Perhaps the defining mark of CA is its commitment to working with what it sees and hears. Or rather, with what the participants in the scene see and hear. It is wary of explaining what's going on in a scene by appeal to things that are hidden from the participants.

The most obvious reason why CA doesn't like to appeal to things like a person's inner feelings or motivation is that we usually can't know what those are - and, arguably, it doesn't matter. 

If you and I meet, we work out our business together knowing nothing about our respective inner lives. What we do know about is what each other's outer life is like: that is to say, what we say and do. Sometimes we display 'inner' emotions or thoughts; sometimes not; sometimes those displays are meant to be accurate; sometimes not. Nobody hears pleased to see you as necessarily accurate (or inaccurate).  

Whatever we do in interaction, we do so as to be understood to mean something. That is plenty to be going on with, both in doing social life and in analysing it.

What is not Conversation Analysis (though it might be another kind)

By now you might have played the episode on the video a couple of times.  Here are some examples of what CA would probably not be happy to say about it.

"Zoe says "the camera's on" because she's annoyed at her mother recording without pre-warning her"

Interesting, but too rushed. It glosses over just how it is an utterance like "the camera's on" could be understood to express annoyance. That is a job CA can do, and we shall see some steps in that direct in the first analysis ("noticing" in the banner link above). What it can't do, and would fight shy of, is then going the further step and putting it down to real annoyance. We shan't ever know whether Zoe was annoyed or not, though we could have a stab at saying that she acted as if she were.

"Zoe and Lyn act the way they do because they are mother and daughter, and comfortable in each other's company"

It might be true that Lyn and Zoe are comfortable in each other's company, but it might conceivably be false (they could have declared a temporary truce, might be disguising seething resentments, and so on). In any case, it is hard to see how this sort of statement, even if we had the means to tell whether it were true, tells us anything interesting about what exactly they are doing in the interaction.

"Lyn doesn't answer Zoe directly because it's a mother's prerogative not to have to answer to her children"

Perhaps an implausible example, but will do to illustrate the kind of claim that's based on generalised ideas about what happens in society, and an unsupported assertion that it's true here, and specifically motivating this particular line of the interaction. None are very defensible. It would be the same for any other general assertion pulled down from the abstract to the particular.

Examples of what CA can say: Noticing and Teasing

All that was rather negative, but might have cleared away a few expectations before we turn to the actual example of analysis: 'noticing'.