Is Activity on Steemit Slowing Down? The Truth Behind the Data - Steemit Business Intelligence
There has been much talk among my Steemit peers that the platform seems a little quite the last two to three weeks. There is a feeling that the activity levels have dropped off somewhat.
Other authors have touched on this in the last few days such as @inquiringtimes with his article 'Are you experiencing Lower Pay on your posts? Is someone picking on you?'
And with @mikepm74 his article 'Has Steemit been slow lately? If so, Why?'
I myself was feeling the same with regards to my posts. I spent 7 hours on an analysis on Sunday which is included in this post, and it only earned $5.12. I'm very surprised with the low engagement and payout on this post. Feel free to check it out 'Positive or Negative? The Data behind the Steemit Trending Page'
For those of you that know me, I don’t normally complain about pay-outs, but when I see my own dropping steadily, then I have to do something.
That something happens to be analyzing the data. With all of this in mind, I have had a good look at the data on a week to week basis from the start of 2017. There was quite a lot involved in this analysis, but I really wanted to see as much activity as I could.
To carry out this analysis, as always, I have connected Power BI to Steemsql held and managed by @arcange
General Activity and % Change on Previous week
Below are 6 charts. Going down the left column of charts first we have
No of Comments (bars) % Change on Previous weeks (line). The number of comments per week peaked in week 26 and has been on a slight downward trend since
No of Posts (bars) % Change on Previous weeks (line). The number of posts also peaked in week 26, however there is a rather consistent posting trend each week.
No of Votes (bars) % Change on Previous weeks (line). The number of votes made each week has been increasing since week 18
In the right column of charts we have
Total No of Transactions (bars) % Change on Previous weeks (line). There is a slight upward trend in the total number of transactions taking place each week on steemit
No of New Accounts (bars) % Change on Previous weeks (line). The number of new accounts peaked at week 24 and has been on a downward trend since then.
Total pay-out – posts and comments (bars) % Change on Previous week (line). Total pay-outs peaked on week 25 with a sharp decline to week 29. From here the decline in pay-outs has continued but at a much slower rate. The trend is remains a downwards trend.
% Changes on Previous weeks
Plotting the % change on previous week for Comments, No of Votes, Posts and Total transactions it is clear to see that the % changes are very much in line with each other. However with the number of new accounts the % change was off the scale and is plotted in the second chart below
Some averages
The first chart below, plots the number of Posts (bars) and the Average comment per post. We can see that although the number of posts is rather level each week, the average number of comments per post has reduced from 5.5 in week 24 to only 4
The average number of votes per post peaked in week 8 at almost 20. This dropped to 2.5 in week 26. The trend has since changed to a more positive trend and is now about 4 votes per post.
In the chart below we can see the trend on a week by week basis for both the average number of votes per post and average comments per post. It is interesting how these have almost come into line in the last couple of months.
As the number of votes are on the increase, but the average number of votes per post is way down on its peak, it would suggest that curators are voting for a wider base of authors.
Year to Date Curves
The first chart below plots the no of comments year to date (ytd), no of posts ytd, no of posts and comments ytd and the number of new accounts year to date. I found the steepness of the curve hard to judge, so in the second charts shows the same data using log instead on linear.
Now it is clearer to see that between week 19 and 25 the growth was at a faster rate
Presenting this data in a slightly different way it is clear to see the growth in new accounts and total number of votes are closely correlated.
What we know so far
The number of posts and votes weekly are on an upward trend. The number of comments left on posts in on a downwards trend. People are posting, but less people are commenting on these posts. However there are more votes – does this suggest less human activity (i.e. reading and engaging) and more bots?
We also know that the rate of growth has slowed down since week 26. And that the average number of comments, votes and pay-outs have decreased per post.
So on a per post basis things are looking worse, hence the downturn I have experienced. But the overall activity on steemit is on the up.
Pay-outs
Pay-outs seem to be the biggest concern, and the ratio for the drop in votes to pay-out on my posts don’t really correlate with all of this data and I think there is a little more going on with the pay-outs and vote value.
In the first chart below we can see the average pay-out per vote peaking in week 24 at 66 cents. This has declined to 11 cents. This is a reduction of 83% on peak
There are a few things that impact the value of a vote, the current price of Steem, the SP held and the voting weight. The second chart below shows the average voting weight per week. The highest was at the start of the year at 81%. Overall the trend is downwards and we are now looking at 63%. This is a reduction of 22%
Looking at the charts below, first we have the average pay-out per post plotted with the average voting weight. Over the last 12 weeks the fluctuation in voting weight is not correlating to movement in average pay-out per post. This would suggest that there is a reduced SP holding.
The average pay-out per post peaked at $16.25 in week 22 and is now at $2.96. That’s a reduction of 82% on peak
The second chart shows the average pay-out per comment and the average voting weight
Finally I have plotted the total posts pay-out with the average pay-out per vote and the total pay-outs on comments with the average pay-out per vote.
It is very clear to see from all of these visualization that what people are earning now per post, is well well down on the peak. Over 80% down. The average voting weight has decreased 22% on peak but we know the number of votes has increased. So this does not account for the drop.
