Sauce #18: Cal AI, TJM Labs
20 new tactics from 2 episodes this week. Here are the 12 with the most detail to copy.
A subscription mobile app · B2C · host's own app
Check that Meta's install and trial counts match App Store Connect before launching
312 trials in App Store Connect vs 276 in Meta; close enough to launch
How: Steve P. Young sends install, paywall view, checkout, trial start and subscribe through the Meta SDK, then compares Meta Events Manager with App Store Connect over the same dates. Downloads ran higher in Meta than in App Store Connect.
“Again, same time period, we have 312 under App Store Connect. On Meta, we had 276. So very close, I feel confident that these events are being tracked properly”
App Masters, 2026-10-20 · Read it in context · Listen from 13:36
Cal AI · B2C · AI calorie tracking app
Run a 9-second ad: one-line hook, the app in use, store link
Ad has run about four months, which the host reads as a winner
How: Cal AI's ad in Meta's Ad Library opens by asking if this is really all it takes to track calories, shows food, then the app in use, and ends on the App Store link. The host infers it performs from how long it has run.
“This creative is—been running for a very long time, so since June, and at this recording, about about four months of running the ad”
App Masters, 2026-10-20 · Read it in context · Listen from 15:01
TJM Labs · B2B · AI agents for pharmacies
Price each AI agent per year, just under one employee's salary
$45K per agent a year; pharmacies average 2.5-3 agents
How: TJM Labs charges about $45K a year per AI agent, slightly less than one pharmacy technician's salary, while an agent does the work of 2-3 technicians. It rejected per-seat and per-task pricing, so revenue grows as pharmacies add agents.
“our average ACV is about 45k a year, right? And if you notice, that is slightly less than the salary of a a pharmacy technician, and that's how we price it.”
Top Founders with Nathan Latka, 2026-10-20 · Read it in context · Listen from 7:47
TJM Labs · B2B · AI agents for pharmacies
Build your first version for one desperate customer you already know
Became a $1M+ a year customer; sole customer for 9 months
How: TJM Labs' founder had years of trust with a pharmacy that had to hire about 200 technicians in four weeks; he offered AI agents so it could hire fewer, and wrote the first code. Pharmacy has no test sandboxes, so trust got him access.
“So instead of hiring 200, maybe you have to hire only 100, right? But it ended up being extremely successful”
Top Founders with Nathan Latka, 2026-10-20 · Read it in context · Listen from 9:18
TJM Labs · B2B · AI agents for pharmacies
Sell AI agents as a replacement for hired staff, not as software
Founder says small pharmacies go from 2% to 8% margins
How: TJM Labs' agents take over technician data entry on e-prescriptions, faxes and paper scripts while the pharmacist double-checks. Independent pharmacies hear a margin pitch; grocery chains hear their pharmacy stops being a loss leader. The margin figure is the founder's claim.
“When you come in and implement a solution for this, they they go from a 2% margin to an 8% margin.”
Top Founders with Nathan Latka, 2026-10-20 · Read it in context · Listen from 13:08
Practitioner pattern · Steve P. Young
Wait to run Meta ads until 5-10% of installs start a trial
Host's agency rule of thumb; 30%+ trial-to-paid and ~2% install-to-paid also required
How: Steve P. Young's agency checks the funnel before any Meta spend: 5-10% install-to-trial, over 30% trial-to-paid, at least 2% install-to-paid. 1,000 installs at 7% gives 70 trials and 21 subscribers, the bare minimum. A rule of thumb, not a tested result.
“you want to at least have 5% to 10% install-to-trial activation, all right?”
App Masters, 2026-10-20 · Read it in context · Listen from 4:34
Practitioner pattern · Steve P. Young
Run Meta ads only once each install is worth at least $3
Host's minimum; Meta installs are expensive
How: Steve P. Young's agency requires at least $3 lifetime value per install, measured around month 11 rather than first revenue, read from RevenueCat, Adapty or Superwall. Apple Search Ads can work below this bar. His own threshold, no results given.
“we're going to stick with at least a $3 LTV per install”
App Masters, 2026-10-20 · Read it in context · Listen from 8:39
Practitioner pattern · Steve P. Young
Copy ideas from competitor ads that have run for months in Meta's Ad Library
Research step; no results reported
How: Steve P. Young searches competitors and complementary apps (MyFitnessPal, Strava for a calorie tracker), sorts by impressions and watches ads that started months ago, noting the hook, demo and end slide.
“It's automatically sorted by impressions, and then you're going to try to find something that has been running for a long time, right?”
App Masters, 2026-10-20 · Read it in context · Listen from 14:11
Practitioner pattern · Steve P. Young
Test 3-5 very different ad ideas first, small tweaks later
Host's launch practice; no results reported
How: For a calorie tracker, Steve P. Young builds distinct concepts: best way to track calories, a weight-loss before/after, and an app that tracks even snacks. Each has a hook, demo, benefit and App Store end page; 3-5 share one Meta ad set.
“I've now come up with three to five different ad concepts that I can just put into one ad set in Meta and let the algorithm do its thing.”
App Masters, 2026-10-20 · Read it in context · Listen from 20:39
Practitioner pattern · Steve P. Young
Set Meta to find trial starts, and budget for 50 a week
$1,000 at $2 installs gives ~35 trials, short of 50; ~$95 per subscriber
How: Steve P. Young optimizes for trials because install-optimized campaigns bring $1-2 installs but few payers, and Meta needs 50 events a week to learn. A first $1,000 mostly shows which creative works. A worked example, not a result.
“Meta's good at getting you whatever you optimize for, so if you optimize for installs, you might get a cheap install, maybe a dollar to $2, but nobody pays, so you're going to try to optimize for the trial.”
App Masters, 2026-10-20 · Read it in context · Listen from 9:17
Practitioner pattern · Steve P. Young
Send each audience's ad to its own App Store page with matching screenshots
How: Steve P. Young ties each Meta ad to an App Store custom product page whose screenshots match it, such as an ADHD screenshot for the ADHD ad. He calls it a later-phase step; at launch, just keep ad promises consistent with the screenshots.
“you can tie each Meta ad to a CPP, a custom product page, and you can say, look, for ADHD, I'm going to show an ADHD screenshot when they land on the App Store page.”
App Masters, 2026-10-20 · Read it in context · Listen from 18:47
Practitioner pattern · Steve P. Young
Show Meta ad users a hard paywall during onboarding
How: Steve P. Young always puts the paywall in onboarding and suggests a hard paywall for all Meta paid traffic, with the free trial on whichever plan, yearly or weekly, has the highest LTV. With a hard paywall he aims for 10%+ install-to-trial. No results reported.
“always have your paywall during the onboarding, and I might argue that you might want a hard paywall for all Meta paid traffic”
App Masters, 2026-10-20 · Read it in context · Listen from 4:34
Browse every tactic on GTM Sauce, each with the conversation around its quote.
Episodes this week
- App Masters: Before You Spend $1 on Meta Ads, Watch This (2026-10-20)
- Top Founders with Nathan Latka: Doctor Strikes it Rich with AI Tool, $10m ARR in 6 Months | TJMLabs CEO Jonathan Adly (2026-10-20)