My Story:
Inspired by this post
As a software engineer, I always wanted to launch my own business. Initially, Saas (software as a service) was the only thing on my mind. Given that my main skill was knowing how to code, I thought this was the only thing that made sense. This limiting idea led me down a long path of building a project, failing, and moving to the next.
Over the last few years I’ve tried to build a marketplace (basically an OfferUp clone), a machine learning stock trading tool, an AI dev tool, a sentiment analysis software, etc. As you can probably predict, none of these worked out which left me in a perpetual cycle of building something, and moving on. The plus side of this cycle is that I got to develop a lot of cool things, each providing its own set of unique technical challenges. I definitely learned a ton and became a better software engineer. Unfortunately, my goal isn’t to become a better software engineer, it is to become a business owner.
Changing My Mindset
This whole time, I had been considering a saas as my only option. This is likely due to all the business advice and podcasts preaching the message, “do what you're good at” or “learn a skill and build a business around that skill.” This advice is probably good. But for me, it hadn’t worked out at all. Admittedly, I think some of my ideas may have worked, it was just me lacking the knowledge and discipline to see it through. Anyhow, one day I was drinking apple cider sitting in front of my computer (probably watching another business podcast) when a thought came to me; “What if I don’t build a SaaS at all?”
I hinted at this thought earlier in this post so it may not seem super profound, but in the moment I really shocked myself. I was now allowing my mind to explore every type of business idea possible, which overwhelmed me so I decided to go to bed and sleep on it.
The Idea
The next morning while thinking about what type of business to start, I remembered a huge frustration I had from one of my previous side projects. A brief summary of this side project: After AI Image technology like Dalle2 and Midjourney became readily available, I saw it as an opportunity for Machine Learning models. ML projects that deal with images and classification need a huge amount of images to train on. So I began working on a software that would be able to generate hundreds or thousands of images for any ML task. The major problem was the distortions. These annoying distortions in faces, hands, and objects made it unreliable for machine learning models to train on, forcing me to ditch the project.
I remember wishing I could just bundle these images up and send them to an image editor to get them quickly edited. The issue was that it was too costly, and finding an image editor online that could produce extremely high quality images in a short amount of time was difficult for me. Many quotes were out of my budget as they would charge around $25 per image (I wanted to edit 1,000+ images!). Additionally, the quality of images I got from the freelancers I worked with was not good enough for ML models to actually train on.
This is when I came up with the idea to create a service specifically designed for AI image editing in bulk. Essentially the plan was for businesses to come to me with a large amount of AI generated images, and I edit them making them ready for practical business applications within a week. A couple days after thinking it through, I officially launched refinebase.com
The Daily Grind
My Routine:
8:30am: Wake up & get ready
9:00am-5:00 pm: Day job as a software engineer
5:30-6:30 pm: Workout
7:00-8:30 pm: Dinner and relax
9:00pm-1 am: Work on RefineBase
Validating The Idea
As this AI image technology is still fairly new, it's hard to know how many people are thinking about using it at scale for their business. I needed to know if people wanted to use AI images for professional business applications, or just to play around and create cool images. I began asking around in a few forums and got some responses but nothing too promising. As I didn’t know who my target audience was, I just took a bet and geared my website towards businesses who need very high quality and professional images such as creative agencies or magazine publishers. I figured these types of businesses can’t afford to have imperfections and distortions in their images as it would ruin their brand. To be honest, I still don’t have this 100% figured out, but since this is a high ticket service I only needed a few clients to get an idea and achieve a decent MRR.
Getting to $15k MRR
This stage was not easy at all. After doing all the setup work for RefineBase, it was time to actually get customers. I am very bad at marketing so I had no idea where to start. Warm outreach wouldn’t work for me because I have no one in my direct network who is generating thousands of AI images. Paid ads didn’t seem smart yet as I didn’t know who I would target and I didn’t even have my first customer. And cold outreach was an option but again, not having a clear target audience made it difficult to build an efficient cold outreach system.
It was time to get creative.
When trying to get my first client I thought to myself, “if someone needed image editing services, where would they go?” My first answer to that question was freelancing sites such as Upwork and Fiverr. I didn’t want to make a listing and just be another freelancer, but I needed a way to access these people coming for image editing services.
This is when I had an aha moment. I realized that if I directly message the image editors on these platforms, I can offer them a commission for every referral they bring to my website. The idea was that if an image editor couldn't take on a requested task due to capacity issues, they would just direct the client to RefineBase. I would give them a discount code to make it a win-win, and so that I could track their referral numbers(For those wondering, I offered a 20% commission per referral). At first, I had no success with this strategy. My theory is that many people thought it was a scam, or just thought it wasn’t real. There were also many issues with the freelance platforms themselves as I would get warnings for violating their terms of service. After a lot of struggle, I refined my pitch to these freelancers to create a trusting conversation between us, as well as to keep in line with the terms of service.
Using this method I was able to get 3 clients, bringing me to $15k MRR.
What’s Next
Keep in mind that I still have a ton of work to do in order to grow the business and improve my profit margins. 3 clients is fairly little, and I think I can grow that number with some good marketing. I still face the problem of not knowing my target audience, as all 3 clients are in completely different industries.
The next immediate steps for me are to figure out my target audience, grow my socials, improve my profit margin, and figure out a good marketing strategy to start scaling. As I still have a lot of work to do, any feedback on the site or business model will be greatly appreciated.
I left out a lot of details in this post such as how I actually fulfill the service, so comment below if you want to hear more!
If you made it this far and are interested in the service, then head over to refinebase.com. If you know anyone who would benefit from RefineBase then shout us out, or email me at hello@refinebase.com and we can work on an affiliate deal. Thanks!
