Is User Sentiment Up? Can Numbers Tell Us?

MORE THAN A FEELING

After reading a couple of recent posts and their comments—to wit, that thanks to newsteem (ie, the EIP and a general striving for a positive change in culture), things were looking up on the blockchain, user sentiment was up, etc.—I got curious about some numbers. Thankfully, penguinpablo has been publishing different data for a long time now, and it became the source material for this post.

I wondered if these numbers could be the reason for any good feelings about the current state of the blockchain. I wondered what the data could prove, disprove, and whether it would really do neither.

I honed in on three different categories:

Daily Accounts Transacting—unique daily users
Daily Number of Votes—posts and comments
Daily Number of Posts—along with the average number of comments per post

There are obviously more numbers to pay attention to, but these are the three that the aforementioned posts and comments by others got me interested in the most. I figured more accounts transacting, more votes and more posts would equate to a higher positive for user sentiment, while lower numbers would lead to the opposite. Even if within these numbers there were undesirable quantities of spam, vote farming, etc.

HOW TO PRESENT

Knowing how to group the data, how large of a sample to pull, etc., is always the fun part when trying to present it. I didn't want to go overboard with it, but at the same time, I wanted a fair representation in an attempt to avoid actual or even perceived cherry picking.

As to methodology, then, I started a year and a half out from today, and then chose periods of time before and after hardforks, particularly HF20 and HF21/22. I then grouped the three different categories for those specific windows in time together. Also curious as to what STEEM value was at any of those particular points, I then added in the ones which more or less corresponded with the timeframe.

The results are as follows:

DATA SET 1—MAY 1, 2018

For the week leading up to when this data was published, the total number of accounts transacting was mainly between 61,000-66,000, with the outliers on both ends being in the lower 58,000 and in the lower 67,000.

During that same period, the daily number of votes ran as low as 706,783 and as high as 779,964. Note the number of paid bot votes, and the percentage totals.

Posts were published at a robust clip in the range of 52,000 to nearly 56,500.

You'll note that this was the week following STEEM reaching its second all-time high, closing at the end of 4-23-2018 at $4.36 USD.

DATA SET 2—AUG. 31, 2018

Four months later, by the end of August, the number of accounts transacting had dipped into the lower to mid 50,000s, rather than the lower to upper 60,000s.

Daily number of votes dropped, too, with the amounts widening from a little over 70,000 vote fluctuation in a week to a little over 90,000. Low end came in at 631,804, with the high reaching nearly 723,000.

However, the category of the three that dropped the most precipitously was posts, falling to basically half of what they were four months ago. Note that the average comment per post held, and actually ticked up a bit.

This particular period of time was the last few weeks leading up to Hardfork 20, but it was also the last time STEEM value closed nearest $1.00 USD.

DATA SET 3—NOV. 1, 2018

A year ago, after most of the dust had settled from HF 20, the amount of accounts transacting, despite the introduction of Resource Credits (RCs)—which hampered most of our ability to do much of anything for a period of time before the allowance was upped by a factor of 10—was pretty similar to what they looked like before the hardfork, though trending down a tad.

Meanwhile, the number of daily votes declined a little more, dipping below 590,000 on the low end, while not quite eclipsing 631,000 on the high. You'll not that paid bot votes continued to fall, in total and in percentage of votes.

Again, posts took a greater hit, sliding another nearly 32% from just two months prior. Average comments per post, as you'll see, stay more or less the same in this period.

DATA SET 4—AUG 16, 2019

This next data set fast forwards several months to mid-August of this year. The significance of this time period is that it is just before what would be not only one but two hardforks, but Hardfork 21 being the one to bear the Economic Improvement Proposal (EIP).

Nine and a half months from the last data set, the number of accounts transacting on a daily basis has fallen noticeably, barely exceeding 43,500 while dipping below 40,000.

Along with it, the number of votes continue to decline, again a little more obvious with more time involved. Low end comes in at just short of 483,000, while the high surpasses 532,000.

Meanwhile, though a slower decline than the previous times given the longer time period, the number of posts per day has again, on the highs, fallen another 42%. Comments per post continue to hold, with a slight uptick.

STEEM value by this point has continued its spiral downward to just above $0.17 USD.

DATA SET 5—TODAY (NOV. 1, 2019)

This is the period in which we are now, a little over two months out from the latest hardforks and the EIP.

