Fadé Adeniyi
Friday 30 October 2026
You already have the users. Here's how to sort them, learn from the ones who care, get the stalled ones using it and ask for money at the right moment, with Claude prompts and ready-to-send emails you can copy.
Most of your 500 users haven't used the product enough to see why they'd pay for it, so a discount email to all of them won't fix it. Sort them into active, stalled and never started, learn from the active ones, get the stalled ones to the moment it clicks, then ask for money straight after it.
Picture your user list right now. A few people log in every week, a big chunk tried it once and drifted off and plenty signed up and never came back. Each group needs a different message. Send all 500 the same email and you waste the attention of people who already said yes once.
Step 1. Sort the 500. Export your users as a CSV from your database, Stripe or whatever tool you signed them up with. Upload it to Claude with code execution and file creation switched on under Settings, then Capabilities. Then paste this.
I've attached an export of my users. Each row has a user ID, email, signup date, last active date, how many times they've used [THE MAIN FEATURE] and whether they pay. My product is [WHAT IT DOES] for [WHO IT'S FOR]. 1. Put every user who doesn't pay into one of three groups: ACTIVE: used the main feature in the last 14 days. STALLED: used it at least once, but not in the last 14 days. NEVER STARTED: signed up and never used it. 2. Give me the count for each group and for paying users. 3. Save one CSV per group with user ID, email and last active date. 4. Tell me anything in the data that looks wrong or missing. Don't change or delete anything in my original file.
Check a few rows by hand. Pick five people you know and make sure Claude put them in the right group before you email anyone.
Steps 2 to 5 work through the groups in order: talk to the active users, bring back the stalled ones, make the ask, then track who moves each week.
The 5 steps
Your active users found the value on their own. Ask them what it is before you write a price, because their answers become your paid plan and your upgrade email.
Use the question Superhuman used. In 2018 founder Rahul Vohra wrote that Superhuman asked users how they'd feel if they could no longer use the product. The benchmark he used, Sean Ellis's, is that companies with fewer than 40% answering "very disappointed" almost always struggled to grow. Superhuman's first score was 22%.
Email 1. Send it to your ACTIVE group from your own inbox.
Subject: Three quick questions about [PRODUCT] Hi [FIRST NAME], I'm [YOUR NAME] and I built [PRODUCT]. You've been using it for [WHAT THEY DO IN IT], so you're exactly the person I want to hear from. Could you reply with a line on each? 1. How would you feel if you could no longer use [PRODUCT]? Very disappointed, somewhat disappointed or not disappointed? 2. What do you mainly use it for? 3. What would make it worth paying for? If talking is easier, grab 15 minutes here: [CALENDAR LINK] Thank you, [YOUR NAME]
Prompt 2. Paste the replies into Claude once they're in.
Here are the replies from my most active users: [PASTE REPLIES OR CALL NOTES] 1. Count very, somewhat and not disappointed. Give me the very disappointed share as a percentage. 2. For the very disappointed group, summarise what they use it for and the main benefit, in their words. 3. List everything people said they'd pay for and how many mentioned each. 4. Suggest a paid plan from this: what goes in it and what stays free. Quote people exactly. Don't invent answers for anyone who didn't reply.
Listen first. If nobody says "very disappointed", fix the product before you charge for it.
Find the one early action paying users share. At Facebook, Chamath Palihapitiya has said his growth team's biggest realisation was getting any individual to 7 friends in 10 days. At Slack, Stewart Butterfield said in 2015 that any team which had exchanged 2,000 messages had really tried it and 93% of those customers were still using it. Upload your events export to find yours.
Here's my events export (user ID, event, timestamp, pays or not). Use the groups from before. 1. List what paying users did in their first 7 days and how often, next to stalled users. 2. Find the early actions paying users took far more often and give me a simple threshold, like "did X 3 times in 7 days". Tell me how many users sit on each side. 3. For stalled users, show the last step they reached before they stopped and where most of them dropped off. 4. Flag anything that only looks important because paying users use the product more overall. Give me our aha action in one sentence and how sure you are.
Only a handful of payers? Ask them what they did in their first week instead.
Clear the path before you email anyone. Fix the step where Prompt 3 says most people dropped off, such as an empty first screen, settings before any value, early invites or a long tour. In Chameleon's 2020 data, three-step tours were finished 66% of the time and completion fell below 50% after five.
Email 2. Send it to STALLED users. For NEVER STARTED, swap the second line for the job they signed up to do.
