Note that when I say "you" below, I intend the general "you," meaning all those writers out there. [waves hand toward the horizon]
1. Be very clear about what you want and what your beta is going to give you. Discuss it beforehand. Do you just want mechanical help? Do you want canon help? Do you need help with a dialect or local knowledge? Do you want someone to brainstorm plot ideas before you've even started writing? Do you want the whole works, absolutely everything the beta sees and notices, positive or negative? If what you want and what your beta does don't fit then part company with no hard feelings. You need to find a good match and your best friend or someone who's "willing" isn't always going to be a good match.
2. Don't use the beta as a crutch, or an excuse to not learn your craft. You're the writer and it's part of your job to learn to spell, to punctuate, to use the right verb tense, to use the right word in the first place, to learn to write transitions, to learn to write a flashback without the lame device of putting *FLASHBACK!* before and after it, to learn to paragraph properly, to learn to control point of view -- all that stuff is your job. No one's born knowing these things and your betas will help you learn, but I'll only work with a writer who's willing to work and learn and improve. Ideally she'll eventually get to a point where she doesn't need me anymore, except maybe for the occasional "How the heck did I miss that?!!" kinds of things.
3. Related to the above, take responsibility for your story. If someone in comments points out an error, don't say, "Oh, my beta missed that." No, you missed that. You're the final arbiter of what your story looks like when you post it. Your betas will help you but the buck stops with the writer. If you acknowledge your beta in your header then everyone reading your story will know that someone else went over it, and yeah, probably should've caught the error. But pointing fingers when some reader finds a problem is a cop-out -- own up to the mistake.
4. And related to both two and three, be an informed consumer when shopping for a beta. There are a lot of absolutely sucktastic betas out there, many of whom have heaps of praise and shiny thank-yous from people they've betaed for, who nonetheless can't write (or edit) their way out of a virtual paper bag. I don't know how often I've seen a writer praising her beta (or her two or even three betas) up, down and sideways in the story header, then gotten down to the story itself and found misspellings and grammatical errors and commas that looked like they were scattered with a shaker-jar, point of view glitches and misused words and whatever-the-heck else, and I had to wonder what the thing looked like before they "betaed" it. And honestly, I wonder why it didn't look a hell of a lot better afterward, too.
This is sort of a consumer issue -- it's like hiring a contractor to work on your house. You probably don't know anything about plumbing or electrical or installing a tile floor, so you hire a pro. But because you don't know anything about the process, you're an easy mark for some bozo who comes in and takes your money and does a shit job. At least with a contractor you can ask for references and actually call them all and hear that the guy's plumbing leaked or his electrical work burned the house down or the family's piano ended up in the basement. With betas, if you talk to other people they've worked with, those writers won't necessarily know that their beta can't do the job properly because the story won't visibly "break" afterward. It's up to you to educate yourself as much as possible so you know whether or not this kindly person who's offering to do you the favor of betaing your story for you is actually any good at it. Asking around isn't any guarantee, but it's worth a try. But always be aware that when you "hire" someone who supposedly knows more than you do, you won't always know whether they're doing a good job. And that your best friend isn't necessarily a good beta.
5. Ask your beta to not only point out errors, but to explain to you why they're wrong and, if appropriate, make some suggestions about what would be right. This does a couple of things. First, it helps you learn how things work and how to do a better job yourself later, and second, it helps you evaluate whether or not your beta knows what she's doing. If she can't explain to you exactly why "Joe had wrote five letters to Bob" is wrong (but she does know that it's wrong) then you might consider getting another beta, someone who can help you learn the rules rather than just make red exes on your paper. If she tries to explain to you why "Joe had written five letters to Bob" is wrong then you definitely need another beta. :P
6. Don't get defensive if your beta finds a lot of things to squawk in your story. That's what she's there for, and if she's saying it then a lot of your readers would've been thinking it if you'd posted it that way. If you've asked someone to beta for you then unless you're paying her you're asking for a favor. Your beta spends a lot of time and effort she could've used writing her own stories, reading, hanging out with friends and family -- doing a lot of other things more fun and relaxing than helping you with your writing. You owe her politeness and at least your thanks, whether or not you take all her suggestions. And yeah, you'd be amazed how many people have asked for concrit, either one-on-one or by joining a fiction workshop, and then gotten nasty and defensive and sarcastic when they actually got concrit. [eyeroll] (The sole exception here is if your beta is nasty or snarky in the way she writes up her critique -- if so then my advice is to dump her ass and find someone else.)
7. Ask your beta how long she'll need to go over your story or chapter and then give her that much time. If you're three days late getting her the file, don't expect her to take three fewer days going over it just because your readers expect you to post every Monday. Asking if she can do a faster job is fine, but "Sorry, I can't" is never a rude response to that question.
8. Ask your beta whether she wants to see rough drafts or stuff you'd post if you didn't have a beta or something in the middle. (Notice a trend here? Talk to your beta ahead of time and work all this stuff out between you. There's no one right way -- what's right is whatever the two of you agree works for both of you.) I prefer the writer to do her best with a story before she gives it to me, but some people are more emotionally invested in something they consider a final draft. If it makes you feel less defensive and more open to a critique to think of the file you send your beta as a "draft" then hey, whatever works. :) And seriously, I'm always willing to answer questions, whether I'm betaing or not; I just don't necessarily consider it to be part of the critique process if the story doesn't at least kinda-sorta look done when you squint at it. But if you'd like someone to bounce ideas and problems off of while you're writing, ask your beta if she's willing to do that.
