I really like indie game bundles. With the Humble Bundle leading the pack, we've seen an explosion of bundles, and I buy almost all of them.
With that in mind, I'd like to look at A) the different methods of selling bundles and B) whether this might influence the sales strategy of non-bundles.
The classic "Humble" sales strategy is a DRM-free download for whatever amount of money you want to pay. DRM-free seems like the standard these days (YAY!) so I won't go into detail supporting it, but let's talk briefly about the "any amount" system.
Allowing people to pay anything is a brave foray into a new kind of sales. However, "pay anything" often ends up meaning "pay almost nothing", probably because our culture hasn't really developed a respect for the actual worth of something. If you pay less for a game bundle than for a meal, you're a total ass. Once our culture begins to realize this, we can expect "pay anything" to pay off better.
Until then, a few strategies have come into being to help nudge people to pay more. A big one I've noticed is the "average paid so far" marker, which lets people know what "normal" is. I'm not sure this works very well just on its own, but it is also possible to split this up into categories. For example, "Unix users pay $11.50 on average, while Windows users pay $6.40". I think this categorization is valuable, as I think it helps urge people to keep up with the best rather than just measuring themselves against the average.
In a continuation of that is the "leaderboard", where the top ten (or whatever) purchasers get their name and donation up for everyone to see. Often used as an ad, but that's fine, too. I think leaderboards can be very valuable, especially if you get a Colbert-style bump from somewhere. With the Humble bundles, the leaderboard frequently featured people who had paid thousands or even tens of thousands of dollars. Obviously, your popularity matters a whole lot, but it's not that hard to code (you can even just update it manually if you need to) and is guaranteed to get you at least a few donations at a higher level than you otherwise would.
Another method (and my favorite) is the "pay more than average" reward. If you're monitoring the average, then offer a bonus for paying, say, at least $1 over the average. It could just be an art pack, or maybe you'll mail them a signed post card or something. Anyway, the existence of this will push people to raise the average, which in turn will raise your overall income dramatically.
Some people have posited some fairly complex schemes such as prices dropping or increasing over time. I think those may have a future once our culture catches up to our abilities, but until then, keeping it simple is probably best.
Some bundles aren't using a pay-what-you-want system. For example, http://www.buygamesnotsocks.com/charges a flat $5.
I think this is a mistake. If you want to do that, you might consider a pay-anything method with a $5 minimum. I think it's a weaker choice than the complex monitored feedback system the Humble bundles use, but it's still better than just a flat $5: I would have paid $20, but it's physically impossible. $15 in lost sales.
What does this kind of thing have to say about individual games?
Well, buying a bundle is inherently less risky and more attractive, just for starters. Half the games in any given bundle just aren't interesting to me. There's nothing wrong with Cortex Command, it's just not my style of game.
To some extent, you can think of a bundle as diversifying a stock portfolio. If you have games of varied genres and audiences, you increase your chance of appealing to any given audience member. The risk of failing to appeal to them is much lower.
If you're only selling one game, there's no diversity, which is naturally going to limit your appeal to a much smaller part of the audience. But that doesn't really mean anything horrible! After all, if you're selling on your own without the Humble-style pricing, you still have the same limited appeal. It's just that you need to realize the weaknesses in your offer: you can't directly compare an individual game to a pack of games.
Also, the bundles often come after the "primary" sales life of the game, so they are often thought of as a kind of end-of-life "bonus" instead of a proper primary sales method.
Despite that, I think that indie games might find themselves doing better with Humble-style "pay anything" pricing, perhaps with an absolute minimum if you can't bear the thought of someone paying $2.50 for your game. As our culture gets more comfortable with this kind of scheme, I think it will become more and more acceptable and profitable to do it this way.
It's also worth mentioning that Kickstarter and similar can often garner you funding at quite a surprisingly high rate. Attempting to fund your game on Kickstarter is not recommended, because most of us know that games cost more than you think and aren't likely to ever get completed. However, if you have a game (or 99% of a game) and want to take it a few extra steps, a Kickstarter project to polish the art or add in additional levels could net you thousands of dollars of investment before you begin, which is another valuable revenue source.
Unlike the Humble-style pricing, Kickstarter-style projects work perfectly for individual games. It's worth thinking about using both Kickstarter-style and Humble-style for your indie game...
I say that, but, of course, I've never tried to sell an indie game. So take the armchair strategy for what it is... do your own research and come to your own conclusions. Just take note: pricing strategies and revenue sources are changing. Indies in particular can benefit.