Revenue and Monthly Recurring Revenue are down significantly since we peaked in February 2026.
Revenue is down 24% since February 2026.
Monthly Recurring Revenue is down 12% since February 2026.
This has happened to me before with an application I made called Girlfriend Plus. In 2017 revenue spiked, but I didn’t do much development on the app. Pretty quickly revenue dropped off. Today it basically makes no money.
Why is this happening?
It’s probably because more people are cancelling subscriptions and fewer people are buying subscriptions. Let’s take a look at the data.
| Month | MRR Growth ($HKD) |
|---|---|
| January 2026 | $9,585 |
| February 2026 | $10,074 |
| March 2026 | $228 |
| April 2026 | -$7,890 |
| May 2026 | -$15 |
| June 2026 | -$5,708 |
| July 2026 | -$19,960 |
| August 2026 | -$5,320 |
Okay it looks like MRR growth has turned negative since April this year. But why?
| Month | New Subscribers |
|---|---|
| January 2026 | 191 |
| February 2026 | 184 |
| March 2026 | 173 |
| April 2026 | 157 |
| May 2026 | 128 |
| June 2026 | 95 |
| July 2026 | 78 |
| August 2026 | 45 |
The rate of cancellation is about the same, however the amount of new subscribers per month has been dropping since February 2026.
Why are we getting fewer new subscribers? I don’t really know, but here are some possibilities.
I stopped building in public. I haven’t posted a revenue graph on Twitter since September 2025. Sharing revenue data is a good way to get attention on Twitter, but for this business the Twitter attention doesn’t really translate into sales. I stopped sharing revenue data because a lot of people started using AI to build their own bank statement converter applications.
People are using AI to convert their bank statements. This works pretty well if you only have a few pages of bank statements to convert. If you have hundreds of pages, then you’ll need to pay for AI to do the conversion. AI chat bots can convert PDF to Excel files, but they’re slower and more expensive. I have a feeling a lot of users are using free tier AI chatbots to do their conversions.
One of my competitors is much better. I have a lot of competitors, because I built this application in public. Perhaps one of them is doing a much better job than I am and is taking away all my customers.
What am I going to do about it?
I’m going to keep working on it. Last time my business failed I think it was because I didn’t do anything about it.
In 2015, I read Founders At Work and was really impressed by the section featuring Evan Williams and Blogger.com. I don’t totally remember what happened, but I think they raised money to start Blogger, ran out of money. Everyone left or was fired and it got down to just Evan working by himself. Somehow he was able to turn it around and eventually he sold Blogger to Google.
Analysing cancellation reasons
When a user cancels a subscription Stripe asks them for a reason. At the moment I don’t really do anything with the cancellation reason, actually this is the first time I’ve ever looked at the data. I’m going to start sending follow up emails after a user cancels, depending on their cancellation reason. For example if a user cancels because they switched service, I can ask them which service they switched to, and what was it about that service they preferred.
Do some marketing
My favourite type of marketing is content marketing. All forms of marketing have a cost, but content marketing is the only form of marketing that can last forever. I don’t really like spending money on cost-per-click advertising platforms because the prices are really high and you only benefit from it while you’re paying for ads.
I want to try out some physical real world marketing channels like tram and bus ads in Hong Kong. I also feel like podcast and YouTube video advertising can be good. The benefit is you pay once and the video lasts forever. So if the video you advertise on is evergreen, then your ad should be as well. I’ve contacted a few podcasters and YouTubers but so far no one has replied.
Talk to users
I still get a lot of customer support messages from users. I’ll keep doing the support and trying to think of features based on the problems they’re facing. The other day a user thought a scanned document was failing to process because it said “Processing” for a very long time. That gave me two ideas. Speed up the processing for scanned documents. Show progress information for scanned documents. “Processing page 22/99”.
Get a job?
Hell no. Last time I worked at a company was November 2020. I love programming, but I really do not enjoy programming at a company. Things were bad in 2020, but I feel like they’re worse now. My friends tell me the job is basically managing AI agents and reading Pull Requests. That sounds like hell, and I’ll try to avoid it for as long as possible.