Paul Kelly’s random thoughts

LOVEFiLM don’t understand averages

April3

LOVEFiLM have had a bit of a redesign recently, and one of the features that has been given more prominence is the chart for each film that shows the ‘Average rating breakdown’ (highlighted below).

Well, except for the fact that it doesn’t.

What it actually shows, is for each rating (on a scale of half a star, to five whole stars), the percentage of users who awarded it that rating. Which isn’t an average by any definition I’ve ever heard of.

Lack of basic maths aside, it’s also a pretty poor way of presenting ratings, which fails to make the best use of the given screen estate. Allow me to explain.

Because the whole range from zero to one hundred percent is used along the horizontal axis, there’s always going to be a lot of empty space. There would have to pretty much unanimous agreement on the exact rating of a film, down to the precise half-star, to get anything even resembling a spike. And because of the very personal nature of ratings (two people could both say they absolutely loved a film, but one would give it a four and the other would give it a five), this ain’t gonna happen. This leaves the graph looking a little, well, limp.

Even for films generally held in high regard, such as The Dark Knight (pictured), which 77% of LOVEFiLM viewers awarded 4 stars or over, the graph doesn’t look particularly impressive.

A better way to present this would be to dynamically adjust the maximum value of the horizontal axis based on the highest percentage. So in the case of The Dark Knight, the majority percentage of 37% would cause the scale to be from zero to forty percent, allowing the graph to fill up the space nicely.

IMDB understand averages. They give you the arithmetic mean AND the median. Rock.

posted under usability | 4 Comments »

Wow, that’s a lot of different types of chicken pie

August21

After being rather disappointed with the service we’ve been getting from the Sainsbury’s online store lately, we decided to give a couple of their competitors a shot. First up, Asda.

Asda

The folks at Asda seem to have decided that the tried and tested concept of logging into a website with a username and password is far too dull, and so have instead opted to go for the no-chance-of-remembering-this eight digit “customer ID”, and a super secure four digit “PIN number”.

So, in we go to the online store. Products are organised by an “aisle” methaphor, listed down the left hand side.

Not sure what the colour coding means here. Does it mean that Bakery & Cakes, etc are all in stock, but they’re all out of Baking, Deserts & Spreads, etc? Who knows.

Right, first on the shopping list, chicken pies.

A search for “chicken pies” and I’m informed that “Your search for chicken pies returned more than 50 products”.

Wow, that’s a lot of different types of chicken pie. I like pie. Let’s have a look at those results.

Hmm, Indian and Chinese cuisine aren’t exactly famous for their chicken pies. I obviously wasn’t looking for chicken pies when I typed “chicken pies” into the search box, so I’ve been given what looks like the search results for chicken OR pies. I really can’t be bothered to look through all those just to find some pie shaped goodness.

Finally I noticed the title of the page - “ASDA @t home”. As we all know, @ is shorthand for “at”, so the site is called “ASDA at t home”. That’s the final straw, I’m out of here!

posted under usability | 1 Comment »