Once you understand Go, the next step would be to pilot migration on one of the services in your application. A common mistake is to take the simplest service as this defeats the purpose of learning anything from the pilot. You should pick something that clearly identifies the Go value proposition for your product.

The ideal service should be something that involves a lot of concurrency/orchestration, such as a web-facade API. They generally also have some persistence, allowing you to experience first-hand Go workings with DBs/ORMs. An API-based service also allows phased dial-up of traffic from the old to this new service—for example, in the travel website example, you can take traffic for hotels in a specific city on this new stack.