Steem Price
The Price of Steem and Steem Power are also contributors to values of votes and posts.
Below is the Steem close price chart for 2017. The peak is in week 23. However from mid-july to present the price of steem rests between $1 and $1.50 with the overall trend being a downtrend.
Conclusion
So there you have it. In general Steemit has not slowed down, although there is a question over human interaction over bots. It would appear to me that human interaction is reducing and bot activity is increasing.
This would explain why the platform appears to be less active. For those of us that joined steemit while the price of steem was over $1.50, obviously we are feeling the lower payments. But rest assured this is mostly down to the price of Steem.
People that joined while the price of steem is close to $1 would not feel this as much, in fact every time it rises to $1.50 they are feeling the gain.
If you feel like you are getting less votes and comments per post, well you are. As more people join steemit, votes appear to be spread to a wider audience, however it is not the same story with comments.
I am part of a Steemit Business Intelligence community. We all post under the tag #BIsteemit. If you have an analysis you would like carried out on Steemit data, please do contact me or any of the #bisteemit team and we will do our best to help you...
You can find #bisteemit on discord https://discordapp.com/invite/JN7Yv7j
This is a very detailed post and well done. One thing that I did not see mentioned is the technical difficulties that many users have been experiencing over the last 2-3 weeks in submitting, commenting and upvoting posts. This may also be part of the reason why human activity is decreasing (at least for the timeframe this has been occuring).
I think you make an excellent point here. Rest assured, many of us are working on projects that will hopefully make the Steem experience much better.
i totally agree with you @boodles17 !
I have to admit, i have not been as active as usual. Im also thinking that maybe US based user activity will kinda slow down during the holiday season coming up soon...
Im am from Ireland, didnt think of holidays
I rarely ever visit the Trending page. I tried one bot once and it didn't work lol Sometimes I'll miss some posts from people, but I'll go back in time or upvote the ones I see when I'm able to take the time and check out their content. Human activity is inconsistent but as opposed to bot activity, it's real. So I'm ok with that. I did notice a bit less of payouts recently also, but it was Thanksgiving here in Canada, so maybe people just weren't able to get on Steemit as much as usual. I wonder if holidays affect this. I guess we'll have to see.
About the Trending page, I got on there twice only, and have no idea why or how, I think maybe a dolphin saw my post and upvoted me. But I think it's fine if it helps promote top posts, but there could also be a Monnows page, where it's some of the more interesting posts from monnows, maybe a post that earns over $5 for example, so that it's not too saturated a place. Dolphins and whales can go there and monnows too, and we can check out these posts and give upvotes, and those posts could end up ont eh trending page. It would be the in-between, the before trending, not yet trending, but sorta trending. And would give some of us who have posts that do wel every now and then to do even better and make even more and gain more traction with more steemit users. What do you think?
I think that is a very good idea. or a personalized trending page for everyone
That could work too. People could select their interests, such as gaming, crypto, etc. And that would come into play for that.
This morning, I sent a message to @arcange, which begins as follows: “Here’s a sad scene for you. About 4 people have read this post I made about 3 days ago: “Can you supply some block-chain data for payment?”. Your post suggests that there must be or than 20 people in the community who might have been alerted to my post, and some would have promptly sent me to your service, which clearly needs to be well supported. New Steemit posts disappear into ‘ether’ [after] 20 minutes, and will turn up on peoples’ Feeds only if they are Following the authors!”
As someone above has said, communities of interest need to be identified, and posts related to Community-of-Interest-X need to be advertised among those members, regardless of who is the author.
Sorry to dwell on this; but it is a key problem. SMTs may solve it; but that process will be very slow in the best of all worlds. It would be a lot faster just to fix the Steemit software; but we can see that the economics of the latter are less compelling for Steem Inc.
Here’s my ‘two cents’, along with my big “thank-you” to the author.
We are now just over a week past the post date, and the payout is $35! Everyone should hang their heads in embarrassment on seeing $35 offered for work of this quantity and quality. So let us all get on our knees and pray that this wonderful author is not even thinking about payout levels here at Steemit, and has succeeded in finding other more helpful rewards from her contributions.
It appears that these days, on average, that your curation/commenting work will be only slightly less “rewarding” than your posting work. That, folks, is very bad news if we expect to see a lot of people coming in with valuable ‘green’ content that takes many hours to produce (like this post). I am ready to bet $100 (fiat money) that the rate of flow of this kind of content (standardized for new-member growth somehow) will go down dramatically as we go forward.
Even $200 on a post I might make that is ‘fully green’ and brings unusual value to the table after three-plus days of data mining etc. is not a great motivator for many authors -- nice but not great. Here’s my point. The designers of the system need basic re-education on what drives authors of difficult ‘green’ content (often with inherently small audiences) to bring their stuff to the table.
