We’re the 10 November of 2022. I’m building 2 apps, but then I saw this tweet while at the gym:
I already edited old pictures for my mom using A.I. (she love genealogy and old stuffs), so I know how to do it and how to get the best quality.
So let’s go, goal of today: having the app ready by 19:00 (I have to go to the restaurant with my GF after that).
FInishing the session and then ready to start.
Ready to start. Finished my gym session, shower taken, gonna eat a very quick breakfast and then ready to work. Since I have no time, I’ll copy/paste the landing page design from MyAIPainting.com and change the copy. Gonna Typeform/Tally/Stripe somehow and fine a way to automate it later. No time to waste.
Let’s go! Gonna write each hour what I did.
Alright, landing page copy is ready (I’m quick at writing copies), and technical tests ready, tried with multiple pics my mom sent me and the results are really impressive.
Took lots of tweaking between the different models but I think I got a process that works well.
I won’t do automatic stuff at the beginning unless I have time to code it today, manual 24 hour delivery should be good enough to validate the concept I think.
Landing page is done!
I added pictures, made a logo, bought the domain and finalized a few more stuff. Gonna use Typeform to accept orders at first.
Gonna make the social media sharing picture and stuff now. It’s still pretty early so I guess I’ll try to automate it? Will probably take more than a day, as there are lots of things I need to learn for that. I prefer to do it manually at first to guarantee the best quality
Website published! https://photorestore.app
Posted the tweet and got an answer from the OP, that’s cool! I will probably post inside Facebook Groups and stuff.
For now I won’t automate it, as I want to manually handpick the quality. For the price, I put a random number, not too high but not pennies either as Stripe has flat fees of $0.30 with the %. I think $10 for 10 pictures is pretty fair, considering you get multiple variations of each.
Alright, focus of this afternoon: Making a “free” version of the app. People can register and make one picture per month for free or smth like that.
FIRST CUSTOMER!! YAY! 🥳🎉🎊🍾
The person I answered the comment just placed an order and added some suggestions 😍
Pictures are done, I sent him a .zip
I really hope he will be satisfied with the quality. If so, it officially mean I have a (micro)business!
I need to add:
- Google Auth
- Credit system
I keep getting this so I’ll need to find another tutorial
Alright, wasted yesterday day trying to make that oAuth shit work but it’s too complicated, I’ll just use my own login system. Really don’t understand why they make it so hard to use fking Google to login.
Alright, added auth, Stripe and the credit system. Gonna add more pricing options tomorrow.
Gonna do some polishing and adding more price options
People love “free” stuff (source), so I added free credits with each option.
Code is garbage but hey, for now it works!
K, gonna start marketing for the app soon so I will have to automate stuff a bit, that’s going to be the focus of this week. If it doesn’t works, I’ll have learned new skills anyway.
Things I want to do (and learn) by Friday:
- Automate the A.I. stuff: Shouldn’t be hard
- Improved user interface where user can see and download their pictures
- Uploading/Downloading photos: No idea how to do that, probably require a CDN? Very useful skill for later
- Google Authentication: Gonna try one more time to make that thing work with Passport 😤
Hopefully I’ll be able to make all that work BEFORE friday, so I can do some marketing and then publish my other side project this weekend!
Spent the whole afternoon trying to make AWS S3 work. Why is it so fking complicated, for real, it’s just uploading a picture, shouldn’t be that hard.
Used ImgKit + express-upload, took me less than 2 minutes to upload a file using Node.
Almost took me the whole day for that lol. Gonna hit the bed, wanna go gym tomorrow.
JUST MADE THE SECOND SALE!!! 🎉🥳
Working incredibly slowly today, I’m very tired somehow. Still managed to have the upload + UI working tho!
It’s not luxury but that will do it for now! If I get more sales I’ll improve the UI later. Just need to find a way to make people download from clicking the button (blocked by Chrome if the picture is not on the same domain 💀) and that will be good!
- User interface: ✅ Done
- CDN to upload pictures: ✅ Done
- Automate the A.I. pipeline: Tomorrow
- Google Auth: Friday (hopefully)
Gonna take a small break for today, gotta see friends for a farewell, and I’m super tired + headache.
Alright, took a few hours off, ready to start the day now!
Goal of today: Automate the pipeline.
Can’t believe it worked on the first try! App is now fully automated!! I need to improve the pipelines, but technically, it seems like it’s working, yay!
Lol, that actually was the easy part, managing the pipelines is melting my brain.
Finally got it working, damnit! Done for today, my brain is dead 🧠
Still need a few fix but tomorrow I should be able to work on Google Auth.
Google Auth will wait, gonna improve the user experience today instead, some things are confusing. I’m going to do my own version of the “Mom test”: If my mom can’t use the app without instructions, it means it’s bad.