9. Remember that it's up to you, in the end, to decide what your story looks like when you post. You don't have to take every individual suggestion your beta gives you. You don't necessarily even have to agree with her that this or that chunk of your story is a problem. It is absolutely your right to cherry-pick from among the corrections and suggestions you get from your beta. If she has a hard time with this, find another beta.
10. Related to the above, if you find that you disagree with a large percentage of your beta's suggestions, take a look at what you're getting on the whole and decide whether your beta is working out. Maybe she will be and maybe she won't -- if you disagree with her canon comments but she gives you valuable advice about semicolons and past participles, then it's probably worth working with her. Maybe you could let her know that you just want mechanics advice. But if you're dismissing most of what she says and the stuff you do agree with here and there doesn't seem to be grouped in any clear category, then maybe you'd be better off with someone else. This, BTW, is another reason why it's not always a great idea to have your best friend beta for you; if you decide you want to "fire" her it could damage your relationship.
11. Keep in mind that different people are good at different things. If you want craftsmanship help, and canon help, and someone to help you with Britishisms or with the inner workings of the computer gaming industry or with police procedures, you might not be able to find that all in one person. Just because someone's good at your fandom's canon doesn't mean she knows the difference between "then" and "than," or "who" and "whom." Or that she has any clue how point of view works.
12. Keep an open mind about whom you accept to beta for you. Specifically, someone who's unfamiliar with your story's canon isn't necessarily going to be a bad beta. Canon knowledge is only a tiny part of what a beta brings to the table, and in my opinion it's not the most important part -- it doesn't even make the top five on the list. My view is that someone who's really a fan of a book or movie or TV show will absorb canon naturally. Figure, if someone's so into the source material that they're writing fanfic about it, they probably re-read or re-watch the source on a regular basis. They're also probably reading other people's fanfic and participating in or at least reading discussions about the canon material. They're going to pick up canon on their own and whatever they don't know now they'll learn eventually, just because studying canon is what fans do whether or not they're writing fanfic. It's all the other stuff, the actual writing, that's hard. People are much less likely to learn that on their own. Half the time they don't even know they're making mistakes. That's the part most people need betas to help them with.
If I used a beta I'd take someone who had excellent mechanics skills and an in-depth knowledge of the elements of writing and story structure but couldn't even name the main characters of my canon over someone who knew my canon up, down and sideways but thought that "bicep" was a word, or thought that "Whether you use 'he' or 'I'" is a good, thorough definition of point of view. Ideally you can find both but if you have to make a choice, go for someone who can help with the hard stuff. 'Cause people who know your canon are thick on the ground and it's easy enough to pick up on it yourself, but a really good mechanics beta is as rare as a four-leaf clover and just as lucky to have. :)
About what I like in a writer, aside from the above, I want someone who has a good "feel" for writing, who's a good storyteller and creates sympathetic characters, but who needs a lot of craftsmanship help. Because the "feel" is the tough part -- it's hard to teach, and some would say it's impossible -- and someone who's already a good storyteller has the potential to be an excellent all-around writer. The craftsmanship -- mechanics and characterization and dialogue and plotting and pacing and point of view and continuity and all the other skills that go into good writing -- is all very much learnable, and I can help a writer with that stuff.
I love watching someone with great potential but rough skills learn and grow and shine. Someone who doesn't have that basic feel for storytelling and characters might pick it up along the way, but then again they might not. At any rate, I don't have a good handle on teaching that part of it.
And on the other side of it, I don't feel I'm accomplishing much if I'm "betaing" for someone who in my opinion doesn't need it. I want to be of use and someone who sends me a clean file, a story where I can only come up with a couple of pages of comments and most of those either positive or just data points, makes me think that I'm not being very helpful.
And number six. Definitely number six. [facepalm] I rarely beta these days because I got tired of seeing baby writers who were very clearly in dire need of assistance get defensive and sarcastic at people who were trying to help them, whether it was me or someone else in a workshop, which is where most of my critiquing experience comes from.
The concept of the "beta reader" didn't exist back when I started posting fiction online, by the way. Some of us workshopped, but most writers just took responsibility for their own stuff. (I and most of the people I hung out with online back then were working on original fiction and wanted to be published. Traditionally writers polish their own work before sending it to an editor. Asking someone to give you money for your story is a powerful incentive to make sure it's as good as it can be. :) ) And the general result was honestly no worse than it is today, despite so many archives and even journal communities absolutely requiring stories to be betaed. It sounds like a good idea, but all it means is that someone's name has to be on the "BETA:" line of the header. There really are a lot of horrible betas out there, and apparently a lot of writers who think that so long as they've checked off the box marked "Betaed?" they've done their job and can shrug at how awful their story still looks. Hence number two.
I'm brutally thorough but never nasty. I want someone who can handle sending me five pages of story and getting back fifteen pages of critique, who appreciates the time and effort that went into it and is happy for the chance to learn and improve.
I do an all-or-nothing critique, so I want a writer who wants that. I comment on everything I see -- positive, negative and just data. If someone wants only mechanics help and I spot plot holes or characterization glitches or roughness in the pacing, I can't just ignore that stuff. Someone who's looking for a partial critique would be better off with someone else. And I don't mean that in a negative way; as I said above, it's important to find someone who fits what you want and how you work.
Finding the right person is three-quarters of the battle, seriously. And I honestly believe that most of the glitchy stories that get posted are the way they are because the writer got their best friend to beta, or took the first person who offered.