Publishing tip: To earn more money, try showing fewer ads

Allow me to offer you a completely counter-intuitive piece of advice – one that’s nevertheless helped me to increase income from the networked ads I publish on my websites.

To earn more money, try showing fewer ads.

You might think that online advertising works linearly: More ads = more money. (This equation certainly seems to reflect the thinking behind many ad-laden newspaper websites I read.)

But placing more ads on your website might actually hurt your ad network earnings – and not just because you’d be driving readers away with a lousy site experience.

Ad networks, such as Google’s AdSense (the network I use most often on my sites), often use complicated proprietary algorithms to decide which ads to show on your website, and how much they’ll charge the advertiser for each click. AdSense uses what Google bills as a real-time auction system to determine which ads show in the AdSense slots on a publisher’s site.

But it’s not simply a case of the highest bidder wins the space. Google’s system is trying to determine:

• what ads are relevant to the content of the webpage, or
• what ads are relevant to the interests of the reader on the page, and
• of those potential ads, which one would most be likely to elicit a click, then
• do the math to figure out if an ad less likely to elicit a click could actually earn the publisher more money, even factoring in the lower chances of getting clicked.
• And after all that, the system has to determine if the selected advertiser is behind or ahead of schedule for the amount of money they’ve been charged from their daily ad budget that day. If they’re too far ahead of schedule, the system ignores their ads for while and this selection process starts over again.

Finally, even after a reader clicks on an ad and the advertiser is charged, if too few of your readers come through for those advertisers after they leave your site – they fail to buy something, register on advertiser’s website or simply view enough pages there – Google might charge those advertisers less money for those clicks from your site. It’s called “smartpricing” and can cripple your revenue.

That’s more variables than I ever had to account for in my high-school calculus class.

If you don’t understand the way your ad network’s software “thinks,” you never can expect to make a living wage income from networked ads. But if you take the time to learn about these systems, you’ll come to realize that certain pages on your site actually can work against you.

If certain pages attract readers who aren’t likely to click on ads – or who are “flaky” and likely to click but never to do anything on the advertiser’s website – you’re likely better off without ads on those pages.

As a publisher, you should want to present the ad network with a series of pages that include content likely to attract readers who are both interested in advertisers related to that content and looking to buy or otherwise engage with those advertisers.

The more such readers you deliver, the more lucrative ads you will get from the ad network, earning you more money. The fewer you deliver, the less the ad network “thinks” of your website, which will lead to less lucrative and possibly less well-targeted ads, as the ad system casts about trying to find something that will engage your readers.

Here’s an experiment I’ve tried with great success:

I divide my websites into multiple channels in my AdSense reports. I’ve set up my content management system to publish different topics and services on the site into different URL paths, which can be tracked easily as channels by the AdSense reporting tool.

Then I look at the eCPM (earnings for every 1,000 pages viewed) for each of those channels, as well as the average eCPM for the site as a whole.

For every channel that earns less than 50 percent of the site’s overall eCPM rate, I take the ads off that channel.

The result? The last three times I’ve done this, I’ve yielded an average 15 percent increase in network ad revenue. That’s not a 15 percent increase in eCPM. That’s increase in bottom-line income. (ECPM, obviously, rises way more than that.)

This analysis also helps you identify which sections and topics on your website are earning you the best return on your reporting, writing and development investment, giving you better information with which to decide how you’ll spend your time and resources in the future.

I’m not arguing that you should ditch all your public interest reporting in favor of chasing the highest eCPMs. Or that you should drop lower-performing content from the website. I love Slashdot’s analogy of a successful website as an omelet. You need the right mix of ingredients for everything to come together in an attractive, tasty whole.

But just as not every section and feature of your website serves the same editorial purpose, several sections and features of your site will serve different business purposes as well. Some sections earn money, and should display ads. Others don’t, and should not display ads. Placing ad slots on unproductive sections of your website can hurt networked ad performance elsewhere on the site.

Obviously, you can’t try my method of culling ads from sections of your site every day. Eventually, you’ll come to a point where you cut too far, and overall ad revenue declines. If that happens, hurry to replace those most-recently eliminated ad spots.

After a while, you might try returning ads to some of those borderline underproductive channels, too. In my experience, some sections that performed poorly in the past can begin to perform better in the future – thanks to your site attracting a different mix of readers, or audience tastes changing. Even if that’s not the case, you might still see a nice increase in ad revenue from returning ads to some of those channels for a few weeks before their presence begins to drag down revenue elsewhere.

If you’ve the skill for it, properly timing the addition and deletion of ads from marginally productive channels can help you maximize ad revenue. Personally, though, I prefer to leave ads off channels once I’ve yanked them. I’ll return ads only if I see a change in traffic coming to one of those channels, usually following some news development on that topic that’s bringing new visitors to the site. Even then, I’ll check the revenue figures after a month or so, and if the channel eCPM remains below 50 percent of the site’s average, the ads are coming off again.