Asher Herbstman & Thatcher Clough - I Built a $30k/mo App That Helps You Quit Smoking

After an uphill battle with organic, influencers, Reddit and SEO, moved to paid ads on Meta (Instagram) and TikTok. Realized the flywheel depends on the product converting well enough to earn multiple dollars per install, so they worked on onboarding and paywall to make paid profitable.

Early on tried many channels: organic content, influencers (with referral codes in onboarding), Reddit and SEO.

Founders script and film their own ads, starring family (Asher's actor dad, his mom). Pick the demographic likely to commit and pay, then apply persuasion basics: strong headline, build trust through the ad, clear call to action. Hooks use hyper-specific, even absurd callouts ('Do you hang out in your basement after school hitting your pen until 3am watching skateboarding videos?'). Some best converters are Asher's mom explaining the problem calmly to an older audience.

Break down paywall conversion by onboarding answers (age, dream outcome, why quitting, triggers). Found users seeking mental clarity, feeling out of control, or quitting for motivation convert higher; 26-30 year olds convert almost double 18-20 year olds; 'saving money' converts at 23% vs 29% for legal obligations. Used this to aim ads at older audiences and to plan onboarding copy stressing money saved vs the $30/year price.

Run a mobile measurement partner so Meta/TikTok algorithms can optimize, but also add an onboarding attribution question (knowing Instagram, listed first, gets some bias). Amplitude dashboards show ROAS after the Apple fee, trials started by channel, and cost per conversion on TikTok vs Instagram, used to rebalance ad spend during the day.

Originally no data was collected until users signed in with a phone number, hiding how many opened the app and quit. Added tracking from app open to see drop-off at every onboarding step and time spent per question.

After tearing down Cal AI, Quitter and Ahead onboardings: intro screen shows the dream outcome (30 days clear-headed), not features. Ask their name, then use it in the next question. Ask short- and long-term goals, then reflect them back with social proof ('81% of users saw improved mental clarity'). Ask pain questions (days a week, money spent), then compare them to others with cited sources. Then show where they'll be in 30 days (money saved), that the program does the work, that 25,000+ people did it, and that users improve within the first 3 days. Frames every slide on the value equation: dream outcome x likelihood / (time delay x effort).

Two questions (what do you want to achieve / what triggers you to smoke) take users the longest, 20-30 seconds each. Rather than cutting them, kept them as 'good friction' that commits users to the product and to themselves.

After the questions, show free personalized feedback: current use level, 'ability to build a strong habit: 100%', money they could save, strengths and 'areas for exploration' (never 'weaknesses', to avoid shaming).

Recently added question: 'How do you plan to use the app?' with a list of features (groups with others, video journaling, etc.), sneakily showing features while collecting data on what new users want.

Copied from the Ahead onboarding: before the paywall, a slide says plainly 'this costs money, we're doing it fair, we have a fair trial', contrasting the cost of staying in the habit with the dream outcome, so users reach the payment page anchored to the habit's cost. Paired with messaging that it's more than a tracker (peer support specialist) to cut social media complaints about being paid (goal: -50%).

Copied from Cal AI's onboarding: a social-proof slide ('designed for people like you', five stars, member count, real reviews) that triggers the native rating prompt before users see the reviews. Placed after 5 minutes of reflective questions and free feedback, so users are happy when asked.

Sign-in with phone number (chosen over Google sign-in despite more friction). Send SMS via Twilio with hand-written, humorous, non-spammy copy (not generic AI 'come back to the app'); two-way: users can text the team and get replies.

Old paywall (inspired by Cal AI) opened with a feature carousel (accountability texts, peer support, '1,000+ members'). New first screen outlines the program day by day (day 0, day 1... first 7 days), with small science-style quick wins to cut the perceived time delay (e.g. 'CB1 receptors begin to bounce back within 48 hours'), without promising a fix tomorrow. Rest of the paywall unchanged.

Blinkist-style trial messaging: 'no payment due now', price shown clearly, 'we'll remind you before your trial ends'. Yearly option with 3-day free trial plus monthly. Reminder push 2 days in, turned on in Superwall. Now working on what users see when they tap that reminder.

Early price tests on the yearly plan at $30, $40 and $50, using higher prices as an anchor; ended on $30/year.

Borrowed from alcohol-quitting app Reframe's 'take our quiz to see what alcohol type you are' ads: pitched 'take the quiz, see your relationship percent with weed' and framed onboarding as a quiz.

Look at creatives the Meta/TikTok algorithms allocate most spend to and those with high click-through, work out why, and carry winning hooks (e.g. the trigger of friends smoking around you) into new scripted ads. Because ads are scripted and filmed, they don't do many one-word tweaks. No formal hypothesis doc yet; recommends keeping one.

University of Michigan partnership pending to offer the app to faculty; exploring addiction and state services. Users ask for supplements and therapy, so planning upsells/downsells to raise LTV.