#8 - The 1 Business Concept That Helped Make An Extra £350,000+
Unstructured vs Structured Research
In reality it’s probably a lot higher than this 300k number but before officially thinking about this concept it was more difficult to split the concepts and a lot of time was realistically just wasted.
Unstructured vs Structured Research
You could also replace the word research with thinking or analysis or any type of process around problem solving in a creative way to find a solution.
Doing the unstructured part properly has saved thousands of hours over the course of the last 3+ years and it makes the structured (more obvious) part a lot quicker to refine.
So What Is It?
Unstructured and structured research are 2 ways of trying to solve a problem that feed off of each other.
The process enables you to solve problems by first using unstructured research (wide/do anything) and then put this into the structured research process that is far more systematically and engineered.
It’s the artistic/creative side, coming up with ideas for the engineer/scientist to solve.
But, it’s just you.
Unstructured Research
Unstructured research is the “you don’t know what you don’t know” part of the process.
During this research period you want to go WIDE. Breadth of knowledge….
Here, you want to go WIDE.
Breadth not depth.
You want to travel and research down dozens of different rabbit holes. You are looking to connect dots, create options and simply to find different angles that may or may not work.
Some will sound crazy strange, others will have the potential to be real in the future, but for now we’re just on about ideas.
This is where people need to be if they say “they have no ideas”.
Unstructured research has 3 core goals:
Find out what you want to look into - Niche/Industry
Find the 3+ things other people do that works.
Find 3+ tangential ideas that might potentially be useful in the future *These should sound a bit weird.
An Example
Unstructured research can be tough to understand as its so broad. This is also the reason very analytical people struggle during this section, they try to approach unstructured research like the process of structured, and hence get no where.
We have to go as wide as possible.
One example is in sports betting, this is a niche I’ve been in for years and until the past couple of years was pretty poor at earning from.
Most people approach sports betting from an analytical standpoint, they look into the types of data that you can gather, try and find the content online that might help, even follow people for “tips”. All of which can (in some unusual cases) work, but in general it’ll probably fail.
Instead, just go down a few rabbit holes really deeply.
The first should be the sport - Go see if there’s a niche sports podcast about the particular sport you are trying to beat.
For example, if I’m looking to get into tennis betting, I’m not going to want to listen to the mainstream tennis podcasts and content, that’ll all be too generic and obvious.
Instead find a podcast or YouTube channel that talks about one part of tennis, or a group of players closely or a group of tournaments (maybe only clay for example), and follow that rabbit hole down.
You can then start to pick up on angles or sentences you hear repeated alot. “Clay tennis is very different” - “Players who grew up playing on clay have an advantage because of XYZ” - Don’t worry if these things sound a bit odd yet or if you don’t think they’ll work later down the line, that’s a problem for structured research.
Another angle is to take the opposite of what people say, or use inversion thinking.
If someone says “only goalies matter in ice hockey” - take the approach that what if they don’t matter at all… Or if players think only individuals who have top 10% serving stats can win grand slams in tennis, take the approach that maybe that’s wrong as well.
Just go down every single rabbit hole related or tangentially related to your project.
Another example is taking information from different industries and moving it across, such as investing or trading. This has helped create tons of angles in the sports betting space, all of which are painfully obvious in trading but most people haven’t converted across to sports yet.
MLB’s Moneyball moment happened in the early 2000’s. That’s about 20-30 years after the same principals were taken into account in the financial markets.
In short, financial markets are very early on creative ideas, you can take some of these across - things such as structure, support/resistance, moving averages actually have applicability to teams performances in sports believe it or not, you just have to play around with how you can make it work in the sports niche.
Structured Research
Structured research is taking these concepts and testing them to see if there’s any validity
For example, how can we be sure that the concept “only goalies matter in hockey” or “reviews matter for amazon rankings” is another common one I’ve heard in my other industry.
The truth is you have to test it. Build case studies, split test those, test them again.
This is where the analytical side comes in, it’s almost like the scientific process. We have a hypothesis, we want to create a fair test or experiment to understand whether or not that hypothesis is correct or wrong, or, as in most cases, partially right.
This is where the power of the unstructured to structured research levels comes into play, you keep bouncing from one to the other and adding things, taking away what doesn’t work.
After you have developed a list of concepts from unstructured, you test these in the structured process.
At a core level we want to answer complex questions in chunks, first coming up with creative ideas and then split testing each one, iterating and adding/removing elements.
For example, one of these complex questions I wanted to answer was “how can I rank a product in Amazon”
The unstructured side started with simply Googling this, going down rabbit holes, watching YT videos, reading case studies and content, all creating multiple ideas, problem was this is all opinion. I also came up with other ideas, based on how other algorithms (Google, YouTube, Airbnb etc) rank.
This gives you a list of variables, ideas, concepts and just an absolute mess of things to test. But from here you create micro experiments.
“If I do this, XYZ should happen”.
Hmm - XYZ did not happen… That’s strange because 95% of people that said this should happen, maybe I did something wrong.
Run the test again for a different product/listing — Same result? - Improves validity.
“Maybe if we try this instead”
Ok that worked really well… But that was only said by this one guy, let’s go back and review his content (back to unstructured research).
Split test more ideas.
Build your confidence in the actual process.
You can only be truly confident about something once you have repeatably done it with the same or similar outcomes. This is why the unstructured to structured methodology works so well, because after it, you have a real, repeatably EDGE.
And Edges,
Are VALUABLE.
But anyway, that’s how I approach things now, hopefully it helps.
Cheers,
Thomas.