Fixing Travel's $20 Billion Problem
It's time to build
I’ve been co-living/working in Hack House. Last week, Hack House ran a hackathon for all its members with some insane prizes— Over $25k in cash as well as world-class retreats.
As a hackathon hacker myself, with 10 career wins, I saw as the perfect opportunity to build and demo something cool in front of VCs and potential customers.
After sifting through a catalog of 320 startup ideas, I arrived at 4 worthwhile ideas:
CI and security suite for smart contracts: Smart contracts are notoriously sketchy, and the CI tooling ecosystem around them is still fairly nascent. This suite would be a tool that checks the code for potential vulnerabilities and bugs before the smart contract is deployed. Similar to Checkmarx and Sonarqube
AI-driven frontend component library: Most web interfaces suck. Almost every site out there asks you to fill out the same information again and again. This component library would integrate a number of AI-enhanced UI features including Optical Character Recognition (OCR), autocorrect, response autosuggestions with Natural Language Processing (NLP), AI image upscaling, email autocorrect, etc. Builds upon the capabilities of my last project, DocIt.
Credit card that “taxes” the user extra for each purchase: This custom card will penalize the user for spending in areas that they should be avoiding and adds a flat extra charge per purchase and donates it to charity. Similar to the Aspiration Card.
VAT refunds for receipts for non-residents: Most countries with Value Added Tax allow non-residents to refund the amount they spent on taxes. So, if you buy something abroad, you can get the VAT portion of your bill refunded. This idea is a cross between a premium Amex travel card and VatBox, but B2C.
The $20 billion problem
After a lot of reflection, I landed on the last option. Last February, I spent a month digital nomading around Portugal. A few pain points stuck out to me:
Traveling internationally when your bank deals with American dollars is uniquely difficult. Not only do credit/debit cards charge you tons of fees, but they’re also extremely deceptive. You can easily lose hundreds in bad DCC conversions, ATM withdrawal fees, and lousy foreign exchange rates.
130 countries around the globe have VAT taxes. 54 of them allow you to get refunded for your spending if you’re a non-resident. VATs usually range between 15-25%, but go as high as 27% in Hungary. (Just imagine, a credit card with 27% cashback). But a lot of people don’t even know how to get their money back since the process is somewhat complicated. For this reason alone, businesses lose $20 billion in unrefunded VATs every year.
The user experience for most cards sucks. I pride myself on churning credit cards, so I’m more than familiar with the various interfaces out there. And the UX for all the apps I’ve tried is just horrendous. Here’s a screenshot from my Schwab Checking account. It has the best currency exchange rates, no fees for any ATM on the planet, and no international transaction fees. But the UI leaves much to be desired.
Accounting Expenses vs Recounting an Adventure
This led to the hypothesis: Spending abroad is much more expensive and clunky than it should be. And this is a prime area for disruption.
The Hackers
As it turns out, publicly sifting through that massive list of ideas caught some attention of the other hackers in the house. And some amazing friends, Tanzeela, Derek, and Andrew, decided to join forces with me on this problem.
We attacked the problem on 3 fronts:
UI/UX: We wanted to imagine a better credit card software experience. Amex, Barclays, Citi, etc. all have “sufficient” application experiences, but we wanted to go a step further.
Rewards: The selling point that caught the most excitement was automatic VAT redemption. We wanted to imagine what a card with 27% cashback would look like, and we had to engineer a solution to reach that point. Building the receipt processor took a lot of machine learning, computer vision, and business logic.
Brand: We wanted to make the card feel real. So, we found a local metal credit card designer and built some incredible luxury cards. We even engraved the names of the judges on them.
With less than a week to go and countless hours spent in front of our screens, we built out Roam - The World’s Best Travel Card.
Here was the idea: Build a travel card with automatic VAT refunds on a user’s purchases, the best exchange and transaction rates, and a lovely, clean UX.
Derek and Tanzeela were our UI wizards. Andrew churned out some amazing Figma sketches. And within a few hours, Derek and Tanzeela prototyped a spiffy React Native app:
Being the data scientist/machine learning engineer on the team, I built an OCR + NLP service to scan receipts that aren’t automatically refundable.
We also jumped into hardware. Andrew and I found a lab that made custom metal cards.
The cards turned out pretty swell, I’d say.
Rome wasn’t built in a day. And neither was our card. But we demoed off several key features: Swiping the physical card with a Square reader, redeeming VAT receipts, and generating barcodes for customers. (I’m surprised we didn’t run into more bugs!)
But most importantly, our goal was to focus on the problem, not the solution.
The Reception
And the hackathon results…?
We lost!! (Big sad).
But am I discouraged? Hell no! In fact, we probably learned more by losing than if we had won.
We tracked down the judges afterwards and asked for feedback. Across the board, the judges loved the idea and the solution. VAT refunds and travel finance are tremendously clunky and we are solving a real problem.
Here’s the feedback we got:
“For investors talk about how you’ll make the credit/banking side happen- how are you managing people’s debt? Do you have partnerships in place or in talks”
“Would love to chat more. We have invested in an adjacent project called SmartHotelRate that I think would be a great collaborator.”
“Super interesting idea! I see so many ways that you could partner with other giants in the travel space (air hotels) and the VAT problem is real!”
“Would love to hear more about the future, other than recommendations”
But the question that was on everyone's mind was: “Is this truly venture scalable?”
Ultimately, I believe this is was what did us in. It’s like missing the forest for the trees. Sure, travel spending is a profitable subsegment of the market, but is this really going to be massive? Like, really massive? For example, Uber started as a taxi cab app, but its broader goal was to become the world’s largest transportation and delivery logistics network.
What’s Next?
While the team loved working together, we all agreed that Roam still has a lot of work to go before any of us feel confident jumping full-time into it. Plus, we all have busy jobs. But I’m going to continue working on validating the idea, and the rest of the crew offered to join in once there’s a clearer vision.
So what am I working on now?
Talk. With. Customers.
I’m about 80% confident that we’re onto something with Roam— Either the VAT refunds, the brand, or something else. At best, this is just a hunch. The plan going forward is as follows:
Talk with the people who signed up. Set up short interviews and “Mom Test” them.
Create landing pages and see what sticks. Realistically, I’ll probably spend ~$1k on design and several hundred dollars on ads and gauge whether customers care enough to sign up.
I’m working with the hypothesis that customers want what I outlined. If customer interviews give me sufficient reason to believe we’re onto something, then we commence build mode. And by build mode, I mean the scrappiest MVP imaginable.
If all 3 of these go smoothly, then we’re onto something and this project goes from hobby hack to a real business.
Stay tuned to see how this experiment turns out. And want to get your hands on Roam early? Sign up here! www.roam.creditcard
