One of the things I bring to the table is my uncanny knack for sniffing out bad business before a contract is signed. Don’t get me wrong, this is a rare occurrence. Extremely rare. I have taken 99% of the business that’s been offered to me, from every single type and stripe of customer. And you should too. 

But every once in a while, you’re going to get an offer you can’t refuse from a customer you should refuse. 
As entrepreneurs, we’re practically programmed to take any and all business that comes our way, especially at early stages. There are innumerable reasons to do this. You need revenue to survive. Revenue is the best kind of cash infusion. You need customer references, the bigger and more well known the better. You have no leverage, especially early on, so you can’t say no even if you want to. 
Just remember that every single one of those reasons can and will work against you. 
We spend so much time doing due diligence on employees and investors and partners, but we rarely spend a second considering whether or not we SHOULD take on a customer. 
Make no mistake, a single bad customer can kill a company, and the risk is greater the younger that company is, especially if it makes up more than, say, 10% of overall revenue. And when I say bad customer, I’m not talking about doing work for bad people or in an industry you don’t abide by. Your moral principles are your own. I’m talking about bad business, when the organization or what you’re proposing to do for them might provide a quick windfall or a nice badge for your website but is ultimately going to cost you in headache, time and even real dollars. 
There are six basic customer types to cast a jaundiced eye at: 
The customer who needs major customization. 
If you’re working on a core product or offering a unique service, blowing it up and adding technical debt for one customer will result in disaster. Notice I didn’t say “may result in disaster.” This never works. You’ll be building two parallel systems in no time with the second being extremely low margin and constantly threatening to pull resources from the first. 
Note that I’m not talking about two distinct products or product lines. That’s a different story. I’m talking about tearing up a single product. 
Charging over the top for the customization may not be the answer either. The reason why independent contractors and consultants charge so much money is because they do something different (or should) with every engagement. The higher fee offsets the lack of reuse. But even if you’re charging double or triple your nominal rate for heavy customization, you’ll need to carry that burden through maintenance and support, which is tricky to anticipate. 
The customer who wants a massive discount. 

Some customers see startups and automatically think your pricing is not only going to be ultra-competitive, but that you should come in at about 20 to 50% of the costs of a known entity, no matter how much better your offering is. They’ll want a “guinea-pig” or “early-adopter” price. This is just going to undermine your value. You should be competitive, but if you can’t sell your product out of the gate without heavy discounting, you’ve got a problem somewhere. 
Note: Some of these customers might even justify the lowball rate by offering to be a launch partner of sorts, promising you that they’ll put up with your early-stage hiccups. They will not. If you stumble or delay or make mistakes, which you will, they will hassle you just as hard as anyone else. 

Oh, and never work for free. Let me repeat that. Never. Work. For. Free. 
The customer who is too big. 
When you’re a startup, corporate dynamics will always seem glacial. The way they operate won’t be congruent with the way you operate. Some large companies can take weeks or months to make a decision on a single feature. And this isn’t always weeks and months spent doing other things, this is weeks and months spent in countless meetings with every member of every department in the company, then running it up through multiple levels of management before the actual decision maker gets to make his or her decision. 
Most of these meetings will include you and/or result in additional work that wasn’t in the plan and that work may never actually see the light of day. 
Furthermore, you will have no leverage. Zero. And you may have to bring in temporary outside help, which shouldn’t be in your model early on. 
The customer is too small. 
As someone who built a business consulting for startups, I’m always skeptical of startups working for startups, and I never made it my core business or let it get to more than 25% of my recurring revenue. 
Sometimes (rare times, to be fair), they can’t or just plain don’t pay their bills. Often, they’ll expect a “family” discount because you’re both fine upstanding members of the startup community or they’re just starting out and thus should be cut a break. Not to mention they’re not always focused on the product. They’re entrepreneurs. They wear a lot of hats. So they can disappear for long stretches to work on other parts of their company and leave you hanging, like during a fundraising cycle or when they’re trying to go out and sell the thing you’re helping them build. 
All of this, by the way, is perfectly valid and normal startup behavior. I’ve done most of it myself as an entrepreneur (except the not paying bills thing). It’s just how we operate. 
The customer is clueless. 
Beware of the customer who thinks you’re awesome but doesn’t know what he or she wants or how best to integrate you. “I know we need this” is a great thing to hear. “But I don’t know exactly why” is a terrible thing to hear. 
You might also run into customers that don’t know their own business. When you offer cutting-edge technology or services using cutting-edge technologies, you can be seen as a hot-button item. These are the customers who will get 90% of the deal drawn up, or 90% of the requirements documented, or 90% of the implementation completed, and one little bit of information about how they operate will come into the mix and undo everything. They’ll think that because you’re 90% of the way there, it will only cost another 10% to fix. They forget that it will cost YOU 190% to fix. 

The customer is evil. 
Remember when I said this wasn’t about bad customers? I take that back. Sometimes the people you’re working with are just plain awful. Sometimes you just have to go with your gut and say no. Make something up. More time with the family. Whatever. 
Oh, and watch for the bait and switch. That’s when the people you interact with before the deal is signed are replaced by an implementation team after the deal is signed. Always meet the implementation folks before you close the deal.