In partnership with
Hey, it’s Jeffrey — back again!
I went into this week's prompt ready for carnage. I've been paying for subscriptions for years and I figured the audit would be ugly. Streaming services we forgot about. Apps I signed up for once. The usual "personal finance blog tells you to cancel everything" energy.
That's not what happened.
I pulled every charge out of Monarch, had Claude look for recurring charges and subscriptions, added in the apps that bill through my phone (worth checking separately — they don't always show up cleanly), and listed how often my household actually uses each one.
Annual total across everything: $3,783. I expected to kill maybe half of that.
Here's the prompt I used:
I want to audit every subscription I'm currently paying for and calculate the real cost per use. Walk me through this one question at a time.
For each subscription I tell you about, ask me:
What is it?
How much does it cost per month (or per year, if it's annual)?
How often do I actually use it in a typical month?
Is anyone else in my household using it too?
After I list my subscriptions, calculate the cost per use for each one and sort them from worst value to best value. Flag anything I forgot I was paying for, anything that's passive (like cloud storage or home security) where cost-per-use doesn't really apply, and anything that might be better thought of as a donation rather than a subscription.
Then give me three categories: keep, cancel, or reconsider. For the reconsider category, tell me what specifically to check before deciding.
Start by asking me to list my first subscription.
Here's what the audit found on my numbers:
My best value across the entire stack is Carrot Weather. I pay $40 a year for it. I check it two or three times a day. That works out to about six cents per use. Sure, it’s a splurge since I can get weather for free, but I enjoy it.
Apple Music Family was the second surprise. Two hours a day of listening, my wife uses it too, $18.18/month. That's thirty cents per hour. For a family plan.
Pickleball came in at $5.83 per session. Great value considering the positive impacts on my health. It stays.
The real find was a soccer newsletter I pay $5/month for. I read it maybe twice a month. That alone wouldn't kill it. But when the audit asked me about it, I'd genuinely forgotten I was paying. That's the actual tell. Forgotten subscriptions are where the money leaks, regardless of the per-use math.
The other cancellation was Calm. $40/year, and I used to use it daily. These days it's twice a month. The math isn't terrible, but the drift is. I've been paying for a habit I don't have anymore. Gone.
Two cancellations. About $100/year back. Out of $3,783.
So: I expected carnage. I got two cancellations, a couple of pleasant surprises, a better understanding of where $3,783 a year is actually going.
The best thing this prompt does isn't save you money. It tells you, concretely, what you're buying with each recurring charge. Most of mine turned out to be worth it. A couple weren't. I wouldn't have known which was which without running it.
Keep going? If you want to push this further, here are three follow-ups worth running:
1. Check your App Store subscriptions separately. On iPhone: Settings → your name at the top → Subscriptions. Android: Play Store → profile → Payments & subscriptions. These often bundle under generic merchant names like "Apple" or "Google" and can quietly dodge the audit. This is where I found four of my recurring charges.
2. Ask the AI to stress-test your "keep" list. Paste your final list back and ask: "If I had to cut 20% of this, which would you cut and why?" Forces a harder look at the stuff you defended on the first pass.
3. Run it again in 6 months. Put a reminder on your calendar. Usage patterns change. The streaming service you used four times a week in April might get one watch in October. Cost-per-use is a snapshot, not a permanent answer.
AI tip worth trying this week: Build a custom AI skill for the exact way you get stuck.
Sometimes I freeze on tasks by mentally jumping to the hardest part. Not the first part. The scariest part. I can be staring at a project where step one is "open a Google Doc" and my brain has already fast-forwarded to the meeting in three weeks where someone might push back on it.
So I built a Claude skill for it. I told Claude exactly how I get stuck, what doesn't help (pep talks, full project plans, reminders of everything else on my list), and what does help (one step so small that starting it feels trivial, and answers to the obvious micro-questions inside that step so I don't have to stop and think).
A real example: My propane heater in the van wouldn't fire up. It told me step one was to lift the tank to see if it was empty. An empty tank is about 17 pounds. Full is 37. That's it. That was the whole first step. Turned out the tank wasn't empty, which ruled out the easy answer in about 30 seconds and told me to look at the heater setup next. Problem solved that afternoon.
The generic productivity advice ("break it into smaller steps") never worked for me because I'd break it into steps that were still too big. The skill forces a specificity my brain won't do on its own.
If there's a pattern in how you get stuck, Claude can be taught to meet you exactly there. It's one of the best uses of custom skills I've found.
Stop optimizing for clicks. Start driving pipeline.
Rising costs. Signal loss. Platform changes. Most paid media fails because it's built for clicks, not revenue.
On April 27th, HubSpot's former Head of Paid breaks down the exact framework for structuring campaigns that drive real pipeline in 2026. 20 minutes. Live Q&A. Free.
How did you like today's newsletter?
One quick note: This newsletter is for educational purposes only and does not constitute financial advice. I'm not a financial advisor — just someone sharing ideas and tools I've found useful. Use what works for you, skip what doesn't, and always do your own research. Some links may be affiliate links or sponsored content for which I may receive compensation.