Loading Profile...
Ideas to improve your algorithms, services
I like the idea of Mr. Tweet but think your algorithms need to improve. For example, I'm really not interested in following someone who follows 1000+ people and especially not interested in those that follow 10,000+ people. I want to know who the people that I follow are following and especially want to know who the people I reply to are replying to.
It's a bit like PageRank; if someone follows 100 people and are active then I become 1 of 100 if they follow me. If they follow 10,000 people then I become 1 of 10,000 when they follow me.
Further if 100 people follow them (and they are active) then they are more likely to interact with me than if they have 10,000 people following them.
Another thing to look at would be how many of their followers reply to them?
Another thing I'd like is for you to discount the people with a following/follower ratio of more than 10 (i.e. if they follow 1000 people and less than 100 people them, they are a spammer and are insignificant.) This could be a sliding scale, i.e. if they follow 5x as many as they have followers it would be worse than a 1-to-1 but not as bad as someone with 10x. After 10x, they should be discounted entirely AND you shouldn't even run their statics for them; you should tell them to get real and come back later.
Also I'm really interested in a local geography so I'd like to see people who live within 50 or 100 miles of me first.
I'm basically trying to leverage my network to learn who else I want to follow, but since I use Twitter to build relationships I have no interested in following Twitter "rock stars" because I can't really build a relationship with them because their attentions are so divided.
It's a bit like PageRank; if someone follows 100 people and are active then I become 1 of 100 if they follow me. If they follow 10,000 people then I become 1 of 10,000 when they follow me.
Further if 100 people follow them (and they are active) then they are more likely to interact with me than if they have 10,000 people following them.
Another thing to look at would be how many of their followers reply to them?
Another thing I'd like is for you to discount the people with a following/follower ratio of more than 10 (i.e. if they follow 1000 people and less than 100 people them, they are a spammer and are insignificant.) This could be a sliding scale, i.e. if they follow 5x as many as they have followers it would be worse than a 1-to-1 but not as bad as someone with 10x. After 10x, they should be discounted entirely AND you shouldn't even run their statics for them; you should tell them to get real and come back later.
Also I'm really interested in a local geography so I'd like to see people who live within 50 or 100 miles of me first.
I'm basically trying to leverage my network to learn who else I want to follow, but since I use Twitter to build relationships I have no interested in following Twitter "rock stars" because I can't really build a relationship with them because their attentions are so divided.
I’m undecided
1
person likes this idea
I like this idea!
Tell me when this idea gets some attention.
The more people who like this idea, the more it gets noticed.
The more people who like this idea, the more it gets noticed.
The company has this in progress.
Create a customer community for your own organization
Plans starting at $19/month
-
Inappropriate?Hi Mike, roger that. I took a look at your reports, and I certainly think that everything you said makes a tonne of sense. You want to build meaningful relationships, and relationships come from having sufficient attention being allocated to every follow they make. Essentially, people who
- follow 10X more than they get followed
- follow 10000 people, even if they have 20,000 followers
- every tweet is ignored
really are not suitable for building relationships, since it is likely that they do not pay attention to who they are following
We will be working on it. We are getting sufficient data points to make a major ranking update, and we hope to get to that once we get some breathing room!
Re: the geography: roger that. Problem is that Twitter has pretty poor location data. hmmm... =(
I’m thankful that we are getting feedback!
-
Inappropriate?Hey, thanks for the quick reply. I think you understood exactly the crux of my wants on this. A bunch of us local startup-oriented entrepreneurs were discussing this problem exactly but none of us really want to tackle that problem so glad you are looking at it.
I get your point on the geo data...really wishing it could be better. Wondering if someone could layer that info on top of Twitter. Hmmm... =)
I’m patient to see what you have pending.
EMPLOYEE
