Hacker News Reader

Comment by austenallred | Hacker News Reader

Twitter’s biggest problem is that Twitter increases or decreases in value along with user’s willingness to curate their feed. Many will never do that.

So Twitter tries to do the work for you, but guessing exactly who/what you’d be interested in without tons of info is virtually impossible.

The closest they can get is the current logged-out homepage: “Here are a bunch of random categories. You like the NBA? Maybe celebrity chefs? Cute animals? Country artists?” (Those are literal examples.) Total shot in the dark.

View on HN

I see this as the biggest problem with Twitter also.

Twitter is my favorite app, but it's because I have spent five plus years curating my feed.

There is a fundamental problem with the way they onboard new users and this is why so many people don't stick around.

The onboarding process asks you to select interests and then recommends people to follow based on those interests.

However, they often recommend celebrities and 'popular' accounts.

These people are often terrible at Twitter, so new users see a bunch of boring, self-promotional content in their feed and don't come back.

How did you go about curating your feed?

I tried to get into Twitter for a while and ultimately stopped using it.

Part of the problem I think is I have kind of a wide range of interests -- sports, software, entrepreneurship, etc. So Twitter probably can't gauge my interests very well. I ended up with a lot of those self-promotion feeds and ESPN writers, but nothing ultimately with substance (except for certain people that I found via other sources -- like if I follow their blog, I'd follow their Twitter too).

It'd be better if it was easier to categorize the feeds I'm following and view by category depending on what I feel like at the moment. I know they have lists, but I only just found that a little bit ago -- and it's so hidden that I actually have to type the URL into my address bar because I can't get there from a link on the homepage. Viewing by category should be, if not the default, then one big link away from the initial page.

Lists are the way to go.

Conference or other temporal hashtags are a good way to discover users who are focused on a single topic, who you can then curate into a list. Once you find a few topic experts, they become a discovery source for other experts.

Exactly. I had several abortive attempts at Twitter before I 'got it'. You have to build up your feed slowly and organically. Start by making the effort to follow a few, very high-quality accounts based on your interests, then use these accounts to find others, and so on (unfollowing as necessary to keep the signal/noise ratio as high as possible). This requires substantial effort on the user's part.

I use TweetDeck. It's like the power user version of Twitter. I generally have 20+ columns on my screen at any given time.

Most of the people I follow I found through people I know retweeting them or through looking at the feeds of other successful/important people (for example, I recreated the feeds of other investors/entrepreneurs) - http://theireyes.austenallred.com/

TweetDeck seems to be exactly what I was looking for!

And that other site (twitter through their eyes) looks awesome. Thanks!

Twitter is like any other social network in that it works better if the people you know personally are using it. The only way to build up a good feed is to explore out through your friends' retweets to find accounts they follow that you also like. Then periodically look at your feed volume as a whole and delete anyone too high-traffic or recently uninteresting.

That's true. Most of the recommendations are crap. Got any recommendations to finding high signal/noise users or lists ?

There are a few methods I've found for finding good lists.

1) Google with "site:twitter.com inurl:lists big data"

Google's listings are decent indicator of higher quality lists. Substitute whichever topic you are interested in.

2) Look at your sources' memberships. If there's someone on twitter you follow closely for a topic, check out their memberships to see who has already listed them for that topic and that list may be one to start with.

3) We have a list search you can try out and sort by various metrics to find higher signal lists

https://www.electoralhq.com/twitter-lists/search

Thank you!

>"guessing exactly who/what you’d be interested in without tons of info is virtually impossible."

That was my intuition as well. But it turns out that you can generate good recommendations without tons of info -- as long as you are selective about what signals you use.

We learned this firsthand when building https://recent.io/ for news recommendations -- the backend is able to generate relevant suggestions that appear in the app after users have read an average of 5 articles. That's based on real-world user testing.

If Twitter recognizes repeat visits by people who don't (yet) have accounts, it should be able to try a similar approach. On the other hand it took a while for us to get it right.

Reply on news.ycombinator.com