By Max Sherman · September 1, 2023
I launched AddictingTrivia.com!
I launched AddictingTrivia.com, and became a Twitch Affiliate!
Grade: F
I deprioritized MinutesGenerator and as a result didn't invest time into pivoting to find a niche for it. More on this below on the last goal.
Grade: B
I watched several Ahrefs youtube videos about how to do content marketing and use their tool to identify keywords with low competition. I would say I learned just enough to be dangerous.
I won't give myself a higher grade here because it does seem like I could have done more, but frankly I think this goal isn't well defined because it's hardly measurable. I think I achieved what was in the intended spirit here though.
Grade: A
I paid for an Ahrefs subscription and used my learnings from the Ahrefs tutorials to identify an SEO niche for my next business, AddictingTrivia. This involved spending 12 hours across two days typing words into their tool until I found what I think is the right balance of low competition keywords, and high traffic.
This process was slow, but educational and I feel much more familiar with the tool. I'll classify it as one of these one-time-cost-asset things because I intend to use this tool for every subsequent business idea I have.
This goal gets an A because I picked SEO, and if things pan out as promised, then I can build a website that has a high likely hood of aquiring traffic from the get go. This dovetails nicely with the maxim that "entrepreneurs should build something that people want". People search for things they want, so basing a business idea off search query data checks this box.
Grade: A
As I said above, I'm putting MinutesGenerator on the backburner. There's a chance I'll come back and try to pivot this into a niche, but for now I think my efforts are better spent elsewhere, rather than trying to squeeze blood from a stone in a saturated market.
There are many competitors in the audio to text transcription space, and many of them are VC funded, and pay for ads, and teams of engineers and designers. I'm not optimistic about a head on competition here, and I'd prefer not to go on a suicide mission trying to win a David and Goliath style battle.
One of the luxuries that I can enjoy as a bootstrapped founder is that I don't need to make a billion dollar company, and I don't need to make commitments to investors about what my business is going to do. I can pivot as much as I want (and I plan to) in order to solve the optimization problem of finding the easiest, most sustainable way to build a business that pays me enough money to live.
I'll leave MinutesGenerator up and running for as long as I can, but for now I am on to bigger and better things!
The basic version of the plan is to acquire traffic via google search. There is a volume of about 400K search queries per month for keywords that are supposedly easy to rank for, that are also some form of "quiz" or "trivia", for example "minecraft quiz".
I'm going to focus on multiple choice quizzes because they are the simplest to generate, but I like the idea of expanding to other quiz types eventually. I am building a system that uses GPT to generate quizzes given a topic, and I intent to have this system be fully automated end to end, starting with a root keyword like "minecraft" and ending by uploading quiz data to my DB, pushing the website, and requesting indexing from google.
My goal is to generate 1000 quizzes using this system, because I think this will lend legitimacy to my website. I believe that nobody will take seriously a quiz website with only a dozen quizzes. This will also cast a wider net in SEO-space, allowing my site to acquire more traffic.
Eventually I will put Adsense ads on the site to monetize it.
The two biggest things to figure out for this approach are backlink acquisition and quality control. Backlinks are hard to get, and I don't know how systematized I can really make this process. I'm punting on it for now, because I think it will be an easier sell to ask for a link to a website with thousands of quizzes rather than a dozen.
Ahrefs comes to the rescue here too, because I can look at my competitor websites and see who is linking to them as well. This gives a reasonable place to start searching for backlinks, but ultimately it seems like a manual process.
The second thing to figure out is quality control for quizzes. GPT hallucinates wrong answers and off-topic questions decently often, which leads to poor quiz quality. My rough idea here is to use explicit and implicit feedback from website visitors to dynamically make adjustments to quizzes. For example, if a quiz has a question that 100% of users get wrong, that is an indicator of the question being bad. For explicit feedback, probably a thumbs up/down button asking to rate the quiz would suffice.
I had a lot of fun doing a costream with BadCop and it led to a few followers on twitch and twitter which was nice. I spend a lot of time streaming and it's rewarding to make more and deeper connections in that space. I definitely would like to do more costreams.