I disagree with the accepted answer, because people who will be disappointed and won't come back were people who probably got served the wrong pitch along with the product.
Drive Your Pitch Home
The points that you need to drive home with a beta are:
- your product is cool and will be cooler soon (which is a nice way to say it's imperfect);
- their help is needed (thus valuable to you);
- their help is valued (thus retributed to them in some way).
People only go away if they expected something grander and they don't get it, and can tell from the beta that they'll never get it.
When I don't come back to a product after a beta, it's that the pitch was way over the line and I know the product will never be delivered as it was advertised.
Build a Community
If the right pitch is delivered, maybe not as many people will sign-up, but those people who do will be the ones who care. They will actually help and actively contribute and invest time and effort in you. And they will value your efforts as well, if you let them see what is involved with developing your product.
These people will be the foundation of your community. That's how it starts. And these things kinda often work on a "first come, first serve" basis. Which is annoying for late-comers, but not as annoying as a crap product or no product at all.
Reward your Hard-Core Fans, Beta Testers and Community Members
These people took a bet on you. Maybe not financially, but they did chose to trust you with something, be it their data, their time, or their input of some sort.
Make sure you do these 3 simple things:
- thank them,
- reward them,
- thank them again.
To be able to say "I'm sorry" and "thank you" is a very valuable skill.
Rewards can be a simple as a discount on the price of the final app. It can be an extra license for one friend. It can be a free t-shirt.
The thing is, the reward buys their time, and also produces a sense of urgency that will engage people in jumping in. It's the same "trick" as for sales or many card games. Offers available for a short-time are attractive so you feel like you will "miss out" if you don't jump in. And once you've invested in something, you don't want to bail and feel like you have to stick around longer than you would even have to (same as betting too high and too early in poker, yet the longer you wait to bail, the harder you'll fall if you took the wrong bet).
That's another very important point: free betas are great. But sometimes, paying Betas aren't a bad thing. They need to be signficantly cheaper than the final product, so that people don't feel cheated but you can still pass on misshaps under the cover of the "beta" tag. Also, they "lock in" the buyer's motivation because of that investment, and it ensures that you won't have to support hordes of freeloaders who'd come to try and bail right away.
Some companies do multiple stages in that spirit, starting with the free "alpha", and then the cheap "beta" (which is even free for those who were part of the "alpha", and then the final release (which is cheaper for "beta" adopters and even cheaper for "alpha" adopters). Just be careful not to implement a "bait and switch" system, as nothing angers people as much as this. It's actual trickery and dishonesty, and you'll lose people's trust.
Another option along these lines is to have the reward be a function of the involvement. Alpha users who just register and don't contribute anything are useless, except to boost userbase numbers for glory. Users who report issues are the ones you need to reward the most. Just be careful to not make it so interesting that people will want to submit crap that will take you time to review to determine its value.
Release early, release often.
So in general I'm all in favor of releasing early. Maybe it's not a beta, maybe it's a community preview, maybe it's an alpha, maybe it's some other strange name. But the earliest you can push it out the door, the better it is. Except if you're already a large company and people generate buzz on future products, in which case secrecy is a marketing tool for you as well.
And release often, to show that you care, that you're alive and hopefully well, and that you listen to criticism and get stuff done not only on areas that matter to you but on areas that matter to them (if that makes sense for you though, you need to be able to say "no" to requests that don't present a long-term benefit and beware of creeping featurism).