Are you sure it's worth it

That's my advice from a business who does the exact same thing you just menetioned, web and graphic design as well as hosting. I do it more as a service because it sure as hell doesn't make much money on a small business scale. You'll obviously need your own dedicated server, you'll need someone to administer it, because now you are responsible for other peoples revenue streams. And when the phone goes off at 3am, guess who's up troubleshooting with the admin! It's worth it to retain your customers so they don't go elsewhere, but if you're doing it to make money, it's probably not going to happen.
A little background on us, we manage about 5 dedicated servers. We stay competitive with the prices, but our net proceeds at the end of the year, aren't nearly worth the headaches
I agree with the above statements. I worked for an ISP for 9.5 years, 6 of those were as one of 3 people that *was* the business dept. We did ISDN - DS3 support along with webhosting. We did this on multiple platforms from Straight windows or linux/FreeBSD hostings, to Plesk on Windows and Linux/FreeBSD. We LOSt money on the hosting in terms of support of the customers, but we kept it specifically because of this all-in-one crap other bigger companies are doing, and our customers were leaving. We would make $700 a month on a $850 a month connection to the internet, and literally lose in support tech wages alone $50 - $100 a month on supporting that custmer's Frontpage designed webpage, or their friend the "web designer", and his calls to support.
I do not work for that ISP anymore, but I have a few friends left. They are right now trying to farm the hosting to another company.
Which brings me to my idea. See if you can find a small ISP locally that has a business dept they want to get rid of. If you want to get close to making any sort of profit any time soon, you need bulk customers fast. See if you can buy another companies Web Hosting book of business. You get an immediate customer base you can convert into customers, and you get servers that are (hopefully) decent. Of course, you may be in for a big headache doing this, or just trying to be a hosting company at all. I handled at least 5 integrations for my old company (handled as in beginning to end, not just delt with the customer's issues, I did that too) and each was a big PITA at some stage. Either finding out the servers sucked, or that the customer's were never paying, and now had to.
In this current market for web hosting, you have 2 choices. Figure out how to sell a quality product really cheap, and make a little bit of money, or sell a product and fail.
If its crap, your going to fail. If its too expensive, your going to fail. If its cheap but decent, you might succeed.
Also, customers in need of web hosting are really finicky. They will walk away for a misspoken sentence, and by nature of web hosting, you do not have the chance to resolve the problem, as they have moved on, and are just calling to make sure they do not get billed again. The service is already gone.
I would support 1000's of T-1 customers by myself, if it kept me away from Web Hosting customers. They pay me a 3 to 4 digit dollar amount every month. They complain only when its down. The web hosting customers pay me a 1 digit dollar amount, and complained every week or so. Sites not fast enough, site is down for maintenance, I want credit. Site does not support this script some guy wrote. Etc.