Holidailies Sells Out

My goal was to make Holidailies self-sufficient. Jette and I are glad to donate our time to the project, and we aren't looking to get paid. However, we do have out-of-pocket costs to run the project. If we could pull enough income to cover the costs, that would be good for us and beneficial for the long-term health of the project.

Server rental is the most significant cost, at $150/month. The server supports a number of projects, so I wasn't looking to cover the full cost, just a proportionate share. When all the costs were added up, I figured a gross income in the range of $200-300 would cover it.

On the other hand, it was absolutely critical that in seeking paid advertising, we did not violate the integrity of the site nor the trust of the community. I didn't want to crap up the site or piss off the participants. That would have far outweighed the benefit of a couple hundred bucks.

So, it was important to Jette and me that we pursue paid advertising carefully. We decided to implement two advertising campaigns: daily sponsorship and text-based ad words.

The daily sponsorship campaign sold exclusive, day-long sponsorships at a rate of $18.50 per day. We would place a graphic banner advertisement for that day's sponsor on the web site and in the RSS newsfeed. To ensure that the sponsorships would not irritate our community of readers, we limited sponsorship to that community. We solicited web writers interested in promoting their personal sites and small businesspeople who participated in the web writing community.

The second campaign, the ad word campaign, would use a third-party service to serve appropriate text ads based on page content. We selected the Google Ad Words program for two reasons. First, we liked their implementation and user experience. Second, they have support for RSS as well as web pages—or so we thought.

The Google Ad Words campaign has not gone well. In fact, I've removed the Google text ads from the site.

There were two problems. First, the RSS advertising service is in beta and we were shut out. Had I realized this up front I would have looked for a different service. By the time time I realized this the web pages were already set up, and I didn't want to bother going back to set up something else.

The bigger problem is that we were not getting good ads placed on the site. Very few people were clicking through, and the payoff when they did was miniscule. Initially, we got ads for things like Wicca supplies and other absurdly small pay-per-click ads. Then we stopped getting anything—just free public service ads. PSAs are great—and we already have several booked for daylong sponsor banners—but I'd rather pick them myself rather than get some Google default.

I'm disappointed in Google for the lack of RSS support, but not so much for the crappy ad problem. That's not really their fault; their model doesn't work well for a portal site such as Holidailies. The problem is that the context for a portal page such as Holidailies is not defined by the content, but by the community. A machine cannot tell what Holidailies is about by reading the text of the home page. It's not surprise that Google couldn't figure out what to do with us.

On the other hand, the thing that makes Holidailies such a terrible candidate for Google Ad Words makes it ideal for sponsorship. We estimated that last year, between the web site and RSS news feed, we averaged nearly a thousand visitors a day. This year may be even better. We had over 1200 visits on our opening day. Clearly, the site is being read by a large part of the online journal-writing community.

The Google ad squirrels couldn't distill the value of our community, but sponsors could.

We though a lot about the sponsorship pricing and talked to a lot of journal writers before deciding on $18.50 a day. We tried to find an amount low enough to allow individual journalers to promote their site, yet result in sufficient revenue to hit our targets. I thought that if we were very, very lucky we could sign up sponsors for about half the month (15 days). That would put us in the upper end of the target range I set. We could then use the other half of the month for comp banners: public service ads, ad swaps, bonus days for multiple-day purchasers, etc.

Here is our current breakdown of sponsorships across the 31 days of Holidailies:

Sponsorships Number
Paid 13
Comp 13
Available 5

Most of the paid sponsorship days (9 of 13) were taken by people who want to advertise product. The remaining four were taken to promote a web site or just support Holidailies. I believe that our pricing structure (the not insignificant cost of sponsorship, plus a "buy three get one free" promotion we were running) may have encouraged the skew towards commercial ads over the individual sponsorships. Still, I'm comfortable with the mix, and I really like the group of sponsors we attracted.

We haven't closed out the sponsorship campaign yet, but already I'm pleased. I thought we needed to attract at least ten days of paid sponsorship to consider this a success, and we've done that. More importantly, I think we've amassed a really cool group of sponsors and banner graphics. Initially, I feared that the ads would detract from the site, but I don't think they do. I do wish we had more individual sponsorships (which is why we discontinued the "buy three, get one free" promotion); individual writers rarely get such an opportunity for exposure.

We aren't even halfway through the first week of Holidailies so it may be premature to be drawing conclusions. Instead, I'll propose some tentative observations:

  • Paid sponsorship is a viable way to (partially) support community portal sites.
  • Conventional ad word mechanisms are a poor match for such sites.
  • Commercial ads are easier to attract than individual sponsorships.