For the most part, you need to represent value far in excess of the total lifetime cost, which includes half of the transition cost to the replacement.
Say your software is licensed at $10 per person using it per year. In addition to the $10, I need to consider the costs of installation, configuration, first level support, entitlement/license management, and what it will cost me to extract the data and munge it into a usable form when we decide in three years that your competitor does it better.
Now consider that most companies offer lousy support for their products, so I also have to do a test roll-out to a small number of people, manage that, and figure out if it will all be worthwhile.
If you discount based on quantity, can I roll the test group into the main group? If you price in chunks larger than per-user, do I fall between your levels in some disadvantageous way?
Is it any wonder that I vastly prefer standard-protocol based, open-source applications with healthy user communities? No, it is not.
It's software that automates a lot of costly auditing and troubleshooting in complex systems, so the core value proposition is hours that humans don't spend (and reduced downtime). If it can reduce the workload of senior/lead engineers by 5-10%, what's that worth? This sort of thing is really hard to quantify, although I'm sure we'll get more data as it grows.
Does per-seat licensing work best, or something modeling the amount of data being processed? Fixed or floating seats? We're definitely looking at higher pricing, from hundreds to a few thousand per seat annually (not unusual in the market). I'm not worried about whether the product will be loved - we're engineers ourselves, totally focused on the suckiest parts of working on enterprise systems. If our product isn't a pleasure to use, I'll take it as a personal failure. But then we have to sell it. And I'm an engineer solving my own pain points, not a sales guy. That part is new to me!
Everybody prefers predictability. That shows up in the comments on this page about people who hate to see "Call for pricing" on the site: it means that the cost is extremely complex.
Here's a question: will people want to use the software because it makes their lives easier, or will people want to have the software running because it makes their lives easier? That's the difference between a good per-seat plan (where customers want more seats because employees want to use it) or a good per-[quantity] plan, where nobody cares who is running it but the company does better as a result.
Feel free to email me at the address in my profile if you want a deeper discussion.
Say your software is licensed at $10 per person using it per year. In addition to the $10, I need to consider the costs of installation, configuration, first level support, entitlement/license management, and what it will cost me to extract the data and munge it into a usable form when we decide in three years that your competitor does it better.
Now consider that most companies offer lousy support for their products, so I also have to do a test roll-out to a small number of people, manage that, and figure out if it will all be worthwhile.
If you discount based on quantity, can I roll the test group into the main group? If you price in chunks larger than per-user, do I fall between your levels in some disadvantageous way?
Is it any wonder that I vastly prefer standard-protocol based, open-source applications with healthy user communities? No, it is not.