Realistic earnings estimator

The original earnings estimator is no longer online, so I’m altering this suggestion to advocate for the return of a realistic earnings estimator. In the mean time SNOs can use this community version to get a good estimate of the earnings potential.

Please note I will NOT grant edit permissions to this sheet. Please copy it to your own Google account to fill in your own inputs.

The sheet predicts the first 10 years of profit based on a best effort estimation of network activity. Additionally it gives an indication of what the maximum earnings potential is. I update the sheet regularly to match current network behavior.


Screenshot made on 2020-08-03, estimates may change

All relevant network statistics used in calculations in this sheet are listen on the right. Let me know if you have any questions or suggestions for improvements. It’d be nice if Storj chooses to provide an official estimator at some point, but in the mean time I want to do my best to keep this one as realistic as possible.

Below you will find my original post suggesting replacing the previous Storj earnings estimator. Note that this criticism no longer applies as the old estimator has been removed, but I will leave the original post below for reference and context of the original discussion.

–Original post for historic reference–

I decided to post this as an idea so it can be voted on. I think we’re seeing a lot of heated debate around payouts that can be remedied by making the earnings estimator more realistic, based on what has been learned about the networks performance thus far. The current version of the estimator is both WAY too optimistic and not optimistic enough depending on the values entered. Bandwidth limits can no longer be set on the node software so they should be an output of the calculation resulting in a minimum requirement instead of an input.
But the biggest flaw is that it doesn’t sufficiently display how earnings will be low for the first few months, but get much better over time. This sets an expectation of immediate income. That expectation is broken and people lose trust in the projected earnings altogether before they even get to the better months. Potential SNOs should be prepared to expect low payouts early on knowing that it will get a LOT better over time, to prevent losing them before it gets interesting for them.

The following is a repost with some small edits from another conversation, with examples of how the current calculator gives the wrong idea.

I don’t think the payout is unfair, but I do think expectations are skewed. The official earnings estimator doesn’t really help here.
https://storj.io/storage-node-estimator/
For example, on a 100mbit (you can’t set it higher) upload, it estimates 8.34TB of egress traffic regardless of how much data you have stored, despite 10% of stored data being downloaded each month being a much more reasonable estimate. The slider for storage goes up to 20TB, but even with that amount stored 8.34TB egress doesn’t seem realistic. It also assumes you will fill up 50% of your storage in the first month and it stays there from that point on. Again, not realistic.

Perhaps it is now possible based on the data collected so far to determine a realistic percentage of egress compared to disk storage. Lets for argument say that’s about 10%. Perhaps it’s also fair to assume you get about 1TB growth per month on a good connection. This would depend on download speed, so you might incorporate that by saying 1TB for 50mbit and lower for lower connection speeds, but not higher for higher speeds as there is no longer an advantage above 50mbit for the most part. We also know that during roughly the first month you only get 5% of traffic due to vetting. Based on this you could build a much more realistic estimator… so I did.

Lets assume these inputs:

  • 6TB available storage
  • 50mbit download
  • 25mbit upload

Storjlabs estimator:

My more realistic alternative:

Most importantly this sets the expectation that earnings will be low during the first few months, but you shouldn’t be discouraged by that because it will get a lot better. However, there are limits to earnings potential in relation to storage size.

For example, if you only share 1TB, it doesn’t matter how fast your connection is, you’re not going to make a lot of money.

Needless to say the official estimator goes a bit nuts on this scenario. Please note, I set it to 1TB, but it assumes only 50% will be used. That’s why it shows 0.5TB.

It also shows that sharing virtually endless storage space doesn’t make you a lot of money right away either. Because it simply takes time to fill that up.

That said, it’s still optimistic on how fast it will fill up. And perhaps egress estimation can be tuned a little. I think this shows that it’s worth it for SNOs to run a node and sets an optimistic enough expectation, without leading to massive disappointment.

I hope something closer to this will be considered. If anyone wants to play with my version of this estimator, I made it available read only at this link. Please copy it to your own account to enter your own values to test with.

–Link moved to the top of this post. This sheet has been extensively updated, many notes in the original post no longer apply–

Notes on this sheet: Egress is capped by upload speed in case people have relatively low upload speed. I used a bandwidth calculator to find: 100 Mbit/s is equivalent to 32.8725 Terabytes (TB) per month. So mbit/3 is roughly the max TB egress you could get. But because egress can be spiky and it’s not realistic to assume your entire upload will be used constantly, I went with mbit/10 as a more realistic option. Assuming that about a third of your connection would be used for egress.

