Come up with a pricing model that finally attracts customers

Ideas & Suggestions
1

Storj is a CDN, geo-distributed, faster than AWS, blablablabla.

We have heard these myths for half a decade now. Since the VC money is running out, finally this is now the time to come up with real pricing models that actually attract customers. Happy to hear your ideas. Please use the same template.

template to copy 

**Who would this be for?**


**What is our direct competition**


**Why would a customer choose us instead of the competition?**


**How does the pricing model look?**

|Product | Customer | Node | Storj profit |
|--- | --- | --- | ---|
|Stored TB per month | $ | $ | $|
|Download per TB | $ | $ | $|

**other changes that would be needed**

Here is my idea:

We make use of the USP of storj, which is “use only unused resources”.
That USP allows us to compete with others. That USP allows us to be cheaper than others.

Who would this be for?
Backup users. Homelabbers or companies that want to backup data over storj native.

What is our direct competition
Hetzner storage box and cheap S3 providers like Backblaze/Wasabi.

Why would a customer choose us instead of the competition?

Unfortunatly unlike a Hetzner storage box, we can’t offer webdav, SFTP, samba, restic, rclone or rsync. We can only offer storj native. But unlike Hetzer we don’t ask for 3.8$ per TB, but only 2$. AWS deep archive is even half the price at 1$ per TB, but it takes up to 12h to download files and the download costs are 90$ per TB.

How does the pricing model look?

Product Customer Node Storj profit
Stored TB per month 2$ 1$ 0.5$
Download per TB 20$ 10$ 10$

other changes that would be needed

  • Providers like Amazon Glacier (1$ per 1TB stored) have limitation like the smallest object size is 128k. Of course you can still upload smaller files, but they will be counted as at least 128k. Something like that or even bigger is also needed. Bigger files create less costs for the storj satellites. Mabye even a min size of 1MB is needed. The native app could create chunks, similar to how Veeam backup works.
  • distribution factor needs to be lower. 1.5 is already pretty high.
  • payment happens in real $. No crypto shenaningans. Nodes get $ in their balance. Payout over switft costs 20$ or whatever it costs storj to do international transfers.
  • hold back model needs to be changed. Keep it simple stupid. Payout is held back 12 months. Always. Forever. It is a rolling payout. As long as I keep my node running, I get my amounts with a 12 months delay. This encourages node to have a working system, and if the drop out before, their lost hold back amount covers the repairs.
  • Uptime does not need to be that high. Again, keep it simple. After a total of 7 days downtime per year, the node gets disqualified. But the higher download costs alone should be enough inventive for nodes to be always online.
2

Your plan dont contain repair costs, as backups usually long terms, it will be needed to repair defenetly.

3

That doesn’t work. The current expansion factor is somewhere around 1.88. Meaning with that alone the profit is down to 12 Cent. Now we also need repair on top. About 10% of the data needs to get repaired each month. You did not specify the pricing for repair egress. I am going to assume $10 per TB? That would increase the total cost to $2.88 per TB. And we also need to finance the satellite itself. Before all the cost saving that was about $1 per TB (all infrastructure costs). I believe it has dropped a lot but I haven’t looked up the current number. But at this point we will be losing more than $0.88 per TB anyway.

Edit: I also need to point out that this goes into the wrong direction. Storj is a premium product. Our current customers pay us $7 per TB for fast storage. So we should compete with other fast products out there. The moment we offer wasabi like pricing it will get hard to explain that we are a premium product. Our pricing needs to be within the range they expect from a premium product.

4

May be it can be lower tier, with lower expansion factor, may be even separate satellite. It like Select, Global, Archive

5

Storj themself said that they are not a backup storage provider. They are a high speed CDN. As a simple backup user you don’t need high speed CDN capabilities. You just want cheap storage that sits there (for you hopefully) unused forever.

6

How would we slow down the lower tier? If the lower tier is as fast as the normal tier all customers would want to take it. The lower tier needs to be actually slower otherwise there would be no incentive for the customer to pay more for same performance.

Lower expansion factor doesn’t work. Customers still expect high durability. We can’t create a tier that translates into store data for a few years with 50% file loss if you just wait long enough.

7

i would support any pricing that is scientifically proven to bring more revenue.
And the only way to prove it is to try it in real.
But how to try it without risking the current network?
i guess we need to create brands, like coca-cola has sprite and others.
There was a short exchange about that for more context:

8

What do you mean? Wasabi is $7.99 TB/month now. :thinking:

9

So even with today pricing storj beats wasabi already. but for this price, they dont change for egress.

10

These tools supports the native Storj protocol: Configure Rclone Natively - Storj Docs and restic works over rclone.

11

Coming up with business models that work is always hard. For Storj, it’s even harder because not only does the model have to work for customers and Storj the company, it also has to work for independent SNOs.

Some of the policies that Storj have developed are a bit anti-customer IMO and are ineffective. For example, having a minimum object size of 50K and minimum duration of 30 days. I understand (I think) the reason: there is a lot of overhead creating and deleting objects, both on the satellite and nodes, and since the fees will be low for small objects and objects that are quickly deleted, these extra fees compensate for the extra overhead.

My question is, does this really work? If the goal is to prevent the behavior, I’d say it doesn’t. Customers are going to store objects however they need to, and if someone is trying to DOS the service, they aren’t going to be paying fees. But more rules will give customers more to think about: Do these rules apply to me? Should I be worried about how much extra this will cost? What are our typical object sizes and durations? (this is very hard to know ahead of time BTW!). Amazon and Google are the kings of complex, conflicting rules and pricing policies spread across many tiers. I’m pretty convinced that the purpose is to make customers feel like they are getting a discount in a dimension that is important to them, such as price-per-byte, while at the same time getting ripped off in another dimension like storage time. It’s impossible to figure it out.

