[azernik]

Price discrimination is a hard problem! There's a conflict between soaking your customers for as much as they're willing to pay, and making your product accessible to customers that won't pay that much. Finding a feature set that the first set of customers absolutely needs and will pay extra for, but the second set can live without, is hard.

[donmcronald]

The problem is that pricing is expected to be the opposite of what it really should be. Companies with 10k users expect a discount, but the reality is they should be paying a lot more because building a product to serve 10k users is much harder than building a product to serve 10 users.

On the flipside, why should I, as a small user, share in the cost of things like scalability, HA, SLAs, etc.? I don't need those things.

Based on what I've seen in larger businesses, those things can be paper features. It doesn't matter if the small business edition has the same scalability and HA as the enterprise edition. Large businesses will still pay for the enterprise edition because they need to be able to show they paid for HA, etc. to cover their own asses if anything every goes wrong.

Charge for scalability and reliability, not features IMO.

[sofixa]

Hard disagree. Scalability and reliability, and SSO[0] are basic features any piece of software should have. Even a one-person startup shouldn't have to accept a tool being down for a month or all of their data lost due to a fire in a datacenter. Charging extra for any of these is a very shitty business model.

0 - sso.tax

[donmcronald]

How much scalability does one person need? And no HA doesn’t mean “down for a month”. It means I have no issue with enduring a few hours of downtime here and there. And HA and backups aren’t even related, so...

[sofixa]

One person could still require significant capacity - imagine an API with flight info, and your one person SaaS that does some magic with it by getting the data real time for all flights around the world.

> And no HA doesn’t mean “down for a month”. It means I have no issue with enduring a few hours of downtime here and there

No HA means downtime is expected and frequent ( baring crazy luck, but let's not base business on hope). If your invoicing, or project management, or note taking or whatever SaaS is down and you need it now to do something for a client of yours, how many hours of downtime can you accept?

> And HA and backups aren’t even related, so...

You said scalability and reliability, good backups are a part of reliability. If a SaaS cheaps on backups and HA for small time clients like one-person companies or small startups, and then a fire burns it all ( but not entreprise customers, because their backups are replicated to another DC), is it OK because they were small?

[alisonkisk]

Is it 1000x times harder to server 10K users?

[higeorge13]

In some cases yes. I worked for a b2b company, which was not designed as multi-tenant and all their customers were under the same db and the same service. Now imagine response time for small business vs enterprises (yes they had a few of those paying premium, lol), sometimes the later was infinite.