I think I can answer this one, as I own the URL shortener Cligs ( http://cli.gs/ ).

There are two main approaches to running a URL shortener: a "basic" approach that simply does a redirect and a more advanced approach with value added services (like analytics).

A basic redirect comes in two types:

1. A proper HTTP redirect (a 30x code, of which only 301 is of proven SEO value).

2.a. A cloaked redirect in which you redirect to a page on the shortening service's website that displays two frames with one frame being the destination page and the other frame being a page you control.

2.b. An interstial redirect in which the initial landing is a page on the service's website that "refreshes" after a few seconds to the destination.

If you do option 1, then you can't make money off the redirects. You can try to make money off the people using your site. If you look at tinyurl.com, that's exactly what they do: find the AdSense block on the right hand side edge of the site.

Both of options 2 offer you a real chance to display advertising. Adjix does that and I have no idea about the kind of revenue they're seeing, but you can see their payout rates at http://web.adjix.com/AdjixLinkerInfo.html .

Option 2b is very interesting but the most annoying for the user IMHO. Linkbee allows you put interstial ads (see http://www.linkbee.com/create.php ).

The value add services on top of redirection are very valuable. There are two kinds:

a. Services that add value to the short URL creator (i.e. the service user). A great example is analytics which tells you a lot about who's clicking through. Competing on analytics is an effective business strategy ;) There are other kinds of value add services, some of which tie back to options 2a and 2b above (i.e. get the user to earn money from their links).

b. Services that are unrelated to URL shortening per se, but are valuable. I've seen a few comments in this thread about the value of the data gained from tracking traffic. I'll leave this one to your imagination - run wild!

There is a third way to make money: white-label services. I won't explain the business logic behind this, so I'll just point you to my recent announcement on the Cligs blog: http://cli.gs/MvDp1s . Cligs is offering for-pay fully branded short URLs with the analytics built right in. I know of at least one other service that will launch with a similar product soon (they're in testing!).

So how can you make money? It's a very competitive market, and the cost of entry is tiny, the user loyalty is almost non-existent, and the traffic can be huge requiring good service architecture. My point from the above is that you will be able to make money as there are ways to create value for your users you can charge for, but expect to get a few bruises on the way.

Finally, a personal note: It's a great market to learn business skills in because it's so competitive and the popular services are run by some really smart people. Can you really value the lessons you learn from competing in this market? It beats any MBA you care to point to.

And yes, you should use Cligs as it's the best around: http://cli.gs/ :)

i am a web developer & startup founder based in Nigeria, West Africa, a place where there is hardly any electricity, so my startup is run on generators & inverters, broadband isnt available, dialup is the norm here. the computers we use are 5 years old, the list goes on. but i & my team still get stuff done, because of the passion we have and i guess its because we are 'relentlessly resourcefull'. Thanks pg for another excellent write up.

The reason HN doesn't need downvotes is that HN, unlike Reddit, kills lame articles. On Reddit, users need downvotes as a way of saying an article is lame. Downvoting is the only way you can get a (nonspam) submission off the frontpage. But on HN you can flag it and if it's bad the editors will kill it.

We can thus safely assume a nonlame set of articles, and we also (so far at least) assume nonlame voters. And if you only have nonlame voters voting on nonlame articles, upvotes should be enough to pick the winners.

Me: Why don't you just do x?

Founders: Could we really do x?

Me: Sure, why not? Can't hurt to try. How hard can it be to do at least version 1 of x?

Founders: Ok [dubiously] we'll try it.

A few days later:

Founders: Pg, check out this demo of x. It was a lot harder than we expected, but we just got it working. While we were working on it, we thought "what if we did y..."

I've just passed my five year anniversary. I would do it again, but the financial instability is still extremely stressful, particularly as I enter my mid-thirties. It's always feast or famine.

My observations:

1. Pay off debt as quickly as you can. When you're broke again, you'll be grateful you did.

2. Never put all of your eggs in one basket. When you only have one or two major sources of revenue (or clients), you're as dependent as an employee, but considerably less secure.

3. No matter how unnecessary it may seem, have a real, legally binding contract for any work you commit to do. I thought I did. I was wrong.

4. Budget for marketing. You can't depend on viral marketing alone.

5. Have health insurance. It's expensive and you can't afford it, and I know you're young and healthy, but when your appendix decides it's time to come out, you'll be on the hook for about $22,000.

6. If you're working from home, set boundaries. I won't work past 8pm, and I try to take one day a week off.

Ideally, you'll have some brilliant idea, already live frugally, have no debt, and a craft a bulletproof plan. Realistically, nothing can prepare you for the experience.

The reason that the web lost faith in charging for stuff is that marginal cost is so low. What is the cost to 37signals of having another customer sign up for one of their services? If I sign up for the $49/mo. Basecamp product, they are offering me 10GB of storage as well as bandwidth and application processing. None of that has a high marginal cost. Maybe $5/mo? Plus, many users won't use all that space so it's a maximum of $5, but the average marginal cost would probably be a good deal lower.

So, while many applications might have very high fixed costs (like paying people to create wonderful applications like those in 37signals portfolio), they have marginal costs that are so insignificant that people would like to charge customers less and get more customers - often ending up with them wanting to give it away for free for maximum uptake (with the whole profit thing put off until later).

Beyond that, some applications you can't charge for. 37signals has a portfolio of applications that don't get better for users as more people use it. That's unlike Facebook, Delicious, Digg, Wikipedia, etc. If those services charged a monthly fee, it would degrade the content that the sites have. Facebook isn't useful if they only retain 10% of their user-base because they're charging $10/mo. So, some websites must try alternative ways of monieization since charging users isn't an option.

Finally, the web often forgoes charging for stuff because they get big paydays for it. StumbleUpon got $75M. Reddit commanded at least a few million. Digg is looking for hundreds of millions while having no plan for profitability. Facebook is looking for tens of billions with no plan to make money! Jaiku was bought for millions. YouTube for $1.5B. I don't need to go on. People get rewarded for the behavior of creating a site with no expectation of profits. As long as that's working, they'll keep doing it.

Personally, I think 37signals has the right idea. But that won't stop people from making free sites with no profit plan which they hope to get bought since it happens enough. Yeah, I'm not going to create visions for myself of taking over the world, but others thrive on that. As long as there is some potential of getting bought for millions and a low marginal cost, people will create those free sites.

Because Fried makes a big deal out of "not doing functional specs' when he's, essentially, doing functional specs, just not calling them that.

He's opposed to BAD functional specs, which is fine, but too many people read his title and decide it's OK to skip to writing the code first.