29
38 Comments

Creating an AI-Powered Logo Creator and Earning $70k/mo

Hi there! Who are you, and what are you working on?

My name is Dawson Whitfield. I recently launched Logojoy — an online logo maker that uses machine learning to make it feel like you're working with a real designer.

Logojoy is for anyone trying to start a business, blog, non-profit, club, or anything that requires a logo. Typically, a logo is the first step for people starting anything, and we're trying to make it easy for these folks to get a simple, good looking logo that they can use on their website, business cards, etc.

I think people choose Logojoy over the 10+ competing logo makers because it's not just a dumbed down Photoshop. We understand your tastes and needs just like a designer would, and we present you with mockups that get you thinking. Even if those designs aren't perfect, we make it simple and enjoyable to see variations and make changes.

I would segment the product into two parts: the logo generator, which takes input and generates logos; and the editing interface, which allows users to see variations and edit things like font, colors, etc. The logo generating algorithm is still in its infancy, so right now it's mostly acting as an idea source for most users. The editing interface is actually pretty robust, my favourite thing about it is that we present options as variations. For example, if you want to see different fonts, we don't give you a drop-down list of fonts — instead, we show you versions of your logo, each with a different font, so it's very tangible.

Designing Logo

An AI-powered logo maker is a fantastic idea. How did you come up with it? And what motivated you to get started?

I came up with the idea for Logojoy when I was designing a logo for a client. I've been a designer for 12 years or so, and I've probably made 150 logos for clients. I would get frustrated designing logos for small businesses because (a) it was so time-consuming to create 30 mockups, (b) it would take weeks to do those small back-and-forths, and (c) the logo would end up being so simple that I felt like that entire process was a waste.

As soon as I had the idea, and couldn't find anyone doing it right, I started working on it.

The main thing that motivated me was just building something cool. I did a quick Google Trends lookup and saw "logo maker" was searched 3x more than "website builder" — which was a big source of comfort during the development process. Whenever I was unmotivated, I would open that tab. I must have opened it 200 times.

Funny enough, another big motivator was how poor the current solutions were (in my opinion). It got me so excited to think that I could offer a product that was 10x better than the next best product, especially because the market was so big.

And lastly, what got me really excited to build a small product that actually generated revenue was Indie Hackers, which I've been following since your debut on Product Hunt :)

Logojoy uses artificial intelligence for its designs. How does it work, exactly? What tech are you using? How long did it take you to build?

This is the fun part. Our logos are just combinations of ingredients — fonts, colors, layouts, symbols, etc. Logojoy uses machine learning to learn which ingredients go better together.

It starts with tracking basically everything our users do. We track things like: the inspiration they selected, the logos they favorite, the changes they make to logos (e.g. changing the font), the logos they purchase, and more. We currently track about 80 types of actions.

Every day, the learning algorithm reads all these actions and weights each one by how many times it occurred. Because of the structured way it reads actions, it's able to define rules based on "heavy" actions.

For example, an action might be defined as "user changed font weight from 100 (light) to 600 (bold)". This action object includes the number of times it occurred, every other preceding action, and all of the logos ingredients. In this case, let's say the algorithm concluded that every time this happened, the color of the logo was classified as "light". The algorithm would presumably define a rule that says "if the color of the logo is light, do not use a light font weight".

Making V1 took about two and a half months. The most challenging this was definitely creating the structure for the system to define its own rules (in other words, the machine learning). My stack is old-school: PHP, MySQL, jQuery.

Finished Logo on Packaging

Where'd you get the money to bootstrap Logojoy?

I financed Logojoy with savings and a little bit of freelancing.

You're up to $15,000/mo in revenue in less than a year. How exactly does your business model work? And what have you done to grow so fast?

We allow users to build unlimited logos for free. Once they're happy with their logo, they pay $20 for a basic version, or more commonly, $65 for the premium package with includes vectors.

We honestly just built a good product. The key to logo design is presentation. Some people pay $65 for an extremely simple plain-text logo solely because we present the logo in a beautiful way.

We launched on Product Hunt and told everyone in our social networks about it. I would highly recommend getting someone with posting power to post your product on PH on Tuesday — it made all the difference.

Since then, our traffic has been almost 100% word-of-mouth. What really surprised me is how quickly the news spread across the world. Within a day, Japan was our biggest market (with 25% of all our traffic). I think we have a tendency to see the US as the whole pie, when if you're creating something that isn't restricted by language, your market is 10x bigger.

We're starting to experiment with Adwords — where we're paying about $0.20 per click. We're spending about $250/day and that brings about 1,800 people to the site every day.

We're seeing about 5,000 people come to the site each day. 25% of all visitors complete the logo-making process and sign up. Of the people that signup, about 0.5% purchase a logo. We're hoping to raise that up to 1-2% within the next month.

The day Logojoy launched, it made $450 USD and it's only been increasing since. In February 2017, Logojoy made nearly $70,000:

Daily Revenue in February 2017

What are your goals for the future? And what are the biggest challenges that lie ahead of you?

I think eventually Logojoy could empower entrepreneurs to design anything from a logo to an ad. I'm really excited now that I can invest all of my time into Logojoy and really make it a world-class product that people LOVE to use.

One thing that will be an ongoing challenge is instilling trust in our users and positioning Logojoy on the same level as designers. Even if Logojoy is eventually able to produce results that are as good as a designer would, people still won't value it as much.

