I have just been having a conversation on a post about how DEV decides what content to promote, so I thought I would ask the question instead of speculating!
- What dictates what is promoted on social media?
- What dictates top comments in the newsletter / weekly post?
- What dictates a top 7 post in the newsletter / weekly post?
For clarity: I am talking about things that the DEV team actively promotes, not what is popular in the feeds or how the feeds work.
I know they say in the top 7s that it is based on traffic or engagement but that doesn't appear to be the case (although I can't see traffic figures to be fair).
They do seem very much human driven (which is fine but then I am asking how are they chosen)?
Just wondered how it all worked and what is actually looked at as the criteria?
Is there a bit of luck involved as to which posts the DEV team happen to read, as I understand you can't possibly read them all?
I am intrigued so I can learn how to game the system write better content that gets promoted šš¤£
Personal motivation
This question was sparked as I released a monster article (one that I believe is truly great quality...for once I didnāt shit post š!) that got great engagement but low views. (currently: 224 reactions to 2500 views)
As an experiment I deliberately wrote a "low effort" post that I released at one of the worst possible times that was designed to be a click magnet / get noticed and....the "low effort" one got promoted! (currently: 55 reactions to 4400 views)
Not complaining about getting a shout out (thanks DEV) it just seems a bit broken that I can predict what is likely to get shared based on the tags and title rather than the quality of the post!
I know that increasing visibility of quality content is something the DEV team is working on, so perhaps a final question is how you find that quality content in the first place (suggestions are: "reactions to views" ratio seems like a decent starting place, maybe "number of comments to views" too?)?
Anyway just asking the question, and from a purely selfish standpoint (until people get sick of me promoting it!) here is the article I believe was "promotion worthy" and the one deliberately designed to get promoted!
the "high effort" post
Article No Longer Available
the "low effort" post
Article No Longer Available
Thanks in advance for any clarity on this!
Update
A great reply from @graciegregory covers all of the points in detail and beautifully, thanks Gracie!
I have included it here in case it gets lost in the comments, it should, in my opinion, get added to the FAQs section as it is that comprehensive and I am sure I am not the only one who wanted to know this!
Hi @inhuofficial .
First off, I wanted to congratulate you for this post. The content team found it to be informative, handy as a reference, and well thought-out. I know youād have these questions whether publicly acknowledged your post or not but wanted to say this up front. Spoiler alert: while we havenāt published the Top 7 yet, we will shortly and your post is one of the articles chosen. I also want you to know that we were not pressured to pick your article because of this discussion. We were planning to pick it yesterday.
Now, on to your questions...
What dictates what is promoted on social media?
We have a team of social media content specialists that scan tag pages and the home feed and then primarily pull quotes from the articles chosen to draft and schedule social posts. The entire goal here is to promote our community members in their words much more often than we promote DEV as a company. Our social content specialists are careful to do what we call ābalancing the feedsā. This means that while they certainly look at popular posts (towards the top of dev.to/top/week), they are also careful to ensure we arenāt only promoting authors within one community, gender, race, or one subject.
What dictates top comments in the newsletter / weekly post?
Hereās how the process of creating the top 7 usually works:
We begin by looking at dev.to/top/week to see what articles are getting a lot of attention and are particularly ābelovedā that week.
We scroll all the way through and (in the very least) scan all of the posts we hadnāt already come across the week prior.
We take a look at particular tag pages like #a11y, #mentalhealth, #career, etc and select posts that are fantastically written, thoughtful, or helpful but might not have gotten picked up from our ātop articlesā algorithm/feed
From there, we hone in on a list of articles deemed to be the āTopā 7 according to the community in general and our content team. Similar to our social media feeds, we use a great deal of editorial curation to comb through our options and pick a list of authors and subjects that are diverse in every sense as much as possible.
I will also say that last week, we actually changed the name of the āTop 7 Most Popular [...]ā to be the āTop 7 Featuredā to better reflect this editorial process. Weāve been doing this series pretty much since DEV began, so it felt like it was time to update the name to match the way the volume of our content (and therefore, algorithm and curation process) has changed.
