Rating Recipes with Stars
In my latest project I’m working with a platform where the core content is cooking recipes. It’s more aimed at cooking enthusiasts and hobby chefs primarily.
In the platform the users have to rate recipes others have posted. Why do we need a rating system? When you are given with 100s of books, movies, products‚…, you need some measure to select one or more from the pile. People uses rating as a common tool for this. On websites it has become a democratical way to decide what content or product is of higher quality.
Popular Rating Mechanisms
Voting: In the web sites like digg, stackoverflow, quora etc. one can give thumbs up or thumbs down for a piece of article. If we look at the type of content these sites are handling are news articles, opinion (question & answers) etc. The votes are like ‘I agree to the article or I don’t agree’, ‘I like or I don’t like’, ‘I recommend this or I don’t recommend this’, so on. Its a binary opinion, discarding the ‘I don’t care’ state of being not voting.
Stars: It has it roots for the printed reviews like Michellin Guide. Commonly 3 stars or 5 stars. When it comes to websites 5 stars is more common. Since this techniques is prevailing since many years the people understands right away. It somehow gained an affordance of grading over the course of time with users.
Score Based: Similar to starts but here the range of rating go beyond 5. Usually on a scale of 1 to 10, 1 to 100 etc. In the website this done either by a scroll range bar or type in.
I Like Button: The Facebook introduced a new and simple grading mechanism, which you either like or not.
Coming to Recipes
My friend @cherianthomas told me chefs are artists and each of their recipes are their art pieces. I agree with him. So how do people rate art pieces? Coming to paintings you may say I like the stars shining giving the light to the town…, blah blah. You will just bring out you emotion. Coming to food you will also have similar kind of emotions. The food and the recipe from which it is made are two seperate entities. The recipe is like a tool which one used to prepare a dish. All the emotions when you eat a dish will just for the dish you ate. So here we just need to rate the tool, that is the ‘recipe’.
If I give you three hammers and ten nails to fix. Say a claw hammer, a sledge hammer and mallet. Which one do you use for nailing? I would pick claw hammer for that purpose, because it just works on nail. See here a utilitarian object of the hammer comes into play. Similarly when you speak about recipes, it should be doable for me, the context, re-usable,… these kind of parameters churned in your brain simply makes you say things like: “I’ve tried this recipe, didn’t work because of so and so… I won’t cook this again”, “It was easy, sure I can use this recipe for breakfast”,…
These are the kind of opinions I need to collect. Then their summaries I have to represent effectively. Which is actually can be done in many ways. Can be with heat bar, stars, textual,… From these I chose stars rating control. If implemented well you just need to see & one click. No dragging, thinking which number to give, multiple. I would call this simple. Other simple ones are voting up/down & ‘I like’ which I rejected because of the bandwidth of opinions possible are not enough.
Say I have five kind of opinions for example:
- I won’t cook again
- May be, may be not
- I may cook again
- I will cook again
- Its become one of my favorite
These can be easily mapped to the 5 stars rating. So the two sides of a rating mechanism: collection & representation is handled.

Interaction with control is simple. Almost simple enough to say just Read & Click

Also the stars rating snippet can be used in different sizes and colors depending on the context without losing the meaning
This is a part of the thought process which I had while I’m working on the rating and resulted in finalizing on stars rating for recipes. Do share your views on this.

comments
You’ve described a couple of the parts of a rating system: collection and representation.
The design of rating systems ideally includes some parts unseen by the end users. Check out Randy Farmer and Bryce Glass’s Building Reputation Systems (<a href="http://buildingreputationsystems.com" website and O’Reilly book) to learn more about calculations on the backend (which means my rating this recipe with 4 stars might be worth someone else’s 4 or more or less depending on various factors), criteria to help you decide which visual treatment might fit your situation, why “don’t like” is typically a poor choice to include, and more.
@Nancy, Yes I totally agree with you. Like an ice berg, the unseen part needs is big in terms of effort and is the driving force in driving the rating system. Thank you for suggestion & the book reference. Sure I will check it.
Here’s a challenge I’m facing ATM. A Chinese living in Germany, if he decides to post a snake recipe chefoke.de will invariably get only 1 star. This in all probability might be a fine dish with the Cantonese.
How can I take this into consideration when aggregating the rating?
This problem is cannot be handled by a rating system. Rating is just a tool to represent the acceptance or liking of an item (recipe, movie, post, news) accepted by the majority of the community using the platform. Posting a Cantonese fine dine recipe in a German cooking platform will not be favoured by the community.