyan seller
Auto-capture felt foreign and unreliable to users who were used to tapping the shutter themselves.
“There was so much writing and I still didn’t know what documents to upload.”
Nigerian seller
Overwhelming text and no visual guidance made sellers second-guess what was required for verification.
Problems Identified
Users didn’t recognize the selfie icon as an actionable step. It failed to communicate purpose, causing hesitation and drop-off at the selfie capture step.
With all the information present, sellers couldn’t confidently identify which documents to upload
The automatic selfie capture broke familiar mental models. Instead of feeling seamless, it made users anxious about whether the photo was taken correctly or taken at all.
Iterating/Improvements after testing
1
Added sample selfie with verification badge. Sets a visual standard and subtly reassures users (“this is what success looks like”)
2
Added extra text to further guide users on exactly what to do to be successful in this step
1
Rewrote the guidance beneath the frame to be shorter and more direct; telling users exactly what to do, in the tone they expect from a native camera app.
2
Added “Why we need this” link
Introducing this link gave users helpful context, reducing skepticism and making the request feel more legitimate especially for high-risk users who are more cautious about privacy
3
Switched from automatic to manual capture.
Users were more comfortable taking control of the camera. Manual capture aligned with how they already use their phone cameras, reducing confusion and building trust in the process
the aftermath
Milestones
- 59 out of 64 businesses successfully verified and approved
- 92% of testers rated the onboarding flow as “clear” or “very clear”
- Onboarded all 3 seller types: low (38%), medium (44%), high (18%)
- 83% of users submitted valid docs on the first try
Next steps
- Continue monitoring product progress in beta
- Implement analytics on Mixpanel with PM and devs to monitor user behavior
- Prepare support team with real user edge cases (e.g. poor internet, ID rejections) so they’re ready to guide sellers smoothly during onboarding
What I learnt from this project
- Design for emotional reassurance, not just functional completion
- Treat edge cases as core cases; they reveal where real people struggle
- Ask more than “can they do it?” and instead ask “how does it feel while doing it?”
- Make the invisible visible; feedback, progress, and support need to be loud and clear
Contact Me
I dont bite...
© 2025 Frankie’s Portfolio