Anytime you are building a web application, you are actually solving a business problem for your client. For example, that problem may be more driving sales or reducing staff time spend on a particular task.
You need to spend time diving into this with them to figure out what your project is really about. Then you can ask targeted questions to the client. For example:
"So, right now you're telling me that Jan and Ella are spending 20 hours a week processing the proposals. If we could reduce that to 5 hours per week, how much would that save you?"
"If we could improve the conversion rate on your flagship product by 10%, how much would that be worth annually in new revenue"
Once you've asked some questions like this, you can then position your price relative to these other values that the client has affirmed for you.
All of a sudden, it makes it much harder for them to feign shock at paying $25k for a system which increases revenue by $250k in the first year.
If you're not working with clients who have valuable problems to solve, find clients who do.