Fridays and Sundays Are the Worst Days to Publish
Read the small print
If you follow me here, you probably already know that I love to analyze data to see if I can find an edge. A few days ago, I read a great article by Cedric Boogaerts in which he concluded that publishing on weekdays had a huge benefit over publishing on weekends.
I decided to dig into my data to see if I could find any pattern that showed which days are best to publish on.
Data deficiencies
Before I share my results, you should be aware of the following deficiencies in the data.
- The sample size is too small to be statistically significant.
- I only analyzed data for August & September 2022 — I planned to analyze all my data, but it’s far too much work.
- I published in different publications — the differing number of views per story could be down to that and not the day of the week.
- The time of day isn’t taken into account — publishing one minute before or after midnight is essentially the same, but will show as different days.
- This is for my own amusement — don’t take it too seriously.
- There are other deficiencies in the data.
If you think the above makes the results unreliable, you’d be right.
The results
The number of stories published on each day was as follows:
- Mon — 10
- Tue — 2
- Wed — 6
- Thu — 8
- Fri — 1
- Sat — 9
- Sun — 3
I excluded one outlier from the data. This was an article that got 9,500 views and was published on a Sunday. Over 90% of views were external. This one story would distort the data. I’ll come back to this later.
The average number of views per story were as follows:
- Mon — 250
- Tue — 169
- Wed — 245
- Thu — 195
- Fri — 90
- Sat — 165
- Sun — 89
Based on the above figures, my data suggests that the best days to publish are Monday through Thursday. Saturday is the next best, with Friday and Sunday the worst.
Average views for Mon-Thu and Fri-Sun were as follows:
- Mon-Thu — 226
- Fri-Sun — 146
This means that stories published on Mon-Thu got 55% more views than those published on Fri-Sun.
A complication
Analyzing data is never easy. As well, as the data deficiencies already mentioned, here’s another one.
One of the stories published on a Thursday got 80% external views. If I exclude those, then the average views for Thursday drop to 108. The average for Mon-Thu drops to 199. That still makes the Mon-Thu figures 36% better than Fri-Sun.
But here’s the BIG complication.
If I included the story that got 9,500 views but only count the internal views (739), it increases that Sunday average to 305. This makes it the best day of the week to publish, based on my data.
It also increases the Fri-Sun average to 191, which is pretty similar to the Mon-Thu average of 199.
So, what does that all mean?
That’s a good question that doesn’t have a definitive answer.
I would summarize it as follow: Monday through Thursday are reliable days to publish. You generally know what’ll get. Friday through Sunday tend to be more hit-and-miss. Most story views seem lower but you can make up for that by a story going a little viral.
I think that for stories that get many thousands of views, the day you publish doesn’t matter. This is because the bulk of the views can come weeks or months later.
For the average story that gets most of its views in the first few days, I’d guess that publishing Mon-Thu is a better bet.
Will I be changing the days I publish on?
Not at the moment. I’m going to see if a similar pattern holds true for this month. I’ll then decide what to do.
I doubt I will change anything though. There are simply too many variables to be able to make an informed decision. The biggest factor that affects the number of views is the story itself, especially the headline.
There is more to be gained from writing great stories with great headlines than there is from worrying about which day of the week to publish. If you publish in publications, you can’t usually control the day of publication anyway.
I’m submitting this story to a publication on a Sunday. It might get published today or tomorrow. I have no idea if that will make a difference to the number of views or not.