August and September '18 New STEEM User Report

At the beginning of every month I was preparing a report containing details of new users for the previous month. In April we broke 1 million registered accounts. However, we have seen a downturn in the users registering, and an increase in bot and spam activity and I didn’t feel a monthly report was providing much value.

For that reason, I had changed to a report every second month, providing details of the two previous months.  

If you missed the last report you can read it here

https://steemit.com/utopian-io/@paulag/june-and-july-18-new-steemit-user-report

However due to HF20 and the changes in account creation, I believe this report will be of value on a month to month basis and will be returning to producing it monthly, with additional information relating to account creation tokens.

Aim of Analysis

The aim of this analysis is:

• Establish how many new users registered

• Establish what % of accounts have posts

• Find out who are the new big players in terms of SP

• See what accounts are acting suspiciously

• Establish if there is a break in any of the trends from previous periods 

 Repository

https://github.com/steemit/steem

Year to Date Data

 

In 2017 to the end of Sept, 0.6M accounts were registered.  In 2018 for the same period 0.42M accounts have been registered.  

The median new accounts per month this year is 60K, however we can see clearly that the first quarter of the year had substantial more registrations the then next two quarters.  New registrations had been decreasing since March, however we had a small upturn in September.

August Overview

   

The above dashboard shows details of accounts set up in August.  Of the 22K new accounts, 18.33% have completed their about section in their profiles and 8.58% have added a link

38.28% of accounts have posted once, with 13.49% posting more than 10 times and 4.16% posting more than 50 times. 

Looking at the activity pie we can see Not active accunts make up 50.6%.  Fully active accounts have both posted and voted.  This makes up 27.4% of accounts registered in August.  Almost 11% of accounts have posted but never voted and 7% have voted and never posted.  Finally almost 4% of accounts are Wallet user only.

9K STEEM was spent on Account creation fees and 621.64M vests were delegated to in the set-up process.

The table on the left of the dashboard shows who set up the accounts.

The chart below shows the number of accounts created each day in August compared to the same period last year.

  

September Overview

   

The above dashboard shows details of accounts set up in September.  Of the 25.5K new accounts, 14.9% have completed their about section in their profiles and 7.27% have added a link to a different source.

33.4% of accounts have posted once, with 8.84% posting more than 10 times and 1.57% posting more than 50 times. Although these values are down on last month, I would not be concerned as it was HF20 month and it did have an impact for a few days on the level of activity.

Looking at the activity pie we can see Not active accounts make up 34.83%. This is down on the previous month.  and Fully active accounts have both posted and voted.  This makes up 25% of accounts registered. Just over 8% of accounts have posted but never voted and 30% have voted and never posted.  Finally almost 1.63% of accounts are Wallet user only.

7.7K STEEM was spent on Account creation fees and 725.65M vests were delegated to in the set-up process.

The table on the left of the dashboard shows who set up the accounts.

The chart below shows the number of accounts created each day in September compared to the same period last year.

  

HF20 Changes

Under HF20 accounts can now be set up by either burning the STEEM and sending delegation so the accounts can transact, or they can be set up with discount account creation tokens where the fee is 0.

For the entire month of September the pie below shows that the vast majority of accounts were set up with a base fee of 0.1 Steem.  Prior to HF 20, this value was multiplied by 30 to get the actual cost so 3 Steem.

   

Zooming in to post HF20 we can see a big change.  Almost 31% of accounts were set up using the account creation tokens and 0 Fee.  56.55% of accounts were set up for 0.1Steem and as this was post HF20, no multiplication was applied.

     

 Conclusion

HF20 made direct changes to the account creation process.  It will be interesting to see moving forward how this impact the number of accounts that will be set up each month and it will also be interesting to see if there is a shift from Steemit Inc to other in the account creation process.

I also found it interesting that the number of accounts slightly increased in September and I am wondering is this a false positive because of account grabbing prior to the HF 

The Data-source and Query

As always to produce this report I connect to Steemsql (paid subscription services held and managed by @arcange ) with Power BI. The SQL query used to gather the data for this report was

Select *

FROM Accounts (NOLOCK)

where   

( created >= CONVERT(datetime,'01/01/2018')

AND created< CONVERT(datetime,'10/01/2018')

To get details of who set up each account I used the following query

Select *

FROM txAccountcreates (NOLOCK)

where   

timestamp >= CONVERT(datetime,'01/01/2018')

and to get details of accounts that used their wallets I used the following

select [from], timestamp

from TXTransfers

where   

timestamp >= CONVERT(datetime,'01/01/2017')

Once I loaded the data into Power BI, I then carried out transformation and calculations using DAX language.