The creative heuristic algorithm can help spark new business ideas for your next side hustle, or even new blog post ideas to make money blogging.
The creative heuristic
algorithm, developed by David Mercer
, provides a mechanism for creativity on-demand
by generating a large array of both closely and distantly related semantic concepts for the brain to associate.
Our brains are weird. They are responsible for logical, procedural planning and creative sparks of problem-solving genius — both very useful attributes in business.
The problem is that creativity seems to be elusive. It comes and goes. Not easy to conjure on demand.
- Bloggers suffer from writer’s block.
- Entrepreneurs struggle to come up with a creative solution.
- Marketers try to work out what will inspire their audience.
- Artists can't seem to find the right inspiration.
Everyone understands the frustration of trying to force creative ideas to flow. Often with nothing to show for the time invested.
First, some scientific background on how and why the creative heuristic works.
Why We Use Heuristic Algorithms
Not all problems are easy to solve. There are beasts out there, lurking in our mathematics — such as Non-Deterministic Polynomial Hard (NP-Hard) problems.
The Multiple Travelling Salesman Problem (mTSP) is a famous type of NP-Hard problem that applies to every-day life. Essentially, you want to work out the most efficient way for a bunch of people to visit a bunch of locations.
Obviously being able to optimize routes efficiently has real-world applications in many spheres of industry. In particular, so-called last-mile deliveries can take nearly 30% of all supply chain costs. Reducing costs here would seriously benefit almost every aspect of every economy.
Solving the Notorious mTSP
Let’s take a very, very moderately sized traveling salesman problem to highlight why this hard to do.
Assume a small company has one delivery vehicle and needs to visit 100 locations.
What’s the most efficient way to do this?
We could test every single possibility and then choose the lowest cost route (i.e. the route that has the lowest overall cost taking into account distance costs like fuel and wear and tear, and time-based costs like paying drivers).
To work out how many possibilities (permutations) there are in the problem we can say the following:
- Picking the first stop has 100 possibilities
- Picking the second stop has 99 possibilities
- Picking the third stop has 98 possibilities
- 100. Picking the last stop has 1 possibility
To calculate the total number of permutations, we say:
100 x 99 x 98 x 97 x … x 3 x 2 x 1 = approx 9.33 x 10¹⁵⁷
That’s a pretty big number.
To put this number into perspective, the Sun is thought to have about 10⁵⁷ Hydrogen atoms.
The Sun is very big. It contains about 99.8% of all the matter in our solar system.
A Hydrogen atom is very small.
10⁵⁷ is an enormous number. Yet it is infinitesimal compared to our 100-stop delivery problem.
How much smaller is this number?
Estimates put the number of stars in the entire known Universe at about 10²³. That’s a lot of stars. Assuming roughly the same number of Hydrogen atoms in each star (our sun is pretty small, all things considered) would give us roughly:
10²³ x 10⁵⁷ = 10⁸⁰
The number of Hydrogen atoms in pretty much everything that exists; the earth, the moon, the sun, the solar system, the Milky Way, the Andromeda galaxy, the local cluster, every galaxy, everywhere is still a negligible fraction of the size of our 100 stop problem.
Even if we underestimated the size of the Universe by a factor of one million that would still bring our total to only 10⁸⁶. There isn’t really language to describe how much smaller this number still is. A vigintillion times smaller is as close as I can get.
Why Traditional Computing Doesn't Work
Given these astronomical figures, it’s understandable that one can’t simply write a script that instructs a computer to try every permutation. Not unless you’ve got some serious time on your hands.
The fastest supercomputers (all of them working together) on the planet would take longer than the remaining time left in the Universe (regardless of whether you favor a heat death, big tear, big crunch, or another as yet to be conceived doom) to come up with the answer.
Of course, quantum computers might be able to handle this type of problem much quicker because they should, theoretically, be able to try an infinite number of permutations simultaneously.
Unfortunately, they’re not quite production-ready at the moment. We’re stuck with classical computing for now.
To solve (by solving I mean, close approximate) something this complex requires a… change of thinking.
Instead of solving procedurally (i.e. take one step after another in a defined way until the solution is reached), we can use heuristic algorithms.
How Heuristic Algorithms Do Work
A heuristic algorithm is not procedural. Instead of following a set pattern of steps, it compares alternatives and tries to improve on them over time to reach a good approximation of the solution.
Nature uses heuristic all the time.
Consider an ant colony that has to scout for food and resources. Initially, ants head off in all different directions. If one finds some food it brings it straight back to the nest leaving a faint pheromone trail for subsequent waves of ants to pick up on.
If those ants find food at the same place, they also bring it straight back to the nest, adding more pheromones to the trail for other ants to pick up on. Over time more and more ants follow the strongest pheromone trails leading to the emergence of very efficient and direct routes between the hive and its surrounding resources.
It’s not only ants, either. Our genes employ heuristics.
They splice up, combine and recombine (mutate) in different ways leading to offspring with different characteristics. If those characteristics are favorable to survival they may be passed on.
In this way, mutations lead to the emergence of fitter individuals adapted well for their environment.
Heuristics aren’t simply a sneaky way to program software; they’re built into Nature on a fundamental level.
The 'Creative Heuristic' Algorithm
Humans have a strong procedural bias.
We identify a problem, come up with a plan of action consisting of a sequence of clearly defined steps and execute that plan to solve the problem. We apply that thinking over and over, again and again for pretty much every problem we want to solve.
