You run automated tests for every build going through your CI system. You pound your new changes with manual testing, automatic testing, regression testing, functional testing, stress testing, load testing, blue testing, green testing…
Then you deploy it to production and something doesn’t work – you discover it 12 hours after the deployment, when a customer calls. Was it broken before the release? Did the release break it? You don’t know. You didn’t test.
All that testing you put into your software is useless if you don’t extend it to production. A problem in production is a problem for your customers. They can’t tell if you missed it in QA, missed it in the deployment, or knew about it all along.
If you are going to bother testing it before you release, at least do yourself a favor and test it after release to know it’s working in the one environment where it really matters.