If you haven't read part 1 of this two-part article, please read it now!
Now we've looked at what your website is for, let's jump straight in to the various types of website:
A high/increasing number of transactions (consumer or business)
If you're already making a net profit, congratulations. If not, you will probably have a business plan that tells people you'll start making x per month within y months. In the meantime you have to pay for overheads. You need to double your efforts to get to this point, including working really hard on milestones, because it doesn't come easy. At the same time I would double the "y" figure (how many months it's going to take to get into profit) so you can have a buffer against every eventuality. Having done this you should have a sum you should set aside for overheads like staff and premises. Include any stock, but try not to hold too much (but you've thought about this, haven't you?). After that, spend everything on building traffic and a high level of usability into the website. If this means £50,000 then do it. If it means £5,000 then do it. If it means £500 then you probably need to have a re-think, because no matter how talented the web developer and how crazy the deal, this isn't going to rocket you into internet stardom and consistent profits.
After 6 months you will start to see a trend. If that trend is flat or trending downwards, something's not working, and if you can't figure out how to turn it around, it may be best to pull the plug. No-one said entrepreneurship wasn't risky. On the other hand if the trend is upwards, isolate what's causing it and create a budget around a more focussed set of improvements. After that of course, you just need to measure and repeat!
Here as with any website, you need to be looking for a web developer who has experience with the kinds of features you want to build. An experienced web developer will know the "architecture" of the planned website before you even meet. Quite often there will be a "base" piece of software with bespoke elements on top to provide exactly what you need. This "base" software, which may be called a content management system (CMS), should have a minimal or zero cost. It's the bespoke pieces you need to worry about.
The law of diminishing returns will eventually help. Say you have asked for features A, B and C, which seem to be no problem at all for your prospective web developer. These cost, say, £7,500. Your other 'dream' features D and E cost £5000 each - but won't obviously make as big an impact as features A B and C. Ask yourself if you need them from day 1 and what benefit they are going to give your business.
Big organisations may benefit more from these "additional" features because their membership will be prepared to - collectively - fork out for them. So while a small trade association may be content with the A B C feature set, larger organisations with thousands of members may be looking at A B C D E F and G (and spending, say, £20,000).
A corporate website for a company whose business is transacted offline (consumer or business)
Having thought carefully about the impact your website could be having on customers, you now need to think about investing in a good website to make that happen. Not only does your website need to function, it needs to attract business, look good, and keep up with web standards. All this means that going to an instant, do-it-yourself web shop is not an option. But there are a huge range of options, so you need to be guided by someone you can trust, and who's done similar things before. Your first year budget should be in excess of £1500 (for very basic sites) and under £100,000 (because normally if £100,000 is starting to be limiting, then something is probably wrong with the planning or the talent). Typically, if your website is information-only, you'll be looking at the lower end of that scale, but you must budget for design, which takes up more time if your business has a large "footprint", and content architecture, which becomes more complex the more content you have.
Information only, or peripheral services
This is a bit of a wildcard - for example you may be aiming to build a micro-site with a game, or a mobile site, or some wacky new idea thought up by marketing. However, if you're at this stage you probably have a good idea of how to get people to the website and a good idea of how much profit you will bring in per person. If you're going to attract 100,000 people and convert half into buying your £10 product, They're going to spend £500,000. Your net profit on the purchases may be more like £200,000, so a website spend of £50,000 would be OK.
On the other hand if it's going to be difficult to quantify the effect, you may want to leave out the bells and whistles and concentrate on what you can afford.