Subject: The part of [PRODUCT] you haven't tried yet Hi [FIRST NAME], You signed up to [THE JOB THEY CAME FOR] and got as far as [LAST STEP THEY REACHED]. The people who get the most out of [PRODUCT] all do one thing early: [AHA ACTION IN PLAIN WORDS]. It takes about [TIME] and I've [WHAT YOU FIXED, E.G. SET UP A SAMPLE PROJECT] so you can do it straight away: [ONE LINK THAT OPENS THAT ACTION] If something stopped you last time, reply and tell me what it was. [YOUR NAME]
Ask straight after the value. In RevenueCat's 2026 report on more than 115,000 mobile subscription apps, median conversion to paid by day 35 was 10.7% for apps with a hard paywall and 2.1% for freemium apps. That's mobile data, so test it on yours, but it suggests a clear ask converts better than leaving people on free forever. Put the upgrade prompt right after the aha action.
Make your first payers founding members. Give a lower price to anyone who upgrades before a real date. The email, the page and the checkout all use that one deadline, the full price applies after it and you never extend it.
Here's my pricing today: [FREE PLAN, TRIAL, PAID PLANS, PRICES]. Our aha action is [ACTION]. Here's what my active users said they'd pay for: [PASTE THE SUMMARY FROM PROMPT 2]. 1. Compare when users meet the upgrade prompt today with when they reach the aha action. Tell me where it should sit. 2. Using only what users told me, suggest what stays free and what goes in the paid plan. 3. Draft a founding-member offer: price, what's included and a deadline of [DATE] at 23:59 [TIME ZONE]. Nothing in it may become untrue after that date. 4. Fill in the upgrade email below in plain words, about what changes for the user.
Email 3. Send it only to users who've done the aha action.
Subject: Your founding-member price ends [DAY AND DATE] Hi [FIRST NAME], Before [PRODUCT], you were [THEIR OLD WAY, E.G. REBUILDING THE SAME REPORT BY HAND EVERY FRIDAY]. Since then you've [WHAT THEY'VE DONE IN IT] [NUMBER] times. I'm opening paid plans and I'd like you to be one of the first. As a founding member you get [WHAT CHANGES FOR THEM] for [FOUNDING PRICE] instead of [FULL PRICE], for as long as you stay subscribed. The founding price is open until 23:59 [TIME ZONE] on [DAY AND DATE]. After that it's [FULL PRICE] and it won't be extended. [ONE UPGRADE LINK] Not right for you yet? Reply and tell me why. [YOUR NAME]
Pick one day a week and run the same report, so you can see people moving from stalled to active and from active to paying instead of guessing.
Build me a weekly conversion report: [OPEN MY PROJECT IN CLAUDE CODE, OR ATTACH THIS WEEK'S EXPORT]. 1. Count users in each group: never started, stalled, active and paying. 2. Show how many moved since last week, above all stalled to active and active to paying. 3. Show the share of new signups who did [AHA ACTION] in their first session. 4. Show replies and upgrades from each email I sent this week. 5. Give me one table I can paste into a spreadsheet, last week next to this week. Don't change any data. Tell me which files you'd change before you change them.
Rather keep it by hand? One spreadsheet row a week works: the four group counts, upgrades and the one thing you changed.
Emailing all 500 a discount. It feels like action, but most of them have never had the value, so a cheaper price for something they don't use changes nothing. It also teaches the people who'd have paid full price to wait for a sale.
I've spent sixteen years building products and six years managing AI products. One Xeruit ad brought in 3,000 users in three days. That taught me a user list is only worth something once the people on it use the product. I've automated 90% of my startup with AI agents, so sorting an export and drafting these emails is the kind of job I give Claude. I'm vibe coding a dating app right now and this is the order I'll work through with its first users.
You've got five prompts and three emails. Export your users today, run Prompt 1 and send Email 1 to your ten most active people before you touch your pricing page.
What this won't do: it won't turn a product nobody needs into one people pay for. If your active users wouldn't be very disappointed to lose it, the work is in the product. Facebook's and Slack's numbers were found by looking back at who stayed, so pushing people to hit a number without the value won't make them stay. RevenueCat's figures come from mobile subscription apps. Claude can misread a messy export, so check its counts against users you know.
Your 500 users already signed up. Get them using it before you ask them to pay.
Related guides: How to get your first 100 customers, Are vibe-coded apps actually making money? and Meet Claude Code.
Sources, opened 30 October 2026: First Round Review, How Superhuman built an engine to find product-market fit (Rahul Vohra, 2018) and From 0 to $1B, Slack's founder shares their epic launch strategy (2015); The Growth Mind, The Ultimate Story of Growth, Part 1, quoting Chamath Palihapitiya; RevenueCat, subscription app trends and benchmarks 2026 and hard paywall vs freemium; Chameleon, benchmark report; Claude Help Center, create and edit files with Claude.
The Audacity to Build is my community for people doing exactly that. It is opening soon. Join the waitlist and you are first in when it does.
If this was useful, follow me. I share what I am actually building, the tools that survive my own testing, and the ones I quietly cancel. New plays every week.
Pop in your name and email and the guide is yours, along with the occasional new guide when I publish one. No spam, ever.