Inspired by this post
As a software engineer, I always wanted to launch my own business. Initially, Saas (software as a service) was the only thing on my mind. Given that my main skill was knowing how to code, I thought this was the only thing that made sense. This limiting idea led me down a long path of building a project, failing, and moving to the next.
Over the last few years I’ve tried to build a marketplace (basically an OfferUp clone), a machine learning stock trading tool, an AI dev tool, a sentiment analysis software, etc. As you can probably predict, none of these worked out which left me in a perpetual cycle of building something, and moving on. The plus side of this cycle is that I got to develop a lot of cool things, each providing its own set of unique technical challenges. I definitely learned a ton and became a better software engineer. Unfortunately, my goal isn’t to become a better software engineer, it is to become a business owner.
Changing My Mindset
This whole time, I had been considering a saas as my only option. This is likely due to all the business advice and podcasts preaching the message, “do what you're good at” or “learn a skill and build a business around that skill.” This advice is probably good. But for me, it hadn’t worked out at all. Admittedly, I think some of my ideas may have worked, it was just me lacking the knowledge and discipline to see it through. Anyhow, one day I was drinking apple cider sitting in front of my computer (probably watching another business podcast) when a thought came to me; “What if I don’t build a SaaS at all?”
I hinted at this thought earlier in this post so it may not seem super profound, but in the moment I really shocked myself. I was now allowing my mind to explore every type of business idea possible, which overwhelmed me so I decided to go to bed and sleep on it.
The Idea
The next morning while thinking about what type of business to start, I remembered a huge frustration I had from one of my previous side projects. A brief summary of this side project: After AI Image technology like Dalle2 and Midjourney became readily available, I saw it as an opportunity for Machine Learning models. ML projects that deal with images and classification need a huge amount of images to train on. So I began working on a software that would be able to generate hundreds or thousands of images for any ML task. The major problem was the distortions. These annoying distortions in faces, hands, and objects made it unreliable for machine learning models to train on, forcing me to ditch the project.
I remember wishing I could just bundle these images up and send them to an image editor to get them quickly edited. The issue was that it was too costly, and finding an image editor online that could produce extremely high quality images in a short amount of time was difficult for me. Many quotes were out of my budget as they would charge around $25 per image (I wanted to edit 1,000+ images!). Additionally, the quality of images I got from the freelancers I worked with was not good enough for ML models to actually train on.
This is when I came up with the idea to create a service specifically designed for AI image editing in bulk. Essentially the plan was for businesses to come to me with a large amount of AI generated images, and I edit them making them ready for practical business applications within a week. A couple days after thinking it through, I officially launched refinebase.com
The Daily Grind
My Routine:
8:30am: Wake up & get ready
9:00am-5:00 pm: Day job as a software engineer
5:30-6:30 pm: Workout
7:00-8:30 pm: Dinner and relax
9:00pm-1 am: Work on RefineBase
Validating The Idea
As this AI image technology is still fairly new, it's hard to know how many people are thinking about using it at scale for their business. I needed to know if people wanted to use AI images for professional business applications, or just to play around and create cool images. I began asking around in a few forums and got some responses but nothing too promising. As I didn’t know who my target audience was, I just took a bet and geared my website towards businesses who need very high quality and professional images such as creative agencies or magazine publishers. I figured these types of businesses can’t afford to have imperfections and distortions in their images as it would ruin their brand. To be honest, I still don’t have this 100% figured out, but since this is a high ticket service I only needed a few clients to get an idea and achieve a decent MRR.
Getting to $15k MRR
This stage was not easy at all. After doing all the setup work for RefineBase, it was time to actually get customers. I am very bad at marketing so I had no idea where to start. Warm outreach wouldn’t work for me because I have no one in my direct network who is generating thousands of AI images. Paid ads didn’t seem smart yet as I didn’t know who I would target and I didn’t even have my first customer. And cold outreach was an option but again, not having a clear target audience made it difficult to build an efficient cold outreach system.
It was time to get creative.
When trying to get my first client I thought to myself, “if someone needed image editing services, where would they go?” My first answer to that question was freelancing sites such as Upwork and Fiverr. I didn’t want to make a listing and just be another freelancer, but I needed a way to access these people coming for image editing services.
This is when I had an aha moment. I realized that if I directly message the image editors on these platforms, I can offer them a commission for every referral they bring to my website. The idea was that if an image editor couldn't take on a requested task due to capacity issues, they would just direct the client to RefineBase. I would give them a discount code to make it a win-win, and so that I could track their referral numbers(For those wondering, I offered a 20% commission per referral). At first, I had no success with this strategy. My theory is that many people thought it was a scam, or just thought it wasn’t real. There were also many issues with the freelance platforms themselves as I would get warnings for violating their terms of service. After a lot of struggle, I refined my pitch to these freelancers to create a trusting conversation between us, as well as to keep in line with the terms of service.
Using this method I was able to get 3 clients, bringing me to $15k MRR.
What’s Next
Keep in mind that I still have a ton of work to do in order to grow the business and improve my profit margins. 3 clients is fairly little, and I think I can grow that number with some good marketing. I still face the problem of not knowing my target audience, as all 3 clients are in completely different industries.
The next immediate steps for me are to figure out my target audience, grow my socials, improve my profit margin, and figure out a good marketing strategy to start scaling. As I still have a lot of work to do, any feedback on the site or business model will be greatly appreciated.
I left out a lot of details in this post such as how I actually fulfill the service, so comment below if you want to hear more!
If you made it this far and are interested in the service, then head over to refinebase.com. If you know anyone who would benefit from RefineBase then shout us out, or email me at hello@refinebase.com and we can work on an affiliate deal. Thanks!