Finished Logo on Sweatshirt

What do you think your biggest advantages have been so far? If you could go back in time, is there anything you'd do differently?

I think my ability to both design and develop has been crucial. When you're able to go from vision to finished product by yourself, you can move and pivot so quickly. I think that my most powerful advantage is the confidence I have to take on any problem. Once you truly know that every problem can be solved by breaking it down into smaller pieces, nothing can stop you.

If I could go back in time, I would definitely get beta users earlier. Once people started to use the site, I was so much more motivated to finish features and fix bugs.

What's your advice for aspiring indie hackers?

What really worked for me is building a small product. The big picture I have for Logojoy will take years. I was able to get really excited about it because I decided on the simplest version of it. Spend time deciding on the most critical features to launch with, and only build those.

For me, it's all about staying motivated. It's hard to stay motivated sometimes, and the two things that helped me the most are a) being close to launching, or starting with the simplest version of what you want to build, and b) getting feedback from users.

Where can we go to learn more about you and Logojoy?

  1. 1

    Wow! This was all the way back in 2016

    There's a new logo creator with AI coming out every week now haha

  2. 1

    Just came to say I bought a logo from your site and am very happy with it.

  3. 1

    Wow, the idea is great and the page is awesome, my first thought was to copy the idea in my country but... men! theres a lot of competitors with really good websites, even android/ios apps for this niche.

  4. 1

    How can you come up with machine learning / ai at first? As a designer, it is not intuitive to adopt such leading technology to improve the process.

    It can be intimidating to most designers without the technical background. This will make them filter the concept very soon.

  5. 1

    Hi, it's a very cool product. How did you create the AI/NN. Can you give us a technical scoop?

  6. 1

    Amazing concept! It's really fun and super easy to use. My logo turned out great for such an affordable price

  7. 1

    what about the license on the premium fonts you're using? does it cost very much?

    1. 1

      Hi Dawson, Could you please answer this. because I am building an application which helps end-user to create facebook cover image wherein I need to have many fonts as you. Did you buy these fonts separately or as a single package have you bought all these fonts?

  8. 1

    Hi Dawson - thanks so much for sharing your story. I created a product in the website builder space and we have a heck of a time finding affordable adwords. Any tips on how you found .20c keywords?

  9. 1

    This is awesome. As a programmer I need to know what you use to generate those mockups? Any specific library?

    1. 1

      To me it looks just like a algorithm that creates combinations from the option templates you have chosen.

  10. 1

    Hi Dawson, how did you manage to legally use the icons from thenounproject.com? Something about your project doesn't fit on what is written here http://api.thenounproject.com/getting_started.html#unacceptable-uses

    1. 1

      why would you be unaware of something they offer you're literally asking for that you could find in two seconds? curious accusation..

    2. 1

      We use their Pro plan, just like Squarespace Logo, Tailor Brands, and others.

  11. 1

    Hi Dawson, your project is a nice inspiration. I'm wondering about your Machine Learning stack, which tools do you use? I'm a machine learning student having really good time learning about those things

    1. 1

      Hi @dawson can you answer this please.?

  12. 1

    Congrat on your project!

  13. 1

    For a one man operation, you mention "we" a lot, I guess that is unintentional ?

    1. 1

      Company names are collective nouns. When referring to Logojoy, "we", is appropriate, despite the fact that it may or may not be a 1 man op.

      1. 1

        Cool! So I HAVE been grammatically correct all this time.

        1. 1

          Yeah, its a good way of trying to represent the company as something large that can handle any new business, even though you might only have like 3 employees. It's actually a common tactic if you're having a potential client come in to rent an office and pay actors to make your company look big.

  14. 1

    There were a few comments made on HN regarding how the $15k/mo figure was extracted from just a few days since launch. Was wondering whether you expect this to gradually slow down post "launch effect" or if adwords can keep this stat stable/growing.

    1. 1

      Very good point. Adwords can 100% keep this stable, and we extrapolated this w/o the HN traffic. Extrapolated with the HN traffic, it would be about $55k :O

      1. 1

        Very very very awesome - Best of luck! Would be cool to see if you could somehow turn this into a recurring rev model.

  15. 1

    I just played around on the site for a little while, that was fun! Showing what the logo would look on business cards/building addresses etc. was a great idea.

  16. 1

    Amazing inspiration. Thanks for the transparency!

  17. 1

    You mention 2.5 months to the initial version... How long did this take from beginning to launching on PH?

    1. 1

      2.6 months ;) Started working Sept 1st, launched beta Nov 15, PH on Nov 22.

      1. 1

        Interesting. The average time-to-launch for people that I've interviewed for Indie Hackers is 10-11 weeks, so you fit right in there!

        1. 1

          That's shorter than I would have thought!

  18. 1

    This is remarkable Dawson. It's really inspiring to see how making a great product and that initial traffic burst has carried you through to such success so quickly. Thanks for sharing your story, and congratulations on your progress so far.

    1. 1

      Thank you!! It's cool to think how I was being inspired by people on Indie Hackers a month ago. I'll still continue to be inspired by other folks on this site, but cool to think about.

  19. 1

    The "tour" link in your website is broken. Great Idea by the way. All the best.

  20. 1

    This comment was deleted 5 years ago

  21. 1

    This comment was deleted 5 years ago

  22. 1

    This comment was deleted 5 years ago