Why 5-star rate system is bad.
And what do I offer to use instead.
The 5-star (or grade or whatever) assessment is something that we are really used to. Sometimes it’s split into 10 grades, using half of a star as a step. But is it doing a good job?
My idea is that it’s actually not. It’s confusing in most cases. What does it mean if a film has 4 stars? Does the critic suggests you to watch it? Or maybe he wants to tell that film could be better but still it’s a must? Or maybe he is a rough guy and everything below 5 stars belongs to the trash bin? Not likely though, but still confusing.
The reason why it is so variable in perception is that most of the times human brain actually does not rank the data using some scale, because it’s a very time and energy consuming process. Try to think of your 5 favorite dishes and rank them from 1 to 5. Not easy, huh? In most cases the assessment is quicker and simply leads to answering the question: “do I like it?”. Yes or no are the options here. So far, we have two categories, good and bad. But we want more than that.
We are also able to assess the level of excitement that arises from a certain object, meaning we can say “it’s okay” if it’s just fine or “I like it a lot!” if the object makes us tremble in ecstasy. And this is another operation that doesn't require a lot of brainpower, isn't it? So we have 4 categories: “very bad”, “bad”, “good” and “very good”.
So, after all, this is our simple scale of assessment:
Those are very intuitive, unlike the stars, and there’s even no need to provide the explanatory text.
What do we want from those rates anyway? Imagine a weather app, it tells you weather. But do you really need to know the weather or you just need to know if you should take an umbrella? The same with rates. The 3 of 5 stars rate tells you almost nothing about the object being rated. It totally depends on the critic’s habits and judgement.
To figure out the action (e.g. to buy the album or not) our brain would need to add the explanatory text to the stars and it most likely will use its own patterns formed by previous experience, which are different from the patterns of the critic.
So how to turn those stars into something useful? Convert them into 4 categories (4c) system. We can do it through 4 different ways:
Now we can clearly see why stars are open to interpretation.
4c is straightforward and based on simple emotional assessments that don’t require excessive brainpower and are pretty much the same for all people.
For example, if we take books, I’d say:
“– –” is “you shouldn't ever read this book”
“–” is “you can skip this book”
“+” is “you can read this book”
“+ +” is “you should read this book”
Nice and easy.
Now we have a finished product (4c) rather than raw material (stars) that needs to be processed. We don’t have to guess the critic’s assessment system, we just need to figure out if we trust his rates at all (which is another huge topic, I don’t want to even touch it now).
And it’s actually more useful even if consumer and critic are the same person, because still, it’s a lot harder to rank objects using some scale. And after some time current you will have to interpret yourself from the past, and that could be a totally different person.
Conclusion.
In most cases, any rate system (5 star was an example of the most popular one) is less effective that 4c system. Of course there are some exclusions. But implementing rate system should be a really big and important decision to be made, because in most of the times it’s not worth the costs.
After conclusion, first update.
The only thing that could fix rate system problem from the inside that comes to my mind would be using text instead of symbols. But then again, text could be confusing (if too short) or too long (but clear), and also it has to be translated to other languages etc. And it kinda stops rate system from being actually a rate system. But it’s certainly the option to be considered.
After conclusion, second update.
One more thing, for those who would like to state that there is such a thing as “zero attitude”, and the model should look like this:
– – / – / 0 / + / ++
There’s actually no such thing as “zero attitude”. Everything could be placed into a “+” or “–” category. Even when the object does not interest brain, it hangs a certain attitude label (positive or negative). Mathematically it looks like this:
“Zero attitude” suggests that we have 0 attitude data after processing, that actually equates our brain to a black hole in a certain sense:
Flattering, but far from true.
