With all due respect to louthy, making a decision to lower your prices based on the notion that a customer finds it to be a good use of their time to save $600 / year to build a copy of your service indicates that the amount of time it takes him to build it is worth less than $600.
If I were him, I wouldn't worry about 1 customer saying they'll build their own. My coworker (a PM, no less) decided, rather than Solr or ElasticSearch, to write his own search engine for a work project. He did it, and it obviously took 10 times longer than estimated and cost way more than was necessary. There will always be a proportion of customers who want it cheaper, but that doesn't mean you're priced wrong: it means they're not measuring value accurately.
For a good engineer, 600 dollars gets you something between 10 and 20 hours of work. To produce a polished project, start to finish, that's not going to happen.
'Less than $600 a year' is different to 'less than $600'.
If you plan to use it for 4 years (assuming team size remains constant), that's a potential saving of $2400.
With all due respect to louthy, making a decision to lower your prices based on the notion that a customer finds it to be a good use of their time to save $600 / year to build a copy of your service indicates that the amount of time it takes him to build it is worth less than $600.
That might not be a great target market.