Inspiration. Desperation. In an original music band, both of those are important motivators when choosing what cover songs to play.
Especially early on when we were a three piece with about 6 vaguely downbeat original songs, covers were a necessity if we wanted to fill out a full set. Choosing which covers to invest time in also has more logistical ramifications in a band as opposed to a solo or duo act, so what you pick often depends upon where on the desperation-inspiration axis you are.
Got a show tomorrow that you don't have enough material for? Make a list of songs the singer knows and that the soloists can fake their way through. Personally I hate this method; I'm far too (over?)cautious to enjoy completely winging something in a public performance and I almost always feel like I did a disservice to the material and the audience when I do that. I'm always left comparing the song that comes out not with the silence that it replaced, but with what it could have been in more capable hands. That anxiety has abated a bit since I've gained more experience, but it's still always tapping away down there.
With a little more time you can make more intentional choices about both what kind of songs you want to play and how you want to play them. You can go searching for inspiration, as it were. There's a tension between wanting to play things audiences are comfortable and familiar with and playing something either off-the-beaten path or in an off-the-beaten way.
The type of gig is also important. For an original band playing a short, one-set show, a well chosen cover can be useful to either engage or re-engage an audience after a slew of songs they might not know.
Longer gigs are a different beast. We can all recall being at restaurants or bars and hearing an acoustic player cover songs we know in the background, but it's such a ubiquitous experience that, even if they are songs we really like, we're probably not super inclined to remember much about the individual rendition.
And to be clear: acoustic covers can be really hard to sustain for any length of time especially as a solo artist. Unless the artist is an incredible guitar player with a flexible voice there are only so many ways to re-arrange and tweak acoustic arrangements before they start to migrate to a similar sound. (Reggie Williams is a good example of someone I've seen sustain a varied and interesting acoustic cover set for the long haul, but he's kind of a freak in the best possible sense. Anyone familiar with his work as R. Lum. R knows his vocal ability and range are phenomenal AND he's one of the best guitar players I've ever seen. P.S. He's been in Rolling Stone a few times lately. Please please do yourself a favor and check him out.).
Familiar melodies from popular songs can buy you some extra time*, but a consistently great cover set in an acoustic environment is a lot of work. So be understanding to acoustic players at restaurants; it's harder to do a really good job than it looks.
Some of the migration toward similarity in a long acoustic set is structural to the environment. Background music is, well, background music and the familiarity of song choice and homogenization of delivery is a good, easy parcel of sound for an audience to passively receive, but it also means it's easy to tune out.
* Mere Exposure Effect, per wikipedia: "The mere-exposure effect is a psychological phenomenon by which people tend to develop a preference for things merely because they are familiar with them." Check out an interesting (though long) article about its effect with music here and an interesting TED Ed video about repetition within songs here.
Being in a band with non-standard instrumentation, like us, gives you more music pallet choices to choose from. We also have the advantage of having two soloists on different instruments (guitar / dobro for Sean, and me on mandolin) to work with.
My personal approach when thinking about songs to cover tends to be more focused, trying to gain audience goodwill by doing something unexpected rather than by playing the familiar. Given the choice between a good rendition that is only a slight variation on what an audience has heard plenty of times before and a decent rendition of something more creatively unfamiliar, I usually lean to the latter. Hence we play "No Surprises" rather than "Creep," and "Miami" rather than "Mr. Jones," and "Porch" rather than "Alive."
There's a whole host of songs that fall between "overplayed" and "this might as well not be a cover." Songs that the audience knows pretty well, but doesn't hear live bands do covers of very often. (I admit, especially early on, a lot of our covers were much farther to the "is this even a cover" side of the spectrum.)
"Take On Me" is one of those best of both worlds songs for us. I remember being at a friend's 30th birthday party at The Lodge downtown and when Take On Me came on; the place went crazy. It occurred to me that I'd never really heard an acoustic cover of it before. I went home that night and found that, mercifully, the distinctive synthesizer riff laid really nicely on the mandolin. We started working it into the rotation and it gets a good response.
We play it pretty straight up to the original version. Since there's less competition in an audience's memory with "Take On Me" than there is with, say, "Wonderwall"* we don't have to go as far out of the way to make it stand out. In general, then, the more unexpected (or bonkers) a song choice is the more freedom you can have to NOT screw with the arrangement to make it stand out. There are a few exceptions to this guideline in our repertoire but there were extenuating circumstances (see below). A part of my brain also resists playing really familiar songs extremely straight up because, frankly, a lot of those kind of songs are in that situation because they're really good and I suppose it's a defense mechanism against making your own material suffer by comparison.
