Why do I write this article?
My clients and students hear people speak about Lean UX all day long. But they do not know if Lean UX is a good method or just a Buzz word. Many of them would like to be enthusiasts but they grew sceptical with time. Each year brings its share of “magical formula to success” and lean Ux might just be one of them. So is Lean UX a powerful tool that you should consider adopting or is it just a passing trend?
We often live on false assumptions.
Many people do not know how to rationally evaluate a tool. They come up with arguments like “I love this new tool”, “Everybody use it”, “It gives results”, “It does what I expect”, “Famous people talk about it” and so on.
Well… I am sorry but these are not scientific arguments. The fact that everybody thinks that the earth is flat does not make it flat. You might love the idea, you might even invite famous people to talk about it but it will not make the earth flatter.
There are 3 simple questions to rationally evaluate a tool (and grow a critical mind):
– Question 1: Who made the tool? Most tools are not free. People make tools either to earn money or to get attention. For instance, when Google creates Google Analytics, they want you to become a fan boy. When a software company creates a new version of its mockup tools, it is with the hope to sell it to you. And just because companies make their tools popular with the help of marketing does not mean they are good or efficient.
– Question 2: What is the theory behind the tool? When scientists build a Geiger counter, they base themselves on nuclear physics. When scientists build a plane, they base themselves on aerodynamics and fluids mechanic. When scientists build a bridge, they base themselves on Newtonian forces. What is the theory behind your tool? Is there a theory? Is it correct? When Google provides you with analytics, how did they choose their samples? What are they measuring ? How did they treat the results? Did they use valid statistics? Are the results statistically significant?
– Question 3: Does my tool accurately measure what it is supposed to measure? Let us go back to the example of Google Analytics. Many managers are frantically addicted to their statistics. But does it mean that these statistics are valid for all that? Are they a good measure of the reality? After all, many users hide from Google with tools like Ghostery. Others use proxy servers. What does Google measure if there is a black hole in the sample? What is the size of this black hole? What do they actually measure? Which part of the result do they emulate or assume?
Let us analyse Lean UX with this set of questions:
1 – First, let us analyse the theory behind Lean UX :
Lean UX is based on that simple idea that, as a designer, putting your drawings to the test before you launch your product can be a good practice. It avoids the old “just toss it over the fence and see what happens” syndrome. As a matter of fact, all ideas seem to be good as long as they remain paper. But when you test them, problems you did not suspect begin to emerge. Better detect these issues before you put your product on the market.
Customer validation is thus a good way to rapidly improve your product. And if you do it regularly as Lean UX suggests, it helps you detect product flaws, market changes and unfulfilled user needs. I am doing iterative testing for 13 years now, I am not going to say otherwise.
Iterative testing is relevant. And Lean UX must be especially good for startups who have heavier constraints than other companies. When they launch a new product on a new market, startups rarely have an accurate idea of how people will welcome their product. Iterative testing can be a cheap good way to keep in touch with a market you are actually creating. Thus anticipating a lot of problems.
But all the good (Lean) UX practitioners underline the fact that testing is not an end in itself. Testing is just a tool. If you do not know what you test and why you test it, using Lean UX is senseless.
And i have no choice but to note that many companies and many agencies test interfaces without any valid hypothesis or model to base themselves on.
2 – Why is it important to base yourself on a model and to have proper hypothesis?
Well… because that is the principle of science. It is a simple trick that ensures you to find valid results and to build a valid strategy upon them.
Let us take an example: imagine you have a headache. You go to the bathroom, you open the closet, you reach for the aspirin but you suddenly realize that all the boxes look the same and that there are no labels on them. There are 20 boxes in front of you, all containing medication but you do not know which one to choose.
If you have no model to base yourself on and no hypothesis, you are going to use brutal force to solve the problem. You gonna proceed by trial and error: you gonna try each medication until you have the feeling that the headache is gone. But you do not know if it is the medications you tried that solved the problem and some of the medication you took might actually cause more troubles than the headache itself. This is how most people use Lean UX today: they blindly test everything and nothing. They spend time and money to solve absurd questions and to find results that they cannot even explain. Not mentioning that people use testing as if it could solve all the problems. As if testing was the cure for all the diseases.
To accurately measure something, you need a model and an hypothesis. First, you will base yourself on medical science to state that since you have high blood pressure and bleeding problems, aspirin is not a good choice for you. Better look for paracetamol. Then knowing that you can detect paracetamol with the reactive of Mandelin for instance, you will use this reactive on all the pills in front of you to detect which one is actually paracetamol. This simple heuristic will save you time and spare you some troubles.
3 – What is the problem with people who sell this cheap-fake Lean UX:
There are 2 problems:
– First, these cheap UX resellers do not have Psychologists or Sociologists in their team. They do not know anything about human behavior. They do not know the models. They thus rarely formulate valid hypothesis on users needs and strategies. In other words, they do not know what they are looking for. They will make you loose your time and money. And when they say they have specialists, please, check the diploma.
– Second, they act as drug dealers. It is not important that they do not find anything as long as you pay again and again to have the feeling to improve your product.