For me, and I’m guessing for many others, I don’t want to have to figure all this out. I mean, it must be important or the storage service wouldn’t have the rule, right? So I have to try to understand how this rule applies to me. Except if there is no rule; then I don’t have anything to figure out.

I think it would be an interesting exercise to figure out how much Storj billed for a month with these 2 rules, and how much they would have billed without them. I’m guessing it’s an insignificant amount of money, but maybe having the rules cost them a significant amount of business.

So when developing and evaluating business models, I’d suggest starting out strictly from the 3 perspectives: customer, Storj, and SNO, evaluate everything from all 3 perspectives independently, and honestly compare the policy to competitors. And always go with something simpler and better for the customer unless it’s just impossible.

That’s what has always impressed me about Backblaze: their policies are customer friendly, and have changed over the years to be even more customer friendly, like dropping transaction fees and having a 3x free egress allowance. Or overnighting a disk containing an entire bucket, and if you send the disk back, they refund your payment. Of course they lose money on that, but it creates a lot of goodwill and I’m guessing they rarely have to do it.

Once you have a customer I’ve found they are pretty loyal unless you screw up pretty big time. But getting a new one, that’s much harder.

12

Doesn’t Wasabi have a similar policy in place with 4K minimum size and 90 days minimum duration?

13

Of course they do. They just don’t let reality interfere with a nice round low marketing numbers. Hot storage. 90 day minimum retention. No contradiction whatsoever.

Remember, this is the company started by the same crowd who founded Carbonite — the “unlimited” backup service that silently skipped large customer files by default because, apparently, unlimited had some implementation details.

But they shove themselves down every reseller’s throat, plaster that nice round number everywhere, and win. Apparently marketing bullshit, repeated loudly enough, is still a perfectly viable business strategy.

There’s a fine line between being commercially aggressive and being a lying asshole. Storj needs to learn how to walk it.

14

Yes, and a 1TB minimum monthly storage fee ($7.99) if you store < 1TB.

Cloudflare R2 has taken the Google/Amazon approach (different details, but many rules):

  • Class A & B operation fees
  • no egress fee
  • no monthly minimum
  • multiple tiers (standard and infrequent access)
  • standard costs $0.015/GB or $15/GB
  • IA costs $0.010/GB or $10/GB
  • minimum duration of 30 days for IA tier
  • data retrieval fee of $0.01/GB or $10/GB for IA tier
  • every fee is rounded up: storing 1.2GB (in total I think, not per object) costs the same as 2GB
  • no minimum object size

I don’t like these rules because:

  • they make customers stop and think about whether to sign up and how the rules will affect them
  • they have to be implemented and tested in the software
  • they’re going to cause customer confusion, extra support questions, and potentially distrust if customers don’t understand the rule details and can never quite reconcile their charges
  • I don’t think the company makes that much extra money on them
  • customers aren’t going to change their behavior based on the rules: if their files are small, they’re small
15

Amazon charges for everything separately. So does google. I actually prefer that. I don’t want to pay flat rate because that means I"m paying a round number for services I don’t use or are subsidized by other users for services I do.

Storj is not a consumer service. Comanies like yours should be adopting it and selling a nice package to the end users. A good example is Arq 7 Premium: there is GCS on the back, but customers pay fixed flat monthly number and get fix flat amount of storage, and don’t have to think about API cost and segments.

Developers – do have to think about it, and they should want to think about it.

16

I took another look at Cloudflare R2 and realized something: it’s probably more expensive to store a backup in the infrequent access class than in the standard class.

With HashBackup, its recommended to setup a daily selftest that downloads and verifies a fraction of the backup every day, with 1/30th being a common fraction so that the entire backup is verified monthly. Each backup file is downloaded for each destination, the blocks are individually verified, and if any block is bad, it is replaced with the same block from another destination if one is available.

That means paying $0.02 per GB because of the IA retrieval fee. On top of that, there is the extra duration charge for IA.

I really hate these pricing gimmicks…

17

We tried to make a pricing with 1x egress. It’s still possible - $14/TB-mo + 1x free egress. But basically it will be the same as $7/TB-mo storage + $7/TB egress. You may just calculate it.
It might be attractive for the customers who has egress equal or more than the stored amount, but in the end it will be the same invoice.
From the other side, if you don’t download, then you overpay for the storage.

So we decided to return to a Standard model - then both customers will pay only for what they are actually using. Of course there is a minimum monthly billing to at least stop losing money on a small usage.

18

On the min file size and min retension fee. We have a small number of customers that are ramping up a significant charge there. It is not a loss if they leave the network. I would say it is working as intended. I would also say after all the infrastructure cost savings we have to revisit these fees one day and adjust them.

21

I am sorry I have no clue. I live in a country that is not following SEC announcements.

22

I understand. You and SNOs don’t particularly want them because they are inefficient for the network, but your fee policies haven’t convinced them to leave, which would be the best outcome. And I’m guessing that compared to overall billing for all customers for the month, the extra fees are insignificant.

How many customers were turned off by these rules? How many signed up but have a slightly less good feeling about Storj? How many continued looking at other providers rather than signing up? How many felt the need to go back and evaluate how those rules affected them before they signed up? Those questions are important too.

Here’s a 7-day discussion that took place on Veeam’s support site regarding 90-day retentions and bucket size limits on Wasabi:

The last post, from a person not directly involved, is interesting:

“I too have been planning to use Wasabi for cloud backup and had this not had a favorable outcome as it did, would have changed my mind as well. Thank you for everyone contributing to this topic.”

You rarely get an email from these potential customers saying “I would have signed up for your service, but didn’t because I didn’t like policy X.” They just leave.