If you invest money in developing (and testing) a new feature, then you don't get the benefit of that investment until it is actually deployed and made availabl

If you invest money in developing (and testing) a new feature, then you don't get the benefit of that investment until it is actually deployed and made available for use. Until then, it is just burnt money with no return. Batching up many changes and features over longer and longer release periods brings so many issues, and definitely greater risks when actually deployed as the scale of change is larger. What the article perhaps doesn't pick up on so well is that poor or difficult deployment processes lead to this behaviour of batching and putting-off the deployments. Do them less often, have less periods of service impact. Is the logic. Which of course is one of the key points about DevOps, feedback loops, and CI/CD processes to simplify things. If you can make deployment easier, you can do them more often. Making the changes smaller can help ease deployments, so can be a virtuous circle. Though not if deployments are not reliable or repeatable.