Mhhm, it’s actually more complicated than I thought.
Since I have different pipelines based on the type of picture, I have to find a way to automate that too. I could ask customers, but asking them if their picture is “blurry” is too vague and they
Best bet might be to code my own model for that, but I don’t really have enough data yet.
JUST MADE A 3RD SALE!! 😍🎉
Not really profitable yet, but still really nice to get validation for the idea!
I also added a live chat to my website, that way if people have question/objections they can ask directly, might increase the conversion rate. The app is called Crisp, really easy to set up!
Seems like I’ve got the pipelines working. It output 4 different results made from multiple post-processing options, which should be enough to cover most use-cases I think.
Spent some time with GF, starting the day with a new sale (🥳🎉), a few messages to answer, and the app who passed the mom test (meaning my mom was able to use the app without any input from me!)
Focus for today (or what’s left of it): Improving the UX based on the feedback I had, publishing the automated pipeline and starting to plan the marketing!
Alright, was able to add a few tweaks to improve the User Experience, fix a few bugs with Stripe (was using the local URL instead of the real one 🙃). Need to run additional tests tomorrow when after that it should be all set.
I’m making a list of features divided like that:
- Essential: Stuff that are essential for the app to have (Delete photos, etc.)
- QOL: Quality of Life, or anything that improve the user experience but is not 100% required yet (downloading a .zip with all photos, password reset, etc.)
Alright, haven’t done as much as I planned, but I still managed to add the finitions! Technically the website is operational.
For the marketing, the first target I’ll try is genealogy.
Time to create my Learning Board:
It’s basically the “theoretical” version of the company. Gonna plan everything from there, write notes, write the result of my experiments, etc. It’s empty at the beginning, the goal is to fill it up as I learn things.
First part is “Product”. It includes stuffs related to the product like benefits, MVP etc. Just gonna write the vision/benefits here, since MVP is already done.
Benefits = Value for the customers
- See what your ancestors looks like
- Surprise your family with restored photos
- “Remember the good old days”
Probably the main angles. It’s late so I’m going to hit the bed, tomorrow will focus on customers / channels.
I love Sundays, people go on weekend so the coworking is empty, perfect to focus!
Today, going to work on customer avatars to make an hypothesis on who are the customers, and which feature/benefit is important for them.
Features are product-oriented, benefits are customer-oriented.
- Feature: Water is delivered frozen in big ice cubes.
- Benefit for someone living in a desert: Getting fresh water
- Benefit for someone living in the north pole: Building an igloo with it (lol)
The feature is the same, but the benefits depends on the customers.
Important things to write down: Before / After the product, What emotions do they buy, what are the possible objections, and where do they hangout.
Next part: Channels, the most important one
Basically listing every single channel you could grow the company:
I classify them by tier to see which one to focus on first based on 2 criterias:
- Which one can I access to RIGHT NOW
- Which one is the most likely to have my target audience
Main ones are probably Reddit, Q&A website and Paid ads.
First thing to do in any channel is to check the kind of content that works the best so I can make good content. Gonna start with Reddit and see the top content for each relevant subreddit.
Based on the research I’ve made, Reddit might not be the best place, as the subreddits I found are filled with hobbyist restorers who seems to hate A.I.
Might be worth a shot with ads tho.
Going to make a Reddit post to test the waters, and prepare the app for ads (Installing Tag Manager, various pixels etc..). I will also add a free trial to the app if people confirm their email, limited to 2 per person, that way they can try it before buying. (will split-test offering more/less later).
Preparing for the ads = Making sure people who never heard of the app or know me will be able to get enough trust and understanding to use it. Best way I found so far is to make a welcome page with a video:
Just saw this under one of my post so I added a French translation using WeGlot. Took me a few minutes to set up but it seems like it’s working!
Also added an upload preview:
Just woke up with a sale! 😍
Results wasn’t looking the best here, I need to make sure the guidelines are clear. I didn’t count it in the 10 credits. I just need to find a way to downscale/upscale pictures (too big = block the pipeline, too small=bad results) and I should be good to go for the automation!
Should be good for the upscaling.
Also added bulk upload
Not a productive day today, mainly because it’s waiting for user feedback. Going to start planning Twitter content as well as another side project.
Ended up not doing much lol. Anyway, tonight is movie night so, time to finish the day!
Fixed a few bugs and made some adjustments based on the feedback I had. Waiting to fix all majors bugs then I’ll start publishing the app. Took way longer than I thought, because it depends on other people
Since I have to wait for feedback atm I’m going to work on a side project in the meantime, no time to waste!
I just made a new sale! I’m now at $39.78, that’s so cool! I’m incredibly grateful to everyone who supported me so far 🥰