Let’s add these observations, so that we all look at the problem squarely in the eye. (1) A YouTube giant came in here with his followers and on his first post, containing little ‘green’ content, he reaped a $15,000+ payment. (2) Another YouTube giant recently arrived with a paraphrase of five short paragraphs from an article that he published in Wired, and within an hour his expected payout and reached close to $700. I could not find anything particularly ‘green’ in the content of those five paragraphs. (3) Another YouTube person reposted her video explaining block chains along with a few expository paragraphs, and within an hour the payout exceeded $130. There are now tons of articles and books on block chains out there in the wild. (4) Yesterday in an interview posted on YouTube, one of the lead engineers in developing Steem declared that our voting system was being ”abused” -- that was the word he used.
Thank you for such a detailed response. I wish I did find a way to be making more $$$ from my posts, but I think saying 'come for the money, stay for the community' holds true.
I just had a quick look at your post. I will come back to you later today with a reply (added to list)
I feel your pain regarding steemit. I too am not sure smt will solve the problem. and from analysing the data, there are a lot of problems.
Thank you again for your comment, I upvoted with my 5 cents
Thanks for your helpful comment and vote, Paula. It was clear that “dollars to be received” was not the force driving you to put together such a long and extremely useful piece.
I appreciate your offer to come back to me about my post. However, I do not want to go ahead with the work that I had in mind; because the conclusions that I thought I might reach and which might be helpful to authors can be dug out of your data and text, along with the follow-up comments.
The only important finding not stated in this web page (as I recall), and which emerges from the work that I had begun in another related post, is the strong suggestion in the data patterns that (a) certain classes of content tend to be significantly better favoured than others in terms of payout level, and (b) within each class you should expect a substantial payment mainly after you have built up a following large enough to bring you close to 200+ votes soon after your post (and probably including at least one whale).
Anyway, I’m going to focus now on discovering the people in my communities of interest and finding a way to be happy with their company. In the meantime, though, I will keep following you and pay attention to your articles.
Please do join us on discord anyway, you might just find other members on steemit with similar interests. There are a number of noted communities on this server, and if you can not find one, we can help you set one up. There are good people here on steemit :-)
BTW Paula, I just saw another author include in his post his Public Key addresses for BTC, ETH, LTC, DOGE, STEEM, etc., while suggesting reader donations. I asked him if he was getting flak for doing so, and tells me "no" and that he sees some whales doing the same.
I think it is a fabulous idea for people like you and I who will take many hours/days to produce stuff with 'green content' and good value for use by a relatively small audience of specialists. I strongly advise you to do that, so that we who understand what sort of effort, knowledge and skill are behind your output can take action to try and make up for the ridiculous "payment" you are getting from the system. Pl. note, that I was already offering to pay people directly for their work to help me in my projects, and that is precisely because I find the payment levels available to us meaningless!
that is very interesting, and worth considering. I have been working online 7 years now @lestatisticien. I am on steemit only a few months. I will tell you my growth on steemit has been more rapid that many other platforms. I live in Ireland, posting on steemit does not pay to make a living here. But look at other countries where $10 a day makes a big difference. That is where steemit can benefit.
but the cream goes to the top - this is annoying too (very)
human activity is reducing 1) the ddos attack and slow response and low communication to members about how things are proceeding/improving has hurt 2) yes, the price is down 3) there are very good curation projects, but good hearted people do great work promoting posts but no one really looks at them anyways. 4) steemit is too confusing for new users 5) there are too many sites / tools / off-shoots... it needs to be way more intuitive
steemit still has the chance to have the "it" factor, but i see currently everyone is hoping SMTs will empower STEEM... i have no idea. but overall on the platform there are a lot of good people, good vibes, and good world outreach but it's all disjointed... we need to move to the next phase soon.
I think they should focus on getting steemit right, and that smt is just adding more when the basics are not ready
I agree with you 100%. I would encourage you and some of your smart business intelligence friends to post on this. I believe too many of the experienced users are avoiding the issue and just navigating via other sites. That is a geek solution but not catering to the average user / new user. You can't make Steemit great if it's not EASY TO USE AND INTUITIVE. I made two flippant posts on this issue recently, but we need more credible people to speak up quick. Peace
Well said, exactly what I was thinking.
While I like the simplicity of Steemit, I need some content tossed into my feed by people who I don't follow to spice things up a bit. I try to look at the new posts, but It's too time consuming with everything else going on.
Outstanding data analysis Paula!
This is great analysis! He said as the resteem button spun and spun.
It becomes increasingly difficult to find good content in any of the feeds as it gets mucked up with garbage posts, so I imagine curators are having a difficult time of it. Going through several tags today and just browsing the posts, I could see why bigger fish would want to creep into a cave and hide.
I wish I understood this stuff half as well as you Paula! Great work, thanks for sharing!!
Well done, I think the data shows that the real decline in payouts has obviously come from the decline in the Steem price. I'd also wager that the decline of comments and increase in upvotes is largely due to bot activity.
Since we have an increase in posts and decrease in comments another possibility could be that the post quality is in decline and not engaging enough to be worthy of commenting on.
This was a great post! Thanks for taking the time to put it all together and crunch the numbers. Definitely worth an upvote, comment and resteem.
i was also going to guess you were an accountant before reading your profile. :)
You LOVE DATA! :)
i would tend to agree w @boodles17 . Tech difficulty has probably pushed ppl away temporarily.