Is there a bit of luck involved as to which posts the DEV team happen to read, as I understand you can't possibly read them all?
There is definitely some luck involved here. We definitely canāt read everything but I will say that we are very sensitive to the fact that a lot of listicles get posted on DEV, some of which are low quality. Many folks (including members of the social media/content team) find particular listicles helpful so this is kind of subjective in terms of the curation process but weāve been having a lot of conversations about doing a better job of being extra careful not to promote things that other team members feel are low effort lists on social media and in the Top 7/newsletter.
Timing of Top 7 candidates
We pick Top 7 content from the previous Monday through Sunday without any weight given to a particular day of the week.
Is there a way to promote a particular post that I want a few more eyeballs on via listings or some other method?
We used to have a āSuggest a Tweetā feature but, very recently, we got rid of it because it needs to be rebuilt and it wasnāt very usable for our social media team. While we consider a new method for getting your work in front of our team in an alternative way, you can certainly use DEV Listings to share your post. A lot of folks do this!
Comments/questions about reading time and article length.
Everyone comes to DEV to find different types of content. Some come here to find helpful resources to quickly reference ā like you mentioned. Some actually want to find more longform articles. I will say that at DEV, we are always excited when we come across a fantastic long read, essay on the history of a particular technology, or in-depth explainer. Itās true that these kinds of posts donāt always pop off the way a listicle with buzzy words in the title get and thatās something we should address. One thought is for the DEV admins to create something like a #longform tag (or some kind of curated editor's tag) that folks could use to indicate these types of articles. Our content team could reference this when creating the Top 7 each week.
I hope these answers shed some light on your questions and concerns. DEV as a platform is not perfect and never finished. Weāre always striving to improve every part of it, including the feed and our editorial curation process. We are listening around the clock and appreciate constructive feedback, always.
Top comments (33)
Hi @inhuofficial .
First off, I wanted to congratulate you for this post. The content team found it to be informative, handy as a reference, and well thought-out. I know youād have these questions whether publicly acknowledged your post or not but wanted to say this up front. Spoiler alert: while we havenāt published the Top 7 yet, we will shortly and your post is one of the articles chosen. I also want you to know that we were not pressured to pick your article because of this discussion. We were planning to pick it yesterday.
Now, on to your questions...
We have a team of social media content specialists that scan tag pages and the home feed and then primarily pull quotes from the articles chosen to draft and schedule social posts. The entire goal here is to promote our community members in their words much more often than we promote DEV as a company. Our social content specialists are careful to do what we call ābalancing the feedsā. This means that while they certainly look at popular posts (towards the top of dev.to/top/week), they are also careful to ensure we arenāt only promoting authors within one community, gender, race, or one subject.
Hereās how the process of creating the top 7 usually works:
We begin by looking at dev.to/top/week to see what articles are getting a lot of attention and are particularly ābelovedā that week.
We scroll all the way through and (in the very least) scan all of the posts we hadnāt already come across the week prior.
We take a look at particular tag pages like #a11y, #mentalhealth, #career, etc and select posts that are fantastically written, thoughtful, or helpful but might not have gotten picked up from our ātop articlesā algorithm/feed
From there, we hone in on a list of articles deemed to be the āTopā 7 according to the community in general and our content team. Similar to our social media feeds, we use a great deal of editorial curation to comb through our options and pick a list of authors and subjects that are diverse in every sense as much as possible.
I will also say that last week, we actually changed the name of the āTop 7 Most Popular [...]ā to be the āTop 7 Featuredā to better reflect this editorial process. Weāve been doing this series pretty much since DEV began, so it felt like it was time to update the name to match the way the volume of our content (and therefore, algorithm and curation process) has changed.
