
A debt payoff app that uses roundups on purchases to automatically pay down debt.
Ditch pivoted to debt roundups in May 2024 after a painted-door test beat everything else, then used a hard paywall and a long, scored onboarding to scale from $1M ARR in two years to $3M within about 140 more days. About 80% of ad spend runs on Meta, with 90% of creative from UGC. Lesson: build only what pulls, prove unit economics small, then spend.
From Luis Andino, Ditch founder interview
Ditch spent two years proving people would pay before spending on growth. The first $1M ARR took two years, the next $1M took 80 days, the third took 59 days, and about 16 days after $3M it was approaching $4M. Those first two years went into onboarding, activation and retention.
Ditch launched in 2023 as an AI debt companion and only took off in 2024 after its pivot. It was venture-backed from day one and could have chased users first, but Luis Andino feared it would then never build something people pay for. "For the first two years of Ditch, we just became obsessed with one question. How can we build a product that a user would want to pay for?"
A hard paywall served as the test of whether the product mattered. Ditch set the barrier to entry as high as possible so experiments taught it something real. Optimizing trial rates at a price below what unit economics need means redoing every test once the price goes up.
Andino notes that a 60 to 70% trial conversion means little if the price is too low. "We wanted to take the approach of let's use this hard paywall as the sort of litmus test for whether we've built something that mattered."
Once payback is short, scaling spend becomes a responsibility. After proving payback at small scale, Ditch raised capital and poured it into growth. Andino tells founders with seven-day or day-zero payback that they should spend more.
He frames the goal as a ubiquitous brand, which requires growth. No single factor drove results, only a sequence of experiments designed to fail in ways Ditch could learn from.
Painted-door tests showed debt roundups beat every other idea. Ditch launched a canvas-like app instrumented to measure every action and put features behind buttons to count taps before building them. Roundups pointed at debt crushed everything else, so Ditch pivoted in May 2024.
Andino called the app an "intent radar" and "intent magnet" and only built what pulled. The mechanic borrows Acorns' roundups for saving and aims it at a problem he calls "way more contemporary": a $4.50 coffee sends a 50 cent roundup to a student loan, credit card or mortgage.
The product delivers progress passively because users won't do self-service work. People who need financial help rarely read charts or education, and an app cannot out-compete social media for attention. Ditch asks for one decision in about 10 minutes of onboarding, then progress compounds on its own.
Andino says budgeting and net worth apps mostly attract people who don't need them or want a pretty number. "We wanted to build a tool that automated progress so that it felt inevitable."
The real user is a 29-year-old professional woman, not someone in crisis. Ditch assumed users in financial crisis, but 70% of users are women, the average user is a 29-year-old woman, and many are nurses, doctors, consultants and tech workers with moderate to high incomes.
Users carry mid to high interest student loans, credit cards, personal loans and BNPL with a clear path out. Andino's LendingPoint data showed heavy debt comes from following the default path of student loans, a first car and a credit card, and 80% of Americans manage some consumer debt. Roundups are not meant for someone with $100,000 of credit card debt.
Pushing users to pay the maximum caused churn, so Ditch right-sizes goals. Early on Ditch pushed users to pay as much as possible, and many panicked after a week and needed the money back. It now models linked checking and credit card data to recommend a realistic payoff amount.
"We were pushing the user to make as much as humanly possible and we were actually inducing churn because of this."
Every onboarding ask is a debit that needs earlier credit. Screens where the user gives something, like name, email or date of birth, count as debits; screens that deliver value count as credits. Ditch assigns point values and sizes each ask to the credit before it, such as asking for account creation or a trial right after a big win.
"You can't expect that the first action that the user does is a debit because there is no balance."
A simulated roundup gives users the payoff moment before signup. Users imagine buying a $3.50 coffee, tap complete purchase in a bottom sheet labeled as simulated, and watch the roundup flow to their debt with a celebration. Ditch measured this screen as a very large credit.
Six or seven pre-paywall screens show personalized ROI. Before the paywall, Ditch uses the debt details users entered to argue the product will pay for itself, comparing interest saved from extra payments to its fees.
Being fintech lets Ditch quantify ROI directly from interest saved.
Removal A/B tests measure what each onboarding screen is worth. A longer onboarding produced a massive lift in trial start and conversion. Ditch then ran reductive tests that remove one screen at a time and deletes any screen whose removal changes nothing.
The tests also give a proxy value for each screen. "If you only go one way and you do that for five years of experiments compounding then you're going to end up with this like massive thing."
From Luis Andino, Ditch founder interview
Organic UGC that wins on TikTok also wins as paid ads. Experts told Andino organic winners fail as paid. Ditch tested anyway: its TikTok UGC program fed top videos into Spark ads optimized first for impressions, then installs, then trials, then onto Meta, and results held at each step.
