DEV Community

Cover image for When's the Best Time to Post on Dev.to?

When's the Best Time to Post on Dev.to?

Marc Nevin on April 04, 2020

Using Python to work out visualise the best times and topics on Dev.to Having an international community on Dev.to means that there are users a...
Collapse
 
helenanders26 profile image
Helen Anderson

Great post, I'm the SQL mod so really pleased to see the tag has lots of engagement.

I've found if I post at midnight New Zealand Time / midday UTC I get plenty of reactions straight away as it's first thing in the morning in the New York timezone.

Absolutely no data behind that, just my 'gutfeel', but it's influenced when I post.

Wow, I’m definitely going to read this when I get a chance. Skimmed it.

Initial thoughts are that it’s definitely affected by when we are awake and working because of how we schedule for Twitter etc. But we’re always evolving and modifying the process so this could change.

I can’t wait to read through this.

Collapse
 
m_nevin profile image
Marc Nevin

Thanks and about the tag, I'm still a little shocked if I'm honest!

Totally get that, most of my posts are around midday UTC (tried mixing this one up just to see), picked in the same way as you, just on gutfeel that aiming for the US would be best, nice to see the heatmaps slightly justified it now!

Love to see some stats but I'm assuming most of the community is US-based?

Collapse
 
helenanders26 profile image
Helen Anderson

Shocked about SQL? Hope that's in a good way :)

AFAIK the founders are all still based in the New York timezone. The team has expanded to all parts of the world and are 100% remote.

As far as users go, the data is only as good at what folks are entering in the free text box in their profile settings. Which can be anything they like, and also can be nothing at all. There was a bit of research done 18 months ago that shows this in action:

I found it interesting that when the Big Thread Badge was launched it took seven weeks for a winner to be named from North America. Another one of those gutfeel things with nothing to actually back it up but interesting nonetheless. Ben named it the Big Thread Olympics in several of his posts.

1 - Australia
2 - Nigeria
3 - Romania
4 - New Zealand (that would be me :D)
5 - Rwanda
6 - Germany
7 - USA
8 - Netherlands
9 - Iran
10 - Canada
11 - USA
12 - Germany
13 - India
14 - Germany
15 - USA
16 - Turkey
17 - Germany

Thread Thread
 
m_nevin profile image
Marc Nevin

Yeah! ... Sure :D

Thanks for the link, pretty interesting, suggesting the biggest portion are US-based kinda backs up what you were saying!

The Big Thread is interesting, might check if, like you - really cool btw, winners were posting for UTC Midday / EST Morning US? Wouldn't be hard, but might explain why took so long for a US winner if they're proportionally bigger!

Kinda hard to verify, wonder how we could prove it out?

Thread Thread
 
helenanders26 profile image
Helen Anderson

To really prove this out I would want to get my hands on the Dev Google Analytics/Big Query data. It's so much more reliable than people entering whatever they fancy into a free text field.

To add some more complexity to this there is a feature which recommends more active users as 'Devs to follow' when new users sign up. So potentially the Devs who bubble up to the top of the list will have more followers who could interact with their posts.

Thread Thread
 
m_nevin profile image
Marc Nevin

Very true, be interesting to find out, that said I've limited exposure to using it, bar short access to my company's GA, I've never really got to take a look at a properly data-rich page!

Ohh I've definitely seen the suggested follow in effect, biggest follower gain I've seen probably came from this, but noticed some users commenting on a lot of their followers are very inactive... longer term I think it would lead to an increase but probably less than that from some of the other factors we mentioned?

Thread Thread
 
helenanders26 profile image
Helen Anderson

There was a discussion in another post about those inactive accounts being potential spam/bot accounts. However, what I've found (more 'gutfeel' stuff) is that the 90:9:1 rule rings true. 90% will lurk and just view, 9% will comment or add reactions some of the time, 1% are active in the comments and posts. I was a lurker with no info in my profile for well over a year before posting anything.

Thread Thread
 
m_nevin profile image
Marc Nevin

Very good point - pretty sure I did the same, lurked then moved up to comments/reacts, then posts, be interesting to know what the average conversation rate/time is..?

Could try:

  • User endpoint -> capture dataset
  • Use articles data set + Users -> associate posts w/ user
  • Use 'date_joined' compared to article post dates to work out the conversion
  • Bonus points: Get comments for each post too -> compare too

Probably deserving of an IP ban at some point through the user capture though I imagine! :D

Collapse
 
m_nevin profile image
Marc Nevin

Thanks!

Location isn't something I've looked into yet but Helen shared this in the comments, hopefully, answer any location-based questions for you!

Collapse
 
thomashighbaugh profile image
Thomas Leon Highbaugh

Answer to the title: after going on Devrant and letting off steam

Collapse
 
m_nevin profile image
Marc Nevin

Sometimes you just gotta ;D

Collapse
 
emma profile image
Emma Goto 🍙

Nice post! I've been publishing my posts when I wake up on Friday (which is around 7 or 8pm UTC on a Thursday). Maybe I'll experiment with publishing right before I go to bed instead.

Collapse
 
m_nevin profile image
Marc Nevin

Thanks! It's probably worth experimenting, never know if it's going or work or not til you try!

I was nearly automatically posting at 12UTC on Monday every day but this has made me mix it up more

Collapse
 
rhymes profile image
rhymes

This is absolutely fantastic, thanks for doing this analysis.

Also, Python for data science FTW 😂

(cc. @ben )

Collapse
 
m_nevin profile image
Marc Nevin

Thanks! And yesss Python is my goto, but R is pretty good too!

Collapse
 
potentialstyx profile image
PotentialStyx

Cool data, I might have to start drafting posts and keeping them until the best day/time