There is definitely some luck involved here. We definitely canāt read everything but I will say that we are very sensitive to the fact that a lot of listicles get posted on DEV, some of which are low quality. Many folks (including members of the social media/content team) find particular listicles helpful so this is kind of subjective in terms of the curation process but weāve been having a lot of conversations about doing a better job of being extra careful not to promote things that other team members feel are low effort lists on social media and in the Top 7/newsletter.
We pick Top 7 content from the previous Monday through Sunday without any weight given to a particular day of the week.
We used to have a āSuggest a Tweetā feature but, very recently, we got rid of it because it needs to be rebuilt and it wasnāt very usable for our social media team. While we consider a new method for getting your work in front of our team in an alternative way, you can certainly use DEV Listings to share your post. A lot of folks do this!
Everyone comes to DEV to find different types of content. Some come here to find helpful resources to quickly reference ā like you mentioned. Some actually want to find more longform articles. I will say that at DEV, we are always excited when we come across a fantastic long read, essay on the history of a particular technology, or in-depth explainer. Itās true that these kinds of posts donāt always pop off the way a listicle with buzzy words in the title get and thatās something we should address. One thought is for the DEV admins to create something like a
#longformtag (or some kind of curated editor's tag) that folks could use to indicate these types of articles. Our content team could reference this when creating the Top 7 each week.I hope these answers shed some light on your questions and concerns. DEV as a platform is not perfect and never finished. Weāre always striving to improve every part of it, including the feed and our editorial curation process. We are listening around the clock and appreciate constructive feedback, always.
I hate sharing my work on Twitter, so I just wanted to say I'm super grateful that you are doing it often enough for me.
p.s. the #longform tag is a great idea.
One idea I had was that mods and DEV team have a set number of "super votes" each week that add certain articles to the pile to be looked at.
It could also have a secondary effect of adding a boost to that post in the feeds.
Finally "super voted" posts could have a slightly stronger boarder / some other feature to make them stand out as "DEV picks". This is an alternative idea to the pinned post experiment you ran, without it always being top of the feed?
Anyway it was just a thought as I am excited by the fact you are putting more into the curation aspect!
Thanks Gracie, you answered everything perfectly!
I edited the article to include this comment - I did say I thought this comment should be turned into part of the FAQs or something as it explains how DEV picks things beautifully and I am sure many others would have similar questions (but aren't loud mouths like me so don't ask š)! ā¤
I'm glad to hear that everything you were wondering about got addressed!
As for the FAQ section, thanks for the suggestion. We will take it under advisement š
The longform tag would be a great addition to see!
I think (based on my observation) that the "reading time" seems to be important as well. articles with less than 3min are almost never picked in the social media. From my 9 "100 collections" only 1 was picked while other articles that I consider less effort are picked because they have more content.
I also found some poor listcle there: twitter.com/ThePracticalDev/status... .. I got surprised and asked the Team for the "why"
Ah that is an interesting point on the reading time! In all fairness to that listicle it isn't a complete low effort post, they did at least use proper headings and insert screenshots! š hehe.
Interesting that only one of your CSS posts got picked...that makes me think that tags do indeed factor into it quite highly, I would have thought a popular series like that would have been ideal for twitter (I am especially surprised as they were all pretty high view counts!) - out of interest which one got the Twitter boost?
Anyway, I was trying not to speculate as I am sure I will get it wrong. I don't want to @ any of the DEV team but I do hope one of them pops up and lets us know what the rules / criteria are!
The first one got the boost: twitter.com/ThePracticalDev/status... and the second one got the Top 7 (but no social boost) dev.to/afif/another-100-underline-...
Interesting as I would have expected the top 7 boost to result in more engagement than twitter.
More questions than it answers now! I know the stats on posts aren't great but did you notice a significant boost from social media? Or was it just that that particular post was just so unique and high effort at the time that you think everyone decided to read it (as when I first saw that post I did think "better up my game" as it seemed to be much higher effort than most posts at the time!)?