*About Wonderwall: There's a running joke in music circles about covering Wonderwall, to the point where there's memes about it. It's generally considered kind of a faux pas, due to the song being really overplayed for a long long time. As my friend Bradd says, for a while if you drafted songs before seeing live music (i.e. guessed which covers an artist would play), Fantasy Football style, "Wonderwall" would be the Tom Brady of acoustic sets. I would guess there's some residual judgment on the song because of Oasis' Gallagher brothers reputations as being dicks, and also some backlash against a popular and, let's be honest, pretty catchy song that everyone has heard 3,283,416 times. (Roughly)
There aren't really any covers in our repertoire that I hate. Dan did play "Margaritaville" once at Attic Door after a request from the audience. Knowing my... let's say... "volcanic" hatred for it, he'll now and then jokingly ask "Should we put Margaritaville on the set?" My stock response is "As long as you don't mind playing it by yourself." There are some things I will not be a party to. I SAID GOOD DAY, SIR.
Most of our covers are 90's covers, partly because most of us in Beemo are children of 90s music and partly because we had to learn a bunch of 90s covers in a very, very short time at the end of 2015 for Orlando Weekly's 25th Anniversary party at Cheyenne Saloon, which was appropriately 90's themed. We cobbled together Weezer's "Undone (The Sweater Song)," Smashing Pumpkins' "Today," and R.E.M.'s "Losing My Religion" in about a week, all of them pretty faithful to the originals because we were under time pressure and they all have nice distinctive melodic parts that fit well on the mandolin.
"Losing My Religion" was a weird one for me, as it may have been the first cover we've played that I didn't have to adapt a part for since the original prominently features a mandolin. I remember thinking, after looking it up online and learning it in about 5 minutes "Huh. This must be what guitar players feel like when they learn a song." It was quite a different experience than trying to mangle a guitar part (or a piano part) onto a mandolin in a way that makes it both sound pleasant and not impossible to play.
Our most WTF covers came out of a show we played at Back Booth several years ago where the concept was that you played your set and then did an encore made up of covers that were outside your genre. We were the only acoustic band that day, followed by two alternative rock bands, a punk band, and a very heavy metal band (Murderfly, who was awesome btw).
Our encore consisted of Nine Inch Nails "Down in It," Digital Underground's "Humpty Dance," and Ozzy Osbourne's "Crazy Train." Dan was the driving force behind the first two. He had suggested playing a Nine Inch Nails song, half in jest and we realized it was a great idea. We didn't have a drummer yet, so I was banged out the bass drum / snare hit rhythms on the body of my guitar while Sean played slide on the dobro.
Dan inexplicably knew all the words to the "Humpty Dance," so we just vamped on a jazzy original song we already had down while he sang over it. (Not super ambitious from the arrangement side, but again we were really pressed for time)
I had the idea to do "Crazy Train" after hearing a rock band play it in a bar and listening to the crowd's reaction to the opening riff. It was another one that was instantly recognizable but I had never ever heard from a band like ours. Also, I knew that Sean could already play the solo note for note from his 80's shredder days, so it was actually lower hanging fruit than you'd think. (Ask Sean sometime about the foolish bet his friend made to him when it came on the radio one time; it's a NSFW story, but it is hilarious.). Again, I got lucky that adapting it to the mandolin wasn't too painful, though the verse part required some contortions that took me a little time to get down.
I very much enjoy the front end task of working out a creative arrangement to a song. Adaptation is kind of a horizon-expanding experience and a really good practice technique. It's very easy to get into a rut, going to a standard set of go-to moves, rather than actively forging a new musical pathway.
I've been trying to make time to learn new songs for its own sake as a pedagogical tool in order to become a better, more flexible player. Taking a guitar or piano part and forcing it onto the mandolin is a mind scrambling experience that forces me to face the choice between note-for-note fidelity to the original (sometimes impossible on the "wrong" instrument) and creative approximation. Either way it puts me in the position of using a fretting, picking, or melodic style that is outside of my default toolkit. It becomes a task of what I call "deconstructive listening" where learning a song is not just a matter of learning the "what" of the notes, but the "how" and "why" they are in a particular spot in an arrangement.
Doing this with already written and quality vetted songs allows me to focus more deeply on these hows and whys; I can devote any creative energy I have to understanding the part rather than writing the song. I'd rather work out a cool mandolin part to someone else's good song than write a cool mandolin part to my own shitty one.
Lately I've been focusing on learning songs by Orlando artists whose work I admire. At first, it was just a way to stave off boredom and satisfy some musical wanderlust, but it's sort of morphed into a project in its own right. Hopefully you'll be seeing some mandolin showing up in some unexpected places in the coming months.
Stay tuned. And please don't ever request "Margaritaville."