Thank you for the effort! The Storj Earning Calculator is very (too much) optimistic :slight_smile:

2 Likes

There are actually examples where it’s much too pessimistic as well. Mostly when you put lower bandwidth limits. This constraints payout way more than realistic. I don’t think it was bad intentioned. It was just based on assumptions we now have more information about.

1 Like

Thank you. I hope this will be considered.
It really pains me to see how STORJ is ignoring or even refusing this request in so many threads and it made me really sad.
Keeping the current, very unrealistic calculator online is of no help to anyone. We will only get more angry posts about people not receiving the amount of money they expected.
I’m not sure why STORJ did so and I don’t want to speculate too much, but personally I can not see any benefit for the network in trying to acquire SNOs that will leave 2 months later because of lower than expected payment, maybe even spreading the word about “STORJ cheating them”.

3 Likes

Your numbers look pretty consistent with my numbers for current earnings and will probably remain that way for the short term, but as the network picks up things will change. The problem with any estimator for something like this is the fact that you can’t accurately predict the data retrieval rate which is where you make the most money. What should be very clear to people though is that your never going to saturate the upload bandwidth of any data connection with only a TB or 2… unless maybe we’re talking about a DSL connection, but let’s be realistic here, haha. There’s just not enough of a pool of data to pull from… not enough user data to generate that kind of bandwidth even once the network picks up substantially. So I’m sorry, but nobody will be raking in $177/mo off a 1TB node. The calculator simply calculates the theoretical maximum as a best case scenario which is highly, highly improbable. If you really want to calculate potential real world numbers, you would have to get an idea of the average retrieval rate of currently existing cloud storage platforms. In other words, how many GB downloaded per month per TB stored. Once you have real world numbers it’s really simple.

T=Total TB stored
R=Average retrieval rate in %
(T x 1.5) + ((T x R) x 20) = $/mo

Currently on my nodes I’m seeing about a 10.7% retrieval rate so let’s calculate for that with a 1TB node.
(1 x 1.5) + ((1 x .107) x 20) = $3.64/mo

So your really not making much money. Even if you had 100% retrieval rate/month which just isn’t realistic to expect, your still only making $21.50/mo. If your just doing this for fun and don’t care then no problem, but this isn’t a get rich quick scheme… at least not on a small scale. You can make real money in it, but (and I’m not a financial advisor) you would need to invest some serious coin! Which people will do resulting in many privately owned datacenters.

So lets have some fun with this since for the right people there’s some real potential here…
100TB Stored
(100 x 1.5) + ((100 x .107) x 20) = $364/mo
1000TB Stored
(1000 x 1.5) + ((1000 x .107) x 20) = $3640/mo
Well now that sounds like fun doesn’t it… to bad that’s a minimum of a $40,000 investment (because you better have redundancy at that scale) and waiting for all that space to fill up. For arguments sake let’s say you didn’t have to wait to fill it, that would make for a 36.5% ROI over 15 months. Yeah… I’ll take that any freakin day of the week, thank you!

1 Like

I understand that the calculator is very very optimistic, as I think it is made to capture people and encourage them to create the nodes, but it is no use if once people have made the nodes, then they are disappointed and leave the project aside.

The first time I saw the calculator, I was surprised by the high profits it gave, but as I always have the “Internet factor” that nobody gives anything for pleasure, because I knew that was not going to be completely real.

I got the node because I didn’t have to invest any money at all, and I have a server that can probably be running a node for a long time with some disks that are not the bloody SMR, so I’m also one of those who thinks it’s better to be more optimistic about the calculator, even if fewer people have to join. I don’t think STORJ will want people running one node, five or ten, and after three months leaving it all behind for the sake of a small profit, either.

And another thing is the issue of the configuration of the nodes within the same IP, because there are many people who want to incorporate many nodes without knowing that it will not affect the profits that they think will be multiplied.

I am new and I have spent many hours reading the posts, and at the slightest detail of a failure, your node is disqualified so you have to start over, and these are things that also have to be explained, because putting an Rbpi and a hard drive as many people do, I do not see it very feasible by the various factors likely to a failure.

3 Likes

It is also suitable for this topic (@Robertomcat also copy it there):

1 Like

Yeah, I’ve been reading it, and I think I’m following it. Thank you.

Your experiences can help other new SNOs.

I don’t have experience, you just have to keep your feet on the ground, nobody is going to give you the money like when you access a website (whoever doesn’t have an ad blocker) and you skip the banner that you’re going to be the millionaire because you’re the 100,000th person to access the website.

I only have one server, and I’ve dedicated 15 TB to the project because I found out about it on the Internet, but I’ve only made one node and I hope I don’t have any problems, although there are always problems.

