As developers, we cannot trust raw statistical data about people. However, for some reason, when we see stats and figures from studies about people (X load time leads to Y increase in conversions), we tend toward blind faith in them. We see people like machines; a value into the system computed result out.
Such a presumption makes sense. In computer science, we make many decisions based on raw data. Computers are mathematical machines, and numbers are the truth in math. Nevertheless, from the mathematical world, we cannot infer this type of truth onto the non-mathematical.
People are not computers; mathematical data around user behavior is often very deceptive because it infers math-as-truth upon a world that is not made strictly of numbers.
For example, do fast, under two seconds to load website speeds cause an uptick in conversions on e-commerce websites? Of course, but the numbers are disingenuous because they disregard other factors like usability, UI, and messaging. The numbers are not “truth” even though they are solid. The numbers are a single signal source and can be static than clarity alone.
We need to recognize. There are other signals of importance. We need them all to investigate and empower human beings in the technical space effectively:
- Empirical study (the numbers).
- Human psychology.
- Philosophical implication.
- Physiological factors.
With each added signal, we can improve the overall strength of our insight. However, even with all these signals, there is still a flaw — our biases.
Bias will impact how we interpret signals. This includes the data signals provide and the signal itself – how and what we collect. Plus, if we dismiss signals because of dogmatism or mindset, we also risk missing out on the opportunity to improve the very process of collecting these signals.
We cannot assume the numbers or ourselves alone can unlock the truth.
So, what can we do to get better signals?
- First, do not infer math-as-truth in the non-math world.
- Second, collect multiple signals and do not be dogmatic about it.
- Third, get the same signals from people unlike you in mind, spirit, strength, and body.
- Repeat the above steps over time and space as randomly as possible with a high level of frequency.
- Last, do your best to make something of the information as a person.
It can be tempting to turn to math for everything in the tech space. Do not. Remember that the human experience is human.