Book Suggestions and Information Technology- Startup Week 21
I added article suggestions a few months ago, but there was something missing. It wasn’t quite appealing to the current users (not to mention the devastation I felt after the development).
This week I’ve decided to come back to it and add another suggestion mechanism.
While coming to the decision, I’ll also touch on the discussion between software and information theory, and how they can bring value to the end users.
In 2020, San Francisco startups almost become ubiquitous to mean software startup. Y Combinator accelerator won’t even accept you unless you have a technical cofounder.
But software people get blinded with fancy UI design or crave for efficiency. Before jumping into creating apps for your idea, we need to figure out what unique information your software brings unique values to people.
Today, I want to discuss how I came to decide on book suggestions from the perspective of unique information values.
Information Value
Software manipulates information, therefore it is called information technology.
Oxford University Professor, Luciano Floridi, described in a book The 4th Revolution that there are 4 types of information:
- Knowledge: Information you have access to.
- Insipience: Information you want to have but have no access to.
- Uncertainty: Information you don’t know you should have, but it’s knowable.
- Ignorance: Information you don’t know you should have, and it’s unknowable
The forth is impossible to talk about (Black Swan problem), so we will just leave it off.
Traditionally schools have been teaching us with the first assumption: knowledge. Teachers give you homework, and you know the answers are there. Upon graduation, the school will certify you with this minimum knowledge.
Google on the other hand operates on 2. It lets you find answers to your burning questions. What time is this restaurant opening? How to get to Highway 27 the fastest? But you still need to know what to search for.
Then the third is uncertainty. This is where tracking type software triumphs. It can surface the information you didn’t anticipate before, and form a new connection in knowledge which you were previously unaware of. This is where I’m going with Kaffae.
What is Suggestion?
A lot of startups especially those with backend engineering backgrounds want to build so-called perfect “suggestion” apps. But those who actually try implementing it find defining the quality of articles is close to impossible.
The best way to make sense of suggestions is suggesting it based on what? I’ve got this hint from some of the health tracking apps. For example, LifeSum is an app where you can keep track of your nutrient intake by recording food recipes. And then it provides tailored recipes based on your data.
LifeSum’s unique approach is not in the recipes they provide. After all, there are millions of quality recipes online. It’s the unique data it contains that can customize to your need.
Another approach is Netflix. Netflix boasts on its suggestion algorithm, but you’re essentially there for its content that’s not available elsewhere.
So an approach to this suggestion problem is one of two ways. One is to find a way to track unique information to base the suggestions off of. The second is even if you have the same trivial data to base off the suggestions off of, but provide unique data.
Suggestion That Connects the Dot
I do not have the resource to create massive content to create. Nor Kaffae has access to unique information to tailer the algorithm (that’d make it any better than what Google does).
So instead, I could mimic the unique content approach of Netflix by connecting data that is not available to users right now. I focused on building a new connection. What’s not suggested now by Google is books.
Book suggestions fit in the overall Kaffae theme of reading. Many people search online sometimes for entertainment, but many times they want to learn something for work. That is when book reading can help dive deeper into the topic: whether it’s about a technical book like Deep Learning or theoretical understanding like Color Wheel. It’ll be a new way to access information with online articles as a gateway.
Other Suggestions?
Before coming to this conclusion, I’ve been testing out with the podcast recommendations. There has been a recent breakthrough in transcribing deep learning that I could have added the podcast recommendations to Kaffae.
It was a tough call. With podcasts, Kaffae needs to provide apps, which would take away a significant amount of development time. And I wasn’t clear how much Kaffae users would like to listen to podcasts. It was also true that the competition was getting fierce with Google just releasing iOS podcast apps.
Book suggestions instead would fit with the overall theme of reading. And it aligns the value of access to knowledge.
The lesson to take away is to go with the theme of your apps, and don’t jump in with the buzz words phrase because giant corporations would be already on the way to work on it.
