Tarpit Ideas - A Personal Experience

Tarpit Ideas - A Personal Experience

Why do dumb ideas so often look great for amateurs?

Published on Thursday, April 11, 2024

In the context of a startup, a tarpit idea refers to concepts that, on the surface, look great but, after deeper exploration, are revealed to be problematic, difficult to create, and unprofitable.

A few characteristics of tarpit ideas:

  • Seemingly Attractive - Tarpit ideas look great at first glance. A large unexplored market that solves "everyone's need."
  • Seemingly Uncomplex—The true complexity is often hidden. This can be due to technological challenges, market acceptance, laws and regulations, etc. The further you go, the harder it becomes to make the system.
  • Resource-Intensive—Tarpit ideas can sometimes require significant resources, mostly in terms of people and time and sometimes in capital.
  • Difficult to Pivot from—Tarpit ideas are often difficult or even impossible to pivot from. Finding the next right niche when you are too deep is very hard.
  • Warning Signs Ignored - It is so easy, especially for inexperienced founders, not to see the early warning signs due to optimism bias and the sunk cost fallacy. "We are almost there, just a bit longer!"

A great video on the topic:

Personal Experience (and lessons not learned)

Almost 20 years ago, during my college days (and a few years after), I wanted to create:

An Online database of biographies with queryable individual facts, wiki data management, user collaboration, and automated knowledge extraction from text. A news aggregator (from a predefined list of online publications) for each person's data from services like Wikipedia pages, Twitter Accounts, etc.; API for personal details, message board, and media.

Basically, Web 2.0 combination of Wikipedia and IMDB. You would get a large database of famous people with many facts organized schematically. I mean, who wouldn't want to find out what Isaac Newton and Blackbeard have in common!?

  • Seemingly Attractive—There are so many people in the world! We would get so much traffic if we could get data from tens of thousands of famous people! Even if it's just a few views a day per person, it would still be so much traffic!
    • Of course, the reality was much different. Ironically, we got "a lot" of traffic from Google, but it was just for a few people. The second problem is that of those few people, most of them, were porn stars. And (Google) AdSense does not like safe-for-work content that is related to porn. We managed pictures carefully, had manual checks, and all images were SFW. However, Google still marked them as violations of AdSense TOS.
  • Seemingly Uncomplex — Technically, it should be simple. You have a Person table, and that's it?
    • We created a schema-less, queryable, document-oriented database on top of the relational database. Our persistence was only matched by our stupidity.
  • Resource Intensive—Of course, it takes time to develop "custom database weirdness," write scrappers, write biographies, check all the pictures, etc. But we were broke students living with parents, and time was the one thing we had enough of. But we did not have the capital, and the cost of servers was piling up.
  • Difficult to Pivot from - Where could we even pivot to?
    • Porn star database!
      • That could have worked. 💸💸💸
  • Warning Signs Ignored - 🤣
    • This is Fine" Meme Analysis | Medium
    • We have ignored so many things that The Lean Startup book looks more like a list of our mistakes.

The worst part about Tarpit Ideas is that I am still tempted to do this again. Only next time, I will succeed!