Creativity is not procedural.
We need to take a more heuristic approach by giving our brain the opportunity to compare many different, potentially creative possibilities.
If a problem is creative (i.e. not procedural) then it is likely an excellent candidate for an heuristic approach.
It’s almost impossible to sit down and think of a game-changing disruptor off the top of your head. Similarly, coming up with a new small business idea or a new blog post idea, or a creative solution to a business problem can be a real challenge.
Association as a Building Block
We can implement an easy-to-use heuristic approach based on one underlying principle.
The building block of creativity is association.
In other words, what we view as a creative idea is often, at its heart, an unusual or counter-intuitive association between two concepts that are connected in some way (either closely or loosely — see semantic similarity).
If the association is a building block, more blocks mean a higher chance of finding something interesting and new. And, if we want to generate associations we need more “things” to associate.
Thinking inside-the-box utilizes the things your brain already has available.
To change this up, we need to add new and different things to allow for ideas that fall out the box.
Generating Arbitrary Associations
To really give our brains plenty of food for thought, we need a method of introducing a wide range of potentially new and unique associations between a distribution of anchor concepts and control concepts.
Fortunately there is a quick and easy way to do this by merging two lists of concepts in a one to many fashion. Here's an example taken from our small business ideas resource.
1. Create an 'Anchor' List
Create a big list of all the things you,
- love doing
- hate doing
- are happy doing
- get mad at
- are good at
- suck at
- would like to do
- would like to know
- are skilled at
- have experience in
- know about
- are interested in
- would love to learn
Do a thorough job. Try end up with 100+ concepts in the anchor list.
2. Create a Control List
The control list should ideally be a diverse list of concepts related to the niche topic you're working in - i.e. blog post topics, side hustle ideas, etc.
In this specific instance our control list can be made out of concepts taken from existing business ideas, startups, new advances, market gaps, new tech... pretty much anywhere.
There are 100+
creative and unusual ideas for you to choose from in our huge business ideas list
Search Google for concepts and ideas. Collate or jot down ideas you come across on a daily basis. Use existing ideas lists or niche resources.
3. Merge the Lists
Every single concept from the control list must now be merged with each concept in the anchor list.
Pretend my lists have two anchor concepts and 3 control concepts:
Anchor Concept 1
Anchor Concept 2
Business Ideas List
Business Idea 1
Business Idea 2
Business Idea 3
Merging these gives me the following.
Anchor Concept 1 & Business Idea 1
Anchor Concept 1 & Business Idea 2
Anchor Concept 1 & Business Idea 3
Anchor Concept 2 & Business Idea 1
Anchor Concept 2 & Business Idea 2
Anchor Concept 2 & Business Idea 3
Notice that total number of merged concepts is equal to the number of anchor list concepts multiplied by the business idea concepts. For this example,
2 x 3 = 6
These numbers grow very quickly. By the time you have an anchor list of 100 and a control list of 100 ideas there are,
100 x 100 = 10 000
potentially new ideas.
Using Semantic Distance
With hundreds, or potentially thousands, of merged concepts we now have a diverse range of semantically related concepts.
Not every single combination of concepts will have a meaningful association. For example, semantically identical concepts have nothing to associate between them.
The most creative ideas will often arise between the least obvious or least connected (semantically distant) concepts.
It takes a little mental jiggling to merge two seemingly unrelated concepts together into a workable new idea. It is this exact process that regularly surfaces unique associations that simply wouldn’t have occurred to otherwise.
That’s what makes the creative heuristic approach effective.
The creative heuristic approach works by forcing your brain to find associations between a diverse range of semantically distant concepts.
The point here is that your brain is actually being used. It’s not being forced to come up with a brilliant idea from thin air (which is exactly what we ask it to do when we sit down in front of a blank piece of paper and try to be creative).
An Example: The SubMerge Technique for Creative Blog Post Ideas
Consider the problem of coming up with interesting ideas for new content (a common problem for businesses, marketers, and entrepreneurs competing in the online space). Those ideas must be related to the company’s niche industry, but offer something unique and unusual to differentiate it from the mountains of competing for online content.
To generate plenty of potentially unique and interesting associations we can use the SubMerge technique, as explained in-depth in How to make money blogging in the section entitled, Write for Growth.
- Take one or more specific SUB headings from a piece of content to create an anchor list of sub-headings.
- Build a list of semantically related concepts (i.e. from a Google search) of those sub-headings.
- For each sub-heading, MERGE every result gathered from every search.
If you do a search for 2 sub-headings and gather 10 related headings for each you end up with a total of 20 merged concept titles your brain has a chance to associate.
Do it with 10 subheadings and 20 results and the list quickly grows to 200 concept titles to associate — or 200 potentially new ideas.
The exact same process can be applied to coming up with new business ideas or finding gaps in the market no-one else has.
The Creative Heuristic 2.0
For more complex creative ideation, make associations of associations.
Create two lists of, say, 200 merged concepts and merge them with another list of 200 to produce 40 000 potentially new ideas. This creates linked chains of concepts that have significantly more variation in semantic distance than only two merged concepts.
Who knows what unusual associations you might come up with out of that many possibilities.
Essentially, like heuristics in Nature and heuristic algorithms in programming, you want to give your brain plenty of building blocks to play with in order to emerge creative new ideas.
Try it out. Have fun. Let me know what weird and unique ideas it sparked.