Jeremy Liew has a very nice piece for anyone interested in online media business models or even for anyone interested in the "build a web property that is popular and figure out how to monetize it later" approach to starting companies. It is a really short summary of the kind of traffic you should expect/need to get to build a big business based on ads depending on the type of content/audience you are hoping to aggregate.

I like the piece for two reasons. First, most consumer Internet entrepreneurs I meet these days do too little thinking about business model, preferring instead to think that either Google or Yahoo will acquire them or that they will just do the "ad" thing. This is the west coast's "build now, monetize later" investment hypothesis taken to the extreme. To that end, the piece does a great job of showing that banking on this approach has the inbuilt assumption that your web property is basically going to be struck by lightning. It happens— but you shouldn't count on it.

I also really dig the post because in looking at the third model for success— vertical content sites that are best positioned through content for certain types of advertisers— I realized what it is that I dislike about the advertising bet in startup business models: as an entrepreneur, it reduces your optionality (is this even a word?).

Having options when the initial plan doesn't work out is a great thing, especially for small companies in quickly developing spaces. But as I looked at the third, and most viable of the ad models, I realized that there is little option if the vertical content you create fails to attract advertisers or a significantly attractive audience. When that happens, you don't have the scale to just pump generic ads (model #1), and you don't have a model that will continue to itself scale without continued investment. You can try a roll-up across verticals but something like this only gives you scale on the sales side of the business and not on the core value propostion.

If you read the comments (and some of the very interesting posts linked therein), you will see people discussing alternative business models that may work better at maximizing optionality in the face of failing to reach a certain scale. My two favorites are product sales (preferably mass customized like Tabblo's because these avoid pricing pressure best) and subscriptions (but very high value like SmugMug's or eHarmony's), but there are others including "freemium" (though you need a little bit of scale to make this one start to work) and marketplace (where you need a very narrow niche or even more scale).

All of these other models seem to offer more levers to tweak when things don't quite go according to plan. Which is almost always— and a good thing to keep in mind when looking at some of the high hurdles that Jeremy lays out.