1 Like

That’s a well thought out post. I haven’t had a chance to dig yet. But, i thought there used to be a network utilization slider on the calculator. This could have been dynamically set based on current network behavior.

If i force my current nodes average ratings, it looks pretty accurate. Storj has the ability to calculate this based on their telemetry collection of storagenode operators.

the only problem is that download and upload speed in this calculator are expected to be the SNOs network bandwith capacity and not the average STORJ network node speed.

So people using this will always use their home internet bandwith (hoping it works like mining…) but can’t know what the storj network average is.

I’m not sure why this should be voted on? Are you voting on your own model or on that we need a better model?? for its not clear if you want to just promote your own model. Until this is clear this thread should be removed.

We can all make our own models and every model will be inaccurate, including the one StorJ has. It’s the same as predicting covid-19 cases.

1 Like

I think it’s pretty self explanatory. I’m arguing for a more realistic earnings estimator. It’s in the title…
My own version is just an example. And yes, if it helps people while there is not a realistic one on the storj.io website. All the better. It should be voted on in order to determine priority. Just like why everything else in the ideas section is voted on.

Could you argue the content instead of being combative for no reason? Do you disagree with anything? Do you think the original one is sufficient? Do you think my suggested version is incorrect? I would like to hear any of that

Trust me, I’m not here for self indulgence. I just notice the disappointment and heated discussions on the forums that could have been prevented were people better informed beforehand.

7 Likes

yes, VERY IMPOTANT. thanks BrithSilence and everyone discussing this.

I´m new here (12h online node just now) and came here for some guy showing the estimator on his site. Gave the impression you can earn 400$ a month easily You put more nodes, you earn twice, triple, and so on. He actually showed a dashboard showing 400$ earnings in a month, for a node almost 2 years old (a lucky month maybe?) I´ve spent hours these last days reading here and in many places, well spent time since I´ve learned a lot about so many different things. I know now these expectations to be very off, I know also there are other people currently buying 200$ machines to set up nodes, expecting to recoup the inversion in 4, maybe 6 months. This is going to end badly.

I think SNO operator are investing in this project. All this could generate very bad karma about storJ on the internet, preventing it from escalating, and causing the same for the revenues.

Anyway, you can´t expect it to be accurate, but… how can it be so wrong???

2 Likes

by the way! Is there any official information on how big is the tardigrade net for real? how many users, how much data?

And also,
i´m guessing that having all the relevant data on their hands (number of sno, how much are they being payed, how much data are they keeping and how much traffic) should be very easy to make a pretty accurate Estimator by their side. Just put a statistician to work with the data.

For that reason I started 2 nodes on the same server. One with the majority of the available data and one small one. That’s the one I always update fist or make any config changes to if needed in order to test. That way if I screw up I only loose the small one. I did actually get that node disqualified from one satellite while I was testing the node on a network share. That didn’t go so well, but hey it was something learned and little lost so no big deal. And 15TB ain’t a bad start!

Hello, good morning. Well, at the beginning of the configuration I had a problem, and the node spent a day offline and now I have a 95% confidence level and it’s hard to go up, so I’m going to wait for next month. But I’m not too worried about it either, it will recover little by little.

In 2020 January Storj was feeling generous did some kind of stress test on the network, which created a lot of egress traffic. In addition, nodes were paid triple the normal rates, so yes, it was possible to get that much in one month, but, so far, that was the only month and only older nodes got that, new nodes (started in January) got pretty much nothing.

The problem with anything, including this, that it does require investment, no matter what Storj says about it. Unless you are sitting on a bunch of unused (but working) hard drives and some other hardware you will need to buy at least something.
Using a 500GB drive will cap your earnings to a low value and may make it not worth the effort. At least with a bigger node and better hardware I can hope for a larger payout (and that did happen in January).
I usually do not keep large hard drives unused and creating the node out of the drives that I have may result in more problems than it is worth, because managing a pool with a lot of old and different-size drives would not be fun and would be trouble if one of those drives failed.

It really depends on the situation. When I started my node I had 12TB of unused space on my existing NAS array. So I could easily run my node of just that. I have since bought additional HDD’s to expand the array, but managed to do that with my actual Storj income. That said, I don’t think spending $200 is all that unreasonable. Last month earnings were $50 before surge. This month is starting to look very similar. But that’s with 10.5TB used. It’s fine to spend some money if it means you can share more space, but I still recommend starting with what you have or spending very little money. Especially since you won’t see anywhere near that income on the first few months.

2 Likes