We have been off cable for a bit over two years now. It has been fine — better than fine for the last eighteen months — and almost everything that went wrong in the first six was avoidable. This is the list I would hand myself in early 2024.
The connection is the service. Not the provider, not the app, not the box. If the picture is bad, the odds are heavily on the two metres between your router and your box, and you should exhaust that before you complain to anyone. I lost four months to this.
Do not buy hardware first. I bought a box before choosing a service, on the reasonable-sounding logic that I would need one either way. The one I bought was fine and the one I would have chosen afterwards was better.
The player matters more than anything on a specification sheet. Same subscription, same connection, two apps, one of them unusable. This is the single largest variable and nobody mentions it.
Count simultaneous viewers, not screens. Then add one.
Test support before you pay. Send a vague question and see what comes back and how fast. It predicts the whole relationship.
Set up favourites on day one. Six hundred channels is not a benefit if nobody can find the eight they watch, and the household will judge the entire thing on how quickly they get to the news.
Test on a Saturday night game, not a Tuesday afternoon film. Everything works on a Tuesday afternoon.
And give it a fortnight before deciding anything. Every strong opinion I formed in week one turned out to be wrong.
If I were starting again with none of this, the thing I would read first is this rundown of what actually matters here, because it covers the connection and the player rather than leading with channel counts.
Two years, one genuinely missing channel, and about a third of what we used to pay.