Twitter's updated discovery efforts

Twitter announced an update to their Discover tab today. 

As you can imagine, this is something I have a lot of opinions and thoughts around. I’m not going to put them all out here right now (mostly because I don’t feel like writing that much right now – or boring you to death), but I do want to at least take a minute to share some of my quick reactions/thoughts:

1. It’s a good start, but it’s not properly described (for what it actually does) and so it will ultimately be ignored or under-appreciated by the masses.

What I mean is that they are saying things like ‘discover information that matters to you without having to follow additional accounts’ and 'surface content that is even more personalized and meaningful to you’…but the reality is their definition of personalized is tightly tied to who you follow (and really nothing else)…and the 'magic’ behind this iteration is really all about popularity (ie. they are basically picking what to show you based on friends-of-friends counts, follow counts, and share counts).

Honestly it’s not enough.

At best, it will be hit or miss on finding stuff 'that matters to me’ and it absolutely isn’t 'personalized’.

To be personalized you’ve got to actually get to know and understand me first (hint: start by paying attention to what I say, do, and engage with…not just who I randomly choose to follow or who the people I follow, follow).

Once you’ve got a reasonable idea of who I am, and what topics I might actually be interested in, then *maybe* you’ve got a chance to help me discover content (and people) that matters to me. Until then, you’ve got a gamblers chance at finding me good content…and even less of a chance at keeping my attention long enough for you to 'figure it out’.

2. Discovery should have more depth.

Twitter has got a ton of amazing assets they can throw at this problem…the two biggest being an amazing staff of talented/smart people AND gazillions of actual data points (not to mention a nice chunk of change in the coffers to help keep focus on the problem).

The problem is, everyone including Twitter is looking for the 'easy’ way to solve the discovery grail…but the big wins in this space are behind the hard work.

Can you imagine how amazing it would be if Twitter actually learned to properly classify it’s users (and then eventually every single tweet)? Even a generic level of classification could go quite a long way in improving the 'discovery’ space.

Once you’ve got classification, you can start to do grouping…and once you start to do grouping, you can start to help me dig deeper into the topics that I actually do discover and care about (ie. one-off articles here and there about randomly interesting things, is not discovery…it’s sampling).

Oh and by the way, true classification is only the starting line for 'good discovery’ (it speaks nothing of the details involved in the listening phase).

Since we are a ways off from this actually happening on any 'good’ level, I am willing to even accept a simple 'hack’ to start…use Twitter trends, hash tags, or even click data to determine what’s likely related to a given article.

Bottom line, if and when you do hit upon something personally interesting to me, help me go deeper (ie. build engagement)…but what you are doing right now is just making your noisy service even noisier to me.

3. No matter what Twitter ultimately offers up in the 'Discover’ realm, it will be a closed world solution…and this means ultimately, it can’t be truly personalized.

On any given social network, or service in general for that matter, I am but a shell of my true and whole self.

If you, somehow, merge them all into one big glob of data and look at it as a whole…you *might* have a shot at properly classifying me. And only by looking at the larger collection of data will you actually be able to determine what I have (and have not) already seen or been exposed to (properly dealing with the 'seen it’ situation is a massive key to good discovery everyone just sort of skips over at the moment).

4. Ultimately I don’t think the Discover tab aligns with what users currently want or expect from Twitter.

Ask yourself why Twitter cares about discovery at all? What does Twitter really want to be? Do they even know yet?

I love and use Twitter all the time. It’s a noisy and chaotic place (much like the internet as a whole)…but that’s part of what and why I love it. I don’t need or want Twitter to filter the noise for me (just like I don’t want Google to skew or personalize my search results).

I want them to focus on letting more noise flow through the system and making it easier and easier for me to put my own noise into the system. Please be a publishing and communications platform…not a destination and not a media company.

5. I think discovery is an important, and missing, part of today’s internet

As such, I think it’s got to be it’s own stand alone service to actually work. I don’t think tacking it on as a feature within a given service (no matter how awesome that service, it’s staff, and it’s data set are) is going to do justice to the problem.

Every site/service has search (and in many cases had it years before Google)…none of it compares to how Google solves the problem.

Eventually I envision every site/service having discovery features (in fact, many already do)…but none of them are going to be 'good’ solutions for what users really need or want when it comes to true discovery.

Note: Mathew Ingram* of GigaOm also posted a few thoughts on the Twitter Discover tab updates that I think are probably common-felt and worth reading.

* The same guy that for some unknown reason refused to ever acknowledge existed as an option in this space (though he did actually have an account at one time)…not that I’m bitter about that slight or anything. ;-)

This post has received 37 loves.


This is the personal blog of Kevin Marshall (a.k.a Falicon) where he often digs into side projects he's working on for and other random thoughts he's got on his mind.

Kevin has a day job as CTO of Veritonic and is spending nights & weekends hacking on Share Game Tape. You can also check out some of his open source code on GitHub or connect with him on Twitter @falicon or via email at kevin at

If you have comments, thoughts, or want to respond to something you see here I would encourage you to respond via a post on your own blog (and then let me know about the link via one of the routes mentioned above).