Week 82 - Referral Feature
Posted on by Angus Cheng
Last Week’s Goals
Plumpbird: Set some harder goals this week you lazy jerk!
✅ 1. Contact 100 users
I contacted more than 100 users last week and scheduled the emails to arrive at 10 AM in whatever timezone the user was in. Sadly only four of the users replied. Some users I emailed had signed up several months ago. It might be better to email users who recently signed up.
✅ 2. Create a YouTube video
I created two videos last week
This one shows how to extract transaction data from PDFs
This one is a video version of last week’s blog post.
✅ 3. Go to the Office Five Times
I really did not enjoy doing this. It might be because I don’t like my office. I might rent a desk somewhere else.
Used FastMail’s API
I don’t like it when companies send me a lot of emails. To follow the Golden Rule I send at most one cold email to my users. To achieve this, I would search my mailbox before sending out an unsolicited email. If any results existed, I wouldn’t email them. This process takes up some time when you’re sending out 100 emails. To save some time, I used FastMail’s API to do this for me.
Now I can run queries like “Get me a list of users who signed up in the last week, that have not been emailed before”.
- Added the year to transaction dates for PNC Bank
- Fixed a missing data issue for Standard Bank
- Fixed duplicated data issue for People’s United Bank
- Fixed a missing data issue for Community Bank
I’ve noticed that some users will never pay for a subscription. Instead of paying, perhaps I could let them earn credits by referring their friends, colleagues and enemies.
I worked on this feature all of Friday and Saturday. I got it into production on Saturday afternoon. Since then how many users have been referred? Zero.
Difficult Edge Case
This looks like a pretty easy table to deal with. However, it contains two problems. The first is the “Earning Credit Rate” is not a transaction. No problem we can configure our parser to ignore rows with the description “Earning Credit Rate”.
The second problem is in the “Earnings Credit” row. When you come across a table that is labeled “Service Charges” you might expect all the records in that table to be debits. TD Bank doesn’t agree with you though. They include Credits records. How do we deal with this? I’m still thinking about a good way to do it. Perhaps some post extraction step that forces a record to be a debit/credit based on its description.
This Week’s Goals
1. Contact 100 users
Last week’s email campaign was a bit of a bust. This week I’ll email 100 users who signed up in the previous week.
2. Cloud Accounting Software Research
I spent some time figuring out how much work it would take to integrate with Xero. I realised it would take a very large amount of work, and I wasn’t sure if it would be worth it. I’ll look at the top ten cloud accounting software packages and figure out which one is the easiest to integrate with. Afterwards I could even create a blog post titled “Which Cloud Accounting Packages are Easiest for Developers”.
3. Get that API customer
I’ve exchanged a few emails with someone who wants to use Bank Statement Converter via an API. The goal is to figure out what exactly they want from the API and get some sample documents from them.
4. Share for credits
You can refer friends, but there’s no way to earn credits by sharing Bank Statement Converter on LinkedIn.
5. Style this Blog
I don’t like the way this blog looks. There’s a few things I want to change. Perhaps I’ll try hiring someone on Upwork for this.
6. YouTube Video
I spent some time working on a feature that I realised was a waste of time. I’d like to create a YouTube video talking about this dumb feature.
SEE YOU NEXT WEEK