About 90% of Ditch's ad content now comes from the UGC program. Winning formats were trending audio with product demos, talking heads and founder ads. The worst case for each test was a few hundred dollars of spend.
A thin brand layer on UGC makes repeat ad impressions compound. Raw UGC varies in setting, creator, audio and text, so a second ad doesn't feel like the first. An in-house editor applies Ditch's standard text, edit style and overlays to every UGC video chosen for paid.
This runs against common advice to let creators run free. Andino says peers spending seven and eight figures a month on ads found the idea useful.
About 80% of the ad budget goes to Meta. Andino calls Meta the best ad network for matching a profile with a relevant ad. TikTok-style video feels native on Instagram because the same people use both.
He says the choice of conversion event largely decides how well Meta finds the right person, and that platforms change auction rules often enough that Ditch tests from first principles.
A processing fee ties part of revenue to value delivered. Ditch charges a processing fee only on payments actually made, so cost tracks the value users get. The average user pays 2.7x their base roundup amount against debt.
Andino wants to add value without raising prices, such as a refinancing marketplace for users overpaying on student, auto and credit card interest.
Always-on AI agents could refinance debt in minutes. Agents would watch a user's debt composition and take them from an offer to a signed refinance automatically, possibly many times a year.
His example is a six-month window where an 8.5% student loan could be refinanced to 4%, which most users would miss.
Community and challenges aim to make a stigmatized product shareable. Admitting to high-interest debt limits word of mouth, so Ditch plans community, challenges and gamification, plus employer benefits as embedded distribution.
Andino cites a Wall Street Journal article on Americans being addicted to watching others pay off debt.
Investing in agentic workflows keeps a lean team highly leveraged. Ditch sets up connectors and builds skills for recurring tasks, asking of nearly every task whether an agent can do it.
Andino warns that with poor workflows, teammates new to AI turn averse after the first agent mistake.
Fundraise in a tight, committed process that creates urgency. Many VCs dislike consumer subscription apps and try to retrofit their own go-to-market ideas, so Andino qualifies advice and holds his vision. He runs a clear process, creates scarcity and sells a big vision.
"You can't be half in, half out. You have to have a very clear strategy."
Reached $3M ARR: first $1M took two years, next $1M 80 days, third 59 days, then about 16 days toward $4M. Pivoted to debt roundups in May 2024, uses a hard paywall, scores onboarding screens as credits and debits, and puts about 80% of ad spend on Meta with 90% of ad content from UGC.
Ditch spent two years proving people would pay before spending on growth. The first $1M ARR took two years, the next $1M took 80 days, the third took 59 days, and about 16 days after $3M it was approaching $4M. Those first two years went into onboarding, activation and retention.
Ditch launched in 2023 as an AI debt companion and only took off in 2024 after its pivot. It was venture-backed from day one and could have chased users first, but Luis Andino feared it would then never build something people pay for. "For the first two years of Ditch, we just became obsessed with one question. How can we build a product that a user would want to pay for?"
A hard paywall served as the test of whether the product mattered. Ditch set the barrier to entry as high as possible so experiments taught it something real. Optimizing trial rates at a price below what unit economics need means redoing every test once the price goes up.
Andino notes that a 60 to 70% trial conversion means little if the price is too low. "We wanted to take the approach of let's use this hard paywall as the sort of litmus test for whether we've built something that mattered."
Once payback is short, scaling spend becomes a responsibility. After proving payback at small scale, Ditch raised capital and poured it into growth. Andino tells founders with seven-day or day-zero payback that they should spend more.
He frames the goal as a ubiquitous brand, which requires growth. No single factor drove results, only a sequence of experiments designed to fail in ways Ditch could learn from.
Painted-door tests showed debt roundups beat every other idea. Ditch launched a canvas-like app instrumented to measure every action and put features behind buttons to count taps before building them. Roundups pointed at debt crushed everything else, so Ditch pivoted in May 2024.
Andino called the app an "intent radar" and "intent magnet" and only built what pulled. The mechanic borrows Acorns' roundups for saving and aims it at a problem he calls "way more contemporary": a $4.50 coffee sends a 50 cent roundup to a student loan, credit card or mortgage.
The product delivers progress passively because users won't do self-service work. People who need financial help rarely read charts or education, and an app cannot out-compete social media for attention. Ditch asks for one decision in about 10 minutes of onboarding, then progress compounds on its own.
Andino says budgeting and net worth apps mostly attract people who don't need them or want a pretty number. "We wanted to build a tool that automated progress so that it felt inevitable."
The real user is a 29-year-old professional woman, not someone in crisis. Ditch assumed users in financial crisis, but 70% of users are women, the average user is a 29-year-old woman, and many are nurses, doctors, consultants and tech workers with moderate to high incomes.
