Feb. 6, 2016

Twitter is reportedly going to change the timeline strategy from time-series to interest-based one. People are expressing feeling against it, saying it is worsening the experience. Why do they tend to be hostile when a system recommends a list of items? Referring to only the Twitter case, it's due to just a change from accustomed way. However, we don't also get used to the same strategy that Facebook adopts.

What I'm actually interested in is the condition that can meet people's preference. For instance, TV news is alike to the interest-base recommendation. That is, it offers topics not in a time-series manner but by picking the facts up and summerising them. Nevertheless, people don't express discredit to the manipulation of TV. What's the difference between Web services and TV?

I think it's an essential issue how to design recommendation/personalization feature in Web services. It's important if people can be satisfied by the offer from a system or not. While they can't understand whole of the situation anyway, they want to convince themselves to be able to be aware of the wholeness. It may come from the belief in authority of experts or possibility of choice by themselves.

Granularity of information may matter; Twitter restricts a post within 140 characters, whereas Facebook doesn't. It can affect the people's perception on each services. Anyway, I'm interested in what differentiates the degree of satisfaction among the media. It's practical issues for me, as an architect working on Web services.