The illusion of progress
Root Cause Essay #2
Second essay of the biweekly series, “Root Cause”.
We have too much “conventional wisdom” that is creating more problems than solutions.
It’s time to challenge it, one concept at a time, to the root.
In the previous essay, Who to blame?, we discussed how, when a growth initiative fails, we usually blame three things as “not working”: Implementation, then plan, then external factors.
These three causes are our first likely culprits because they are the most visible ones. But what if we are missing something, a fourth culprit that we don’t see?
And even more worrying: What if that fourth invisible cause, one that is baked into the way we’re taught to think about growth in Startups, just gets us into a loop where it looks like we are making progress, but it’s just an illusion?
Let's start exploring that with a story.
1. Two ways to solve a Rubik's cube
I recently bought a Rubik’s cube for my son; I thought that it could be a fun challenge to solve together. Of course, he really enjoyed the part where all the perfectly aligned sides became a multicolor mosaic, and then he got bored and left the challenge to me.
So, I started to solve it in the way that looked more logical: one side at a time. The first side was tricky, but by trial and error, you start to learn some ways to make progress. The second side is, of course, harder because you have many limitations if you want to keep the first side in place. But it was doable. Great, just 4 more sides to go.
By the third completed side, when trying to solve the fourth, it felt like I was not making progress anymore; every time I solved something, I was also breaking something else. It felt like running in circles. So, I started to suspect that I was using the wrong logic; I was following a rule that made logical sense to me (one side at a time), but maybe my logic was off.
After some research, I found out that you actually solve the cube by what the experts call “layers,” that is, by rows and not by sides. With that new logic and some instructions on how to align each layer, it just took an hour to get it solved.
What surprised me the most was the fact that when I was solving the cube by sides, it looked like I was getting closer to solving it. It felt like I was making progress until I was not. Then, once I understood the right logic to solve it, it became immediately obvious that my past progress was just a mirage, an illusion. I was just getting into a trap that was harder and harder to deal with.
2. Invisible models
As per Wikipedia, a mental model is “an internal representation (model) of external reality: that is, a way of representing reality within one's mind.”
What I called “my logic” to solve the cube is a mental model (just “model” from now on), one that turned out to be wrong. That model was not that important, it is just a game, but there are models that affect how we make much bigger decisions, like the models we have about “how startups grow”.
On the one hand, models are useful to help us deal with the complexity of the world; on the other hand, they are dangerous because we don’t see them; they are just “how things work,” facts of life, indisputable truths.
You know, like the fact that the sun rises on the East, hides on the West, and flies underneath our beautiful flat earth until it rises on the East again. And that was the “truth” until it was not (well, for some, it’s still true, but that is a story for another day). Models are useful but invisible, so we don’t challenge them, and they could be wrong.
3. The current model
If you remember from the previous essay, we had a team that made an assessment of the situation that we called the point “A”. That assessment is the byproduct of looking at some indicators and interpreting them based on their model about “how startups grow”:
Their product has many good reviews and a significant MRR (indicators), so they understood that they had a solid product, one that people prefer over competing alternatives (assessment). Now, they just have to show other people that they have a better product, and they will do that with Paid Ads (plan).
The previous paragraph makes total sense. Also, it totally omitted the model. That paragraph logic is built on top of a model that we don’t see anymore. So, let’s unearth it. It's simple; it just has 2 assumptions:
People will switch to a “better” product (10x product)
We know that there are many similar alternatives to choose from, so having a “better” product is not enough. To make people switch to us, we have to be much better; “10x better”.
Then, we have to persuade people to Try the product
Some people will realize we are 10x better and will just switch to us, but a lot of other people are busy or distracted and need some “extra motivation”. We have to “persuade” them to Try the product so they realize what they are missing. Then, they will also switch to us.
This is it; “10x product + persuade to Try”.
4. How we use the model to make decisions
This deep belief about how things work, the current model, affects how we assess each situation and the plans we make to deal with it:
If we are not growing, we will think that this is either because we are not 10x enough; therefore we will add more features. Or because we are not persuading enough; therefore we will “increase persuasion” with better ads, timed trials, reminder emails, offers… And if after that, we are still not growing, we will repeat the process.