Users carry mid to high interest student loans, credit cards, personal loans and BNPL with a clear path out. Andino's LendingPoint data showed heavy debt comes from following the default path of student loans, a first car and a credit card, and 80% of Americans manage some consumer debt. Roundups are not meant for someone with $100,000 of credit card debt.
Pushing users to pay the maximum caused churn, so Ditch right-sizes goals. Early on Ditch pushed users to pay as much as possible, and many panicked after a week and needed the money back. It now models linked checking and credit card data to recommend a realistic payoff amount.
"We were pushing the user to make as much as humanly possible and we were actually inducing churn because of this."
Every onboarding ask is a debit that needs earlier credit. Screens where the user gives something, like name, email or date of birth, count as debits; screens that deliver value count as credits. Ditch assigns point values and sizes each ask to the credit before it, such as asking for account creation or a trial right after a big win.
"You can't expect that the first action that the user does is a debit because there is no balance."
A simulated roundup gives users the payoff moment before signup. Users imagine buying a $3.50 coffee, tap complete purchase in a bottom sheet labeled as simulated, and watch the roundup flow to their debt with a celebration. Ditch measured this screen as a very large credit.
Six or seven pre-paywall screens show personalized ROI. Before the paywall, Ditch uses the debt details users entered to argue the product will pay for itself, comparing interest saved from extra payments to its fees.
Being fintech lets Ditch quantify ROI directly from interest saved.
Removal A/B tests measure what each onboarding screen is worth. A longer onboarding produced a massive lift in trial start and conversion. Ditch then ran reductive tests that remove one screen at a time and deletes any screen whose removal changes nothing.
The tests also give a proxy value for each screen. "If you only go one way and you do that for five years of experiments compounding then you're going to end up with this like massive thing."
Organic UGC that wins on TikTok also wins as paid ads. Experts told Andino organic winners fail as paid. Ditch tested anyway: its TikTok UGC program fed top videos into Spark ads optimized first for impressions, then installs, then trials, then onto Meta, and results held at each step.
About 90% of Ditch's ad content now comes from the UGC program. Winning formats were trending audio with product demos, talking heads and founder ads. The worst case for each test was a few hundred dollars of spend.
A thin brand layer on UGC makes repeat ad impressions compound. Raw UGC varies in setting, creator, audio and text, so a second ad doesn't feel like the first. An in-house editor applies Ditch's standard text, edit style and overlays to every UGC video chosen for paid.
This runs against common advice to let creators run free. Andino says peers spending seven and eight figures a month on ads found the idea useful.
About 80% of the ad budget goes to Meta. Andino calls Meta the best ad network for matching a profile with a relevant ad. TikTok-style video feels native on Instagram because the same people use both.
He says the choice of conversion event largely decides how well Meta finds the right person, and that platforms change auction rules often enough that Ditch tests from first principles.
A processing fee ties part of revenue to value delivered. Ditch charges a processing fee only on payments actually made, so cost tracks the value users get. The average user pays 2.7x their base roundup amount against debt.
Andino wants to add value without raising prices, such as a refinancing marketplace for users overpaying on student, auto and credit card interest.
Always-on AI agents could refinance debt in minutes. Agents would watch a user's debt composition and take them from an offer to a signed refinance automatically, possibly many times a year.
His example is a six-month window where an 8.5% student loan could be refinanced to 4%, which most users would miss.
Community and challenges aim to make a stigmatized product shareable. Admitting to high-interest debt limits word of mouth, so Ditch plans community, challenges and gamification, plus employer benefits as embedded distribution.
Andino cites a Wall Street Journal article on Americans being addicted to watching others pay off debt.
Investing in agentic workflows keeps a lean team highly leveraged. Ditch sets up connectors and builds skills for recurring tasks, asking of nearly every task whether an agent can do it.
Andino warns that with poor workflows, teammates new to AI turn averse after the first agent mistake.
Fundraise in a tight, committed process that creates urgency. Many VCs dislike consumer subscription apps and try to retrofit their own go-to-market ideas, so Andino qualifies advice and holds his vision. He runs a clear process, creates scarcity and sells a big vision.
"You can't be half in, half out. You have to have a very clear strategy."
FinanceA finance automation platform combining corporate cards, bill pay, procurement, treasury and accounting in one place.
FinanceA crypto exchange and platform for buying, storing and moving digital assets, including stablecoins.
FinanceA commission-free retail investing app for stocks, options and crypto.
FinanceA global digital bank app offering accounts, cards, trading, crypto, credit and business banking in one place.
FinanceA Latin American digital bank that started with a no-fee credit card in Brazil.
FinanceAn app that helps people find and claim class action settlements.
A revenue-based financing company that advances cash to app developers against store and ad network receipts.
FinanceAn international money transfer and multi-currency account provider that sells low-cost, fast cross-border payments to people and businesses.