What should you charge when you start? Don’t go for two times the AdSense CPM. You do not want to leave money on the table if it is there. You’ll find it much more difficult to raise rates than to cut them, and retain your advertisers. Generally, the only time you can get away with a price increase is when you’ve sold out your inventory and are turning away advertisers.

Let’s start, instead, with the six times AdSense figure. How does that compare with the CPM being charged by those other websites you examined? Is your figure competitive with their rates? If so, go with that. If it’s too high, lower it to a figure that you believe will be competitive. But don’t undercut the market by too much. (See my rationale above about raising prices.)

Now that you have a CPM rate for each ad slot on your site, it’s time to put them into packages. I’ve found it much easier to “sell” and advertiser, and to close the sale, when you offer a limited set of ad options. Don’t just give them a CPM rate and leave them to do the math on how much the want to spend, for how many ads.

On my personal sites, we sell ad packages for $100, $250, $500 and $1,000. Those price points get advertisers a fixed number of ads, at our various CPMs. Having a fixed set of defined packages makes your rate card easier for advertisers to comprehend, and will help you to better track your billing and payments.

Step 5) Reality-check time

Now you’ve got some numbers – how much you are earning from each ad slot from AdSense, as well as how much you would charge for each ad slot through direct sales.

Let’s put those numbers into reality: your financial reality. How much money do you need to earn from this site each month to make it viable? Sum up the expenses that the site’s revenue needs to cover: Hosting charges, reporting expenses, business fees, your wages, etc. Now divide that figure by the number of page views AdSense said your site served.

Multiply that number by 1,000 and you’ve got your target page CPM. That’s the amount of money that your site needs to earn to for you to cover your expenses and make the living you want.

How’s that number compare with your current page CPM from AdSense? How does it compare with your retail CPM that you just decided to charge your advertisers?

If your current AdSense page CPM is more than half of your target per-page CPM, congratulations. You might not need to learn about direct ad sales, after all. You might be able to make your site financially viable by putting some extra effort into boosting traffic through posting more frequently, creating more “evergreen” content and better SEO, with the expectation that greater traffic could raise your AdSense revenue. If you are more comfortable going that route than moving immediately into direct ad sales, try that first. This article will remain here, if you decide to give direct sales a try in the future.

If your target page CPM is between two and six times your current AdSense per-page CPM, then direct ad sales might be able to bring your site up to viability. Skip ahead to step three and hang in there.

If your target page CPM lies between six and 12 times your current AdSense CPM, you’re going to need to take both routes. You’ll need to boost your traffic, using those methods I just mentioned, as well as pursue direct ad sales. No worries, though, other publishers have done this and made it work.

What if your target page CPM is more than 12 times what you are making per page on AdSense? Now it’s gut-check time. A gap that wide between your needs and current reality is telling you something that you need to hear, however unpleasant that may be. Are you targeting an audience that actually exists insignificant numbers? Are you providing content of value to that audience? Are your pages coded efficiently, so that search engines can find a properly index your content, or do you need to learn about search engine optimization (SEO)?

I’d suggest working to improve the AdSense page CPM of your site before proceeding to direct ad sales. Read up on SEO (click the website design and usability links in OJR’s archives to start). Try 10,000 Words, A List Apart and Webmaster World for more tips. Sharpen the focus on each page and talk with people in your community, offline, to see if your site is heading in the right direction to engage readers.

Or, maybe, you need to be more conservative about your expenses. Drop the freelance or outside help and plan to do more on your own. Or, look for less expensive hosting or how to cut other expenses.

Once you’ve got the AdSense page CPM within that 12:1 ratio with your target page CPM, then you’re ready to proceed.

Step 6) Readership survey

You’ve got market data, a rate card, and you’ve passed the reality check. But there’s one more step to be taken before soliciting your first sale. You need a compelling portrait of your readers to take to potential advertisers.

That portrait is a readership survey. Your website analytics will provide much of the data, including:

But you need more, including:

And possibly:

The easiest way to collect this addition demographic information about your readers to sign up for the Blog Reader Project survey and to direct your readers to fill it out. You’ll see links to these surveys from time to time on many top blogs, including Talking Points Memo and DailyKos.

You don’t need to know about technology, beyond including a tracking code on your page and directing your readers to a link. The Blog Readership Project will track the data and report it back to you.

The problem with this data is that it is self-selected. Ideally, you want a random-sample survey of your readers. Quantcast provides some random-sample data for sites its tracks (log in to your Quantcast user account to see it), but for most sites, the sample’s just too small to be accurate.

You can create a readership survey using SurveyMonkey or some other online tool. (The format of the Blog Reader Project’s survey is fine, if you’d like to use that as a model.) But how to get the random sample?

I’m sorry, but I haven’t found a non-techie way to do this yet. (And if an OJR reader knows of one, pleas respond in the comments, or by clicking my byline and e-mailing me.) Here’s how I do it [warning, geek talk ahead!]:

Start by figuring out how many completed surveys we need. Let’s shoot for 400, which would give us a margin of error around five percent. Not great, but acceptable. If we assume that 10 percent of the people we ask will complete the survey, we need to ask 4,000 people to take it.

We should conduct the survey over a week, as not to skew the results toward readers who visit only during workdays, or evenings, or weekends. So you need to look up how many absolute unique visitors your site gets in a typical week. Divide 4,000 by that number, and that’s the probability that your should ask a given reader to take the survey.

We’ll use cookies to determine who gets the survey invite. If you know how to program within your publishing system, you can do this. If not, you can’t. And if you have fewer than 4,000 readers a week, you might as well just direct readers to the Blog Readership Survey anyway.

If you can program, though, use a random number generator. If the probability of getting an invite is .15, for example, generate a number between one and 100. If the random number is between 1 and 15, set the survey cookie to “invite.” If not, set the survey cookie to “no.”

Place a section in the site’s template to display the survey invitation. (I’d put it at the top of the center editorial column.) Check for the presence of the invite cookie. If it is not there, set it, using the random number generator. If the cookie is present, and the cookie reads “invite,” show the invitation link. (e.g. “Please help us to continue publishing this site by taking a short, anonymous reader survey” or something like that.) If the cookie reads “no,” don’t show the link.

Once an invited reader completes the survey, reset the survey cookie to “no.” Otherwise, keep it as is, showing the selected reader the invitation on every page until he or she completes the survey, or the week is over.

Next week, I’ll write about what to do with this data and how you can use it to start making ad sales.

Learn how ads have performed in specific situations on other websites by looking at Google’s eyetracking “heat map” for online ad placement, as well as its tips for AdSense implementation.

You’ll note that I am writing under the assumption that you already have a site up and attracting readers. I think that you’ll find it darn near impossible to sell advertisers on reaching your readers if you do not have any readers yet, and haven’t established a track record of attracting readers in the past. That’s why it is so important that you start your website before you need the income from it. Start while you still have another “day” job, ideally, or else you’ll need to rely on financial support from another source (such as a spouse, savings or investors) while your build a readership that you can take to advertisers.

You might also note that I haven’t written a thing yet about actually approaching a potential advertiser and asking for a sale. (That’s what I’ll write about next week, in part two.) An ethical journalist-entrepreneur must do a substantial amount of reporting, about his or her readers and his or her market before he or she even thinks about asking for or accepting a sale. You’ve got work to do now. Go do it.

I’ll see you back here next week.