RE: STEEM User Behavior Analysis Report — March Edition 2025
You're absolutely right—I should clarify. By "withdrawal," I meant any outflow of STEEM from an account, not just transfers to exchange platforms.
Initially, I was tracking transfers to exchanges like Poloniex and Binance (bdhivesteem). However, during the analysis, I realized that focusing only on transfers to exchanges could lead to misleading statistics. For instance, a user might first transfer STEEM to another personal account before ultimately withdrawing it. That would still qualify as a withdrawal, just with an extra step involved.
I also realized early on that it’s not logical to directly compare withdrawals and power-ups. A more meaningful comparison is between STEEM earnings and power-ups, which provides better insight into user behavior and engagement. The analysis took considerable time—as I reviewed 1,931,854 accounts—so I decided to share my findings to gather feedback.
I truly appreciate your input! Please let me know if anything I said is inaccurate.
Also, regarding the report I shared at a friend's request—do you think I should delete it?
cc: @steemchiller
There are many automated/reoccurring internal transfers done by different services and exchanges. You will need to overthink the current logic completely in my opinion. Couldn't you just do a diff from month to month? Store balances etc. for each account once every month and then compare to last month's data?
Yes, I can compare user balances month to month, but that alone doesn't provide an accurate picture. The problem is that I can’t always tell whether a user is powering up or cashing out just by looking at their balance. For example, someone might earn 500 STEEM and withdraw it, yet their balance could stay the same. This makes it hard to understand what’s really going on.
In my mind, the better approach is to track how much each user earns, which club they belong to (e.g., club5050, club75), and how much they power up or power down. I've noticed that some tutorials calculate club eligibility incorrectly. They only consider transfers and power-ups but ignore power-downs, which gives inaccurate results.
For instance, a user could power up 100 STEEM, power down 100 STEEM, and also transfer 100 STEEM. Some club5050 tools might still mark them as eligible, but that's misleading if the power-down isn't counted.
My plan is to store this data monthly, possibly on a server, and compare it over time. I’ll first share the results publicly here. By focusing on actual earnings, power-ups, and power-downs, this method should provide a more accurate and transparent view of user activity and club qualification
CC: @moecki
You should definitely clarify in the post what data you checked. You didn't compare withdrawals but transfers with power ups. As Steemchiller already wrote, there are so many transfers that are not withdrawals (for example from binance-hot2 to ...).
I would perhaps compare power ups with power downs (
fill_vesting_withdraw).Yes, my bad. I’ll update the post.
I’ll also compare power-ups and power-downs along with the earnings for each month.
Yes, because rewards payout in SP are power ups too.