The number of accounts transacting on a daily basis continues to fall. The lowest day in the week, October 25, ended below 29,000, with the highest on October 30, a touch above 33,500.

Likewise, the daily total number of votes has dropped, at what appears to be its greatest total, from a high of 506,464 on September 4 to a high of 391,455 on October 29, a difference of over 115,000 votes.

Posts actually fair the best in this data set, perhaps because there isn't as much left to go. The high on Halloween fell short of 8,900, while the low for the week did not reach 7,600.

STEEM values have pretty much gone sideways over the last two and a half months, coming up some from the lows south of $0.13 USD to rest at $0.157-ish.

SUMMARY

The last year and a half has not been kind to the number of accounts transacting, the number of votes cast and the number of posts made.

During this period, the number of accounts transacting daily went from a high of 67,216 on April 27, 2018 to a high of 33,528, a decline of approximately 50%.

And the number of daily votes had a high of 779,964 (or 769,257 non-paid bot votes) to a high of 391,455, again, a drop of nearly 50%.

The number of posts have been hit the hardest—from a high of 56,489 on April 25, 2018 to a high of 8,882 yesterday (October 31, 2019), a fall of 84.3%.

While the total number of comments has gone down, too, the average total per post has pretty much remained within the same range it started at 18 months ago.

STEEM fell from $4.23 to $0.157. That represents a decline of 96.3%.

CAUSE AND EFFECT

I'd be hard pressed to argue that STEEM price is not a significant reason, if not the number one reason, for the across the board declines. However, I could also argue that the most recent hardforks, including HF21, has not helped to staunch the bleeding.

Not yet.

The fact that the decline since just before HF21 has been less than any other similar period could signal a reversal will occur in the coming weeks and months, or it could just be there isn't as far down to go.

But does any of this speak to user sentiment one way or the other?

Well, just knowing the numbers are down, and still heading south doesn't help me any, regardless of what my personal situation might be. And for the record, I would say the payout amounts are up somewhat from before the latest hardfork on my posts that don't find their way to one of the curation trail. While I'm happy with any increase, it's not so much that it improves my outlook on things over how I felt prior to the advent of newsteem.

That said, there are numbers that might help out if they were available somewhere.

UNKNOWNS

The total number of posts, unfortunately, does not tell us the total of unique accounts publishing them. Some users post a few times a week, some daily, some multiple times a day. So, the post totals don't represent the total number of unique posters.
Nor does it tell us how much of the decline is due to spammers throwing in the towel and how much of it was earnest, honest users who have simply given up. The former would lead to positive user sentiment, the other would have the adverse effect.
The number of accounts transacting does not tell the entire tale when it comes to the number of daily transactions on the blockchain. Each user only needs one transaction to be counted, but many, it would seem, will have more than one per day. Knowing where the total amount of daily transactions stands over the course of 18 months would be helpful.

CONCLUSIONS

All of that notwithstanding, these numbers, and their decline, are good to know. I still feel along with whatever the ratio of spammers, vote farming, alt accounts, etc, that have stopped posting and voting, there is a large group of normal, well-intended users that have disappeared or curtailed their activity. That doesn't make me feel very good.

Either way, there is certainly less to look at as far as posts and comments go.

One could surmise that with fewer posts, and fewer comments (along with fewer users upvoting comments thanks to the convergent linear curve and its higher requirements for payouts), that anything less than a comparable drop in voting would mean more posts getting more votes.

But if that is true, is that enough to buoy up user sentiment, to stop folks from quitting, to keep things more positive?

If the number of creators are down but the number of curators are up, that could ultimately help the mood and be a good thing for creators, but only if curators can accomplish whatever it is they're hoping to accomplish with curation, which I assume would run the gambit from wanting to support people they know to maximizing ROI.

The former seemed to be happening regardless the circumstances, the latter apparently needed an EIP to pick up the curation efforts of those in it mainly for ROI, and we're still early days as to whether their curation pace will continue, or whether we'll be hearing about something else someone decides we need to make things work better.

Maybe that's SMTs and Communities.

After all of this, I wonder what the numbers will look like in March or April next year when we finally get to SMTs. I wonder what my personal sentiments will be like by then.

_{Images sources—penguinpablo's daily statistical reports and worldcoinindex}