I would say yes because the two collections that got the most reactions/views are the ones shared the most on social media (not necessarely by DEV)
https://twitter.com/search?q=https%3A%2F%2Fdev.to%2Fafif%2Fi-made-100-css-loaders-for-your-next-project-4eje&src=typed_query
https://twitter.com/search?q=https%3A%2F%2Fdev.to%2Fafif%2F100-underline-overlay-animation-the-ultimate-css-collection-4p40&src=typed_query
Well I have been putting it off, but I think it is time to return to social media, 10 years without using it properly (I was ahead of the curve realising it was bad for you!)...but business over personal interest, hopefully I can pay someone else to manage it pretty quickly so I don't get sucked back in!
Super useful, the last kick I needed to pull my finger out, thanks as always!
I have been fortunate enough to have been promoted on social media 5 times, I want to make it clear that I am grateful for that! Thank you once again!
I just wonder if certain tags and post types are more promotion worthy as it does make a difference to views and I don't mind writing posts that get promoted so I can add links to my more serious posts within my "crowd pleasers" and give them a little boost!
Speaking of "a little boost", I have one last question, I am fortunate enough that DEV has given me a few credits for listings, is there a way to promote a particular post that I want a few more eyeballs on via listings or some other method? I couldn't see a category / way to do that but it could just me I am missing it!
Nice timing! I'm taking an analysis on this (by tracking top 7 posts, bla bla). So far, I found a few things:
I think we are crossing from what DEV picks to what is popular in the feeds here. I am talking about the stuff DEV puts in the weekly newsletters, on social media etc.
For where things appear in the feed you can examine the source code of Forem on GitHub. github.com/forem/forem/blob/c842d1... might be one place to have a poke around and explore from there!
I look forward to seeing what you come up with for working out which posts gain natural traction etc. I would suggest looking at times they were posted using the API and which tags they use in addition to the analysis so far.
inhu.co/dev_to/analyse/timeofday.php may be useful (the table at the bottom) as it is a very rough analysis on what time of day posts were published vs number of likes (with the top 5% and bottom 5% of posts removed to avoid the skew). It doesn't take into account when popular authors post etc. but hopefully it gives you some more information to work with! The data is based on every single live post on DEV (about 3 months ago, I need to update the data).
What DEV picks and places in social media and stuff can't be analyzed properly ā It can only be tracked. My best guess is it mostly contains stuff geared at beginners.
That code seems intresting... I'll surely check it out!
As for what I'm tracking, I plan to track stuff in the feed (in an incognito window of course), the weekly newsletter, and some other stuff using the APIs.
I remeber that post which you're talking about. It's some intresting (but not the most useful) data. I'll check that out too
For a good idea of popular posts, you could go to the inifinity tab
Added a bit for clarity as I obviously wasn't clear (my fault I reread the first part and can see how I worded it badly! š¤¦āāļø):
Thanks for helping me realise that I needed to make it more clear to avoid confusion!
Super interesting topic!
I don't really have any concrete answers, but I can share my 2 cents here:
There is a week gap between the two, but when comparing both posts after the first 24hrs - the first one got much more attention. I believe it is related to the target audience (which is much broader than just javascript/react developers).
Great stats there. (Still not quite the point of the article as I want to see why DEV picks the things they do) but interesting.
Your "set up a new Mac article" hit one of the magic numbers, 100 reactions (roughly). Once that happens it is high enough up the week feed etc. that it continues to gain readers for several days after. Initial traction is what matters for the feed.
There are loads of little things like that (a snowball effect if you will, hit a certain reaction count, get more visitors, get more reactions, increase your chances of a share on social media, repeat).
One sneaky tip (if you aren't already) is the second you release your article give it a heart, unicorn and bookmark and leave a comment. It encourages people to click in top the article and the snowball potentially starts there.
I look forward to seeing more posts from you, the two you have released are well written!
Thanks for your feedback! I definitely plan to continue writing :-).
Here is another question - how are "weekly stats" calculated? Is it "Monday-Sunday" or every day a new "weekly counter" starts?