So, the company from our thought experiment decided to run ads as the way to persuade more people to try, but they didn’t work. If they believe the reason was the implementation, then they will try again to see if a different Paid Ads Expert is better at persuading. If that doesn’t work either, they will start looking into the funnel. Maybe they could gamify the onboarding to make it more persuasive. If that doesn’t work…
With each one of these changes, it feels like we are making progress. We are replacing something that is not working with something else that could work, so we should be closer to a combination of elements that work in unison. But careful: We will be closer to a combination that works IF, AND ONLY IF, the model we are using is right.
But, like with the Rubik's Cube, if we are “making progress” completing sides, then we are not really closer to that combination. We are just trapped in a loop where everything we add solves something only to break something else.
And based on experience, where we tend to see how an experiment improves a metric, but the revenue numbers remain flat, that could perfectly be the case.
5. The 4th culprit: The Model
So here we have our potential 4th culprit: The model. It’s an invisible culprit because it’s so ingrained in our minds that we don’t see it anymore; every assessment and every decision starts from it without mentioning it.
It’s so ingrained in our minds because this is how we were told things work. Nobody told us that “10x + persuade to Try” is one potential option out of many, a mere hypothesis to explain how the world works. Instead, we were told that this was the truth, and all of us believed it
And we believe it because it seems to make intuitive sense, so what evidence do we have to doubt it?
Well, consider that the model was not always like that. We went through several “truths,” like the Earth that was flat, until it was not.
During the dot-com boom, the model was “build it, and they will come”: If the product was web-based, then it was assumed to be better than any offline alternative, until it was not.
After the crash and coinciding with a drastic reduction in the cost of releasing a new product online, the model evolved into “it’s not about being better, it’s about being faster” (Lean startup MVPs, Facebook “move fast and break things”) until breaking things stopped looking like a good idea and MVPs became less “minimum.”
Then, the new tracking capabilities, combined with the prevalence of the subscription business model, led to LTV>CAC as the formula to follow. The way to increase LTV: “10x product.” The way to reduce CAC: “Better persuasion.”
Also, during this time, consumer behavior also changed. From “how to discover new products” to the tiredness of the “persuasion everywhere” to trust issues towards tech Startups as a byproduct of continuous horror stories in the media.
That is more than enough evidence to force us to stop and challenge our models.
6. Challenge, not refine
On each iteration of the “replacing something that is not working with something else that could work” loop, we adopt the latest technique that seems to be working for others to maximize the chances that, this time, it works.
One of these new techniques is Product-Led Growth (PLG). So, could that be the way to make real progress?
If you look carefully at PLG, it is just a variation of the “persuade to Try” model. PLG is “persuade prospects through easy access to the product” instead of “persuade prospects with a salesperson before they even see the product” (Free trial vs. Book a demo).
Techniques like this are simply refinements over the same model. So if the model is wrong, or it has missing assumptions, then we will do an erroneous assessment, we will include the wrong technique in the plan, and when it doesn’t work, we will blame “implementation” or if we trust the person who did the implementation, then “prospects have now easy access to the product, but they are not buying it because it’s not 10x enough”, and back to the loop.
7. Looking forward
So we are trapped with all these new techniques that worked for somebody else, iterating implementation and planning until we give up and move on to the latest one. But we keep at it because it feels like progress. We keep replacing something that is not working for something that could work, all of it while maybe we are trying to solve the “startup cube” by sides instead of by layers…
So, let’s challenge the model. Look, the fact that things like PLG worked for some and not for many others is already telling us that there are other factors that we are not considering, other models about “why people choose one solution over another” and “how companies grow”. New models that will help us choose the right technique at the right time, and even create new techniques that make more sense under this new understanding of how things work.
We need a better map to guide us so we can escape this illusion of progress and, instead, see how all our hard work becomes real progress.
8. Key Insights
When a growth initiative fails, we usually blame: Implementation, then plan, then external factors.
There is a 4th potential cause: The mental models we were taught about “how companies grow”.
The current model is: “10x product + persuade to Try”.
There is enough evidence to think that the problem is the models we use, models that lead us to use the wrong technique at the wrong time.
We will start challenging the model in depth in the next essay.
And remember, you solve a Rubik's Cube by layers, not by sides.
Great insights!