How to structure presentation to avoid getting questions that will be answered later in the presentation?

I'm constantly being asked questions that will be answered later in the slides.

You have to understand that this is a great problem to have. There is nothing worse than giving a presentation and facing a bunch of blank stares. Questions mean that

1. People actually understand what you're talking about.
2. They care enough about your idea to actually ask.
3. The problem you're explaining is not trivial.
4. You actually have an answer to the question - later in the slides!

You mention both conference talks and lab talks. In the first case you need to be a bit more prompt (always respect the time assigned to you in a conference, you're not alone and you'll put everyone in an awkward position if you go over time): just say "this is a great question, can we wait for the next slides?" or "can we take it offline?". I have yet to have seen anyone respond to this by insisting that you answer them right now.

But there are sooo many questions being constantly asked and it completely throws off the presentation because constant tangents are being drawn and one person's question confuses another person.

This is less a problem of structure than a problem of you not controlling the discussion. If there are many questions and people start discussing amongst themselves you need to take command of the room - say, "it's great that we're having this discussion, but I'd really like to move forward!", or something to that effect.

I believe that neither of your approaches (if I understand them correctly) is as effective as it can be. In conference presentations you need to focus on one big idea and explain it well - the objective of the presentation is to get people to read (and cite) your amazing new paper. Avoid giving broad overviews, and getting into every little detail. The audience needs to know why they should spend time reading your paper, and that's it.

In lab presentations make sure that the group knows what the big plan for today is. If there is a constructive discussion that's great. Tell them in the beginning the big points and then try to cover everything else. Give references to stuff you didn't cover so that the group knows where to look for things that weren't covered in the presentation.

Good luck!

Make it clear at the start that you will take questions at the end.

Then, when someone interrupts with a question, don't answer, but say: "please keep that question for the end" and continue.

If you stick to it then it will work, if you answer just one question you have failed...

I completely agree with the answer given by Spark. It is as much of a luxury problem as it ever gets in academia (maybe with the exception of having too much grant money and too few projects/staff).

To add on to that answer, I would like to give a couple of tips/insights on presentation techniques which may help you see this from a different angle.

First off, remember it all makes sense for you and for you alone. Often enough, connections and relations that may seem obvious to you will not be as obvious to someone else who has been thinking about other problems up until (and possibly even during your talk). So you need to take your audience by the hand and guide them through what you want to say. (Also related, but slightly off-topic, there is a saying in literature "The author is dead" meaning when you are reading something you cannot ask questions since the author is not there. However, during a talk you can very much ask questions about bits and pieces that do not make sense for you. Some people do that, and at time over zealously)

So, what can you do? I for one appreciate the introspection you seem to show and that you have tried different approaches. Either approach is fine, I'd say, but as you have commented neither is a solution alone.

I would suggest having some "meta-talk" within your talk. Overview/Agenda slides are particularly good for this purpose. You can tell people what you intend to go through/cover in your talk, how the concepts relate to each other, without going into any detail, in the beginning of your talk. That way you can prepare them for what is to come. This is also a good time to point out that some details will be revealed later, perhaps after you introduce the concepts but not their relationships to each other. This is a highly effective way to keep people interested, actually.

If questions arise, then as Spark also pointed out, you can quickly swat them away by saying something like "I am glad you caught that, let me come back to that in a minute". That way you acknowledge the wisdom and relevance of the question, without getting derailed. It takes some practice to keep your composure and stay confident, when your flow is interrupted, especially it happens multiple times.

I am not sure if this is an issue for you, but by getting frustrated or confused you might leave the impression that you are not actually sure about what you are saying and some people (especially on a conference setting) might see that as a sign of weakness and keep prying into it.

Finally, if despite your best efforts people keep interrupting you in your own lab meetings, you can bring it up with whoever is chairing these meetings and tell them that maybe questions should be kept for the end, or wherever the belong, in order to not hurt the flow of the talk. I cannot imagine an environment where people don't care about the fact that they are hurting a colleague's flow.

After all what is the point of going to a talk if you don't allow the speaker to speak?