In my case - both posts published on Sunday. If this is the last day of the week - reactions/views on Monday will start reset, and there is less chance that those posts will get picked as "top posts" for that week.
What do you think?
Also - thanks for the tip. Do you comment regularly on every post that you write after publishing it?
Weeks do run Monday to Sunday for the purposes of the top 7 posts, I have seen posts published on Sundays make it onto the top 7 posts so I don't think it has a huge impact (who knows, another good question if anyone from the DEV team answers my question(s)!)
In terms of the week and month feeds those are done in days so it has no impact what day you post (assuming the exact same performance of the post...when you post does appear to have an impact on reach!)
Yes, I always comment under my own posts and I make an effort to reply to everyone as there is an element of the initial ranking on the feed that is to do with comments. Plus I always find the comments are where the really interesting parts of an article tend to be, not that it has anything to do with your question I just thought I would add that! š¤£
Haha š¤£
Yes, replying to comments is basic š my question is mostly about just commenting something (not replying to others... which is a must). It's an interesting approach!
As for the "top 7 posts" - really interesting, and it's indeed a good question.
The 68 minute read time might have something to do with it. I skip anything of that length regardless of title, tags, etc. I actually remember scrolling past your post the other day. I remember it not for the title or any other content indicator, but for the reading time.
If I'm going to spend over an hour reading something, it will be something like an article in a refereed journal, or a couple chapters of a book. If I'm browsing Dev, I'm not looking to invest that much time in one thing.
So you think DEV would deliberately not promote the post on social media due to the fact it is long and in-depth? That would be an interesting take on it!
As for personal preference on reading length, interesting that you would say that.
It raises interesting questions about why you would be willing to read it elsewhere but not on DEV?
Is that due to the fact you don't expect the content to be the same quality here (which would make total sense) or something else?
There can certainly be high quality Dev posts. But the type of thing I'd spend an hour reading is entirely different than what one would normally find on Dev, Medium, etc. If I'm browsing Dev, I'm spending at most 30 to 45 minutes total while drinking morning coffee or tea. I'm more looking for things along the lines of "I found this or that useful and why", concisely written, or ShowDev type posts showing off cool things people built. And in that 30 to 45 minutes, I want to check out several posts.
I don't know how the Dev team picks posts to promote. But I imagine that they probably look at reading time as one of several factors. I don't think they would necessarily skip the long posts. But I imagine that they would at least ask themselves the question "Why should I recommend reading this one hour post instead of multiple shorter posts?" before doing so. Additionally they would have needed to spend that time reading the hour long post themselves to begin with. And with limited time and a ton of content posted daily to wade through, something special would likely need to draw their attention to it.
I would hope that nobody reads a listicle like that in one sitting, but rather sees it as a reference piece. But it is useful feedback nevertheless that shorter articles in a series may be a better way to get exposure.
Also interesting that due to where it was posted you wouldn't even attempt to read it.
Not something I had considered but I suppose if you associate DEV with short content then long content isn't going to appeal.
It is interesting as I had never thought like that I just look at whether something interests me and either read it or bookmark it, but it is super useful to know people approach it that way!
Thanks for the thought provoking perspective, I will be taking it into account when designing larger posts!
68 minute read?!
LOL š That must be "the bible of coding" or similar
Nah, it was a listicle of course (no really!), isn't that the typical length of a listicle or am I doing it wrong? š¤£
That's sound very typical! šš
If you dare to read it rather than use it as a reference piece it will probably take about that long, 16500 words! I did say "high effort", I wasn't just talking about writing it šš¤£
This is a very helpful article that covers the essential skills and best practices for mastering frontend development. I appreciate that you explained the core concepts of HTML, CSS, and JavaScript, as well as the popular frameworks and tools that can enhance the frontend development process. I also liked your advice on how to write clean, responsive, and accessible code, and how to use debugging and testing tools. Thank you for sharing your knowledge and resources!