Making Banner Ads Suck Less

The unusually-seen Kurt Gray wrote this; it's funny, to the point and more honest than may make everyone comfortable. Everyone knows banner ads suck; Kurt tells you a little more about why they're still around, explores some things that might make them better, and generally straightens the dope. We're doing this in conjunction with K5, who's also got the story. So, hop back and forth, and we can all get a merry meta-discussion going.

My name is Kurt Gray, I'm the lead programmer for OSDN's ad system which serves ad banners on sites like Slashdot, Freshmeat, SourceForge, Themes, and partnering deal with Kuro5hin, etc. I open sourced our ad delivery code sometime last year and have been maintaining it in-house here as well. My quest now is to create a better ad banner delivery system, not only better for you the audience but also more useful to our sponsors. So I have some ideas about our ad system that we want to pitch to you all sitting out there reading this and your feedback on these ideas would be of great value to us. Note that this is being posted to both K5 and Slashdot because we want to get feedback from everyone we can.

First let me address two issues that have been discussed on Slashdot just recently on Micropayments instead of ads and Ad banners may soon get bigger and how these issues pertain to ad banners on the OSDN sites.

Why run ad banners? What about a tip-jar, or subscription fees, or micropayments, or donations, or bill-the-ISPs instead of ad banners?

When you're running a web site, depending on your content, your audience, the size of your staff, your overhead costs, the size and nature of your audience, and many other factors, it might be possible to get by on just subscription fees, or micropayments, or some other revenue model that does not involve selling banner ads. But the size of the audience on OSDN's web sites and the nature of the content within is such that the subscription models break down. For a network of this size and content ad banners are the only realistic way to cover costs and hopefully earn a little profit (someday we hope). Another way of looking at it is to ask yourself why does Yahoo, CNet, and ZDNet still rely on banner ads? Because for a web sites that have a lot of traffic no one has proven that there is a better way to earn more revenue with less overhead. In any large media company, advertising is it. Even with print magazines the subscription fees and cover prices don't come close to covering the costs for a large circulation magazine: the subscription fees and cover price is just a barrier-of-entry to assure the advertisers that the readership paid to read the content and therefore is the right audience to see their ads.

....but ad banners don't work! There's too many ad filters now days!

Yes, a lot of people, even entire ISPs, have ad filters and proxy rules to block out banner ads but even still there are plenty enough ad impressions delivered every day. In fact those who filter ads are doing web publishers and advertisers a favor by making sure that no time, bandwidth, or impressions are wasted on people who definately will not respond to any kind of ad. So please, filter the ads out if you feel that strongly about it, in fact, I'll pitch you some ideas further on in this article in which our ad system could help you filter out the ads which is why I'm posting this.

...but too many people ignore banner ads, and nobody clicks on them! Advertising sucks! Free your head! Prioritize, man!

Yes, many people, including myself, scroll right past banner ads and ignore them completely. But chances are you did glance at many of the ads in a web given page, perhaps you saw a logo or brand name. In that sense the ad delivered just what it intended. It's called "branding": advertising for the sake of increased brand recognition and its most of what large advertisers hope for when advertsing in any medium including the web. Smaller advertisers will obsess over response to each ad, whether that be a click, or even a sale, and thus they become very unhappy when the click-thru is not to their satisfaction. So just because click-thru percentages are low across the board doesn't mean Internet advertising is doomed, but rather advertisers expectations and ad pricing schemes are changing accordingly. The smallest fish in the pond may be doomed but the pond remains.

What can we expect from OSDN web sites as far as ad banners? Bigger fatter ads? More ads per page? Flashing noisy ads that will read my browser cache and report all suspicous keywords to the NSA?

As you might expect, we are debating internally what OSDN sites can do to stay competitive in the ad banner business. Right now we are not competitive in many areas: we only accept the most basic ad formats, most OSDN sites only accept one ad size, our average click-thru rate is as low as anywhere else, and our rate card prices are higher than most. We've been able to get away with it so far because our web sites are very well known and our audience has just the kind of demographics advertisers drool over, but lately its become a buyers market, the ad budgets are drying up and the few big advertisers still spending online are having their way with the web publishers left groveling for the business. It's times like these when advertisers can force outrageous new ad formats down the throats of the web publishers, and other web publishers are stepping up their ad offerings to entice advertisers to their space -- it's a free market economy after all.

So what are we doing about it? First we're telling our sales people to go after more main stream advertising accounts: entertainment, auto makers, food and beverage, whatever we think fits our audience. Second, we're looking at which newer ad formats and what we're willing to accept. Third, I have to rewrite our ad delivery system to improve our ad targeting: platform targeting, geotargeting, and topic targeting at the very least. Along these lines I also have some ideas I want to bounce off you there reading this here article...

Let the users control the ad delivery. User preferences. Ad filtering. User feedback. Interactive, or as George W. would say "Interactivfulness"

Here's a few scenerios, ideas I've been pitching around:

Comment forums for each ad banner:

What if you could comment on the ad banners, such as each ad banner has its own discussion forum? So if an ad bothers you, offends you, confuses you, entices you, anything about that ad, you can speak and be heard. Let's face it, many ad banners suck because nobody tells the ad agency that the creative needs improvement. On the other hand the ad may be messing with your browser and you just want someone to know about it. Or maybe you wanted whatever was being advertised, you clicked, and you still didn't get the information you were looking for, the ad feedback forum would be the place to get a response on that.

Turning off annoying ads:

Suppose you become absolutely sick and tired of seeing that "Fawking DSL!" ad or that "Punch the monkey" banner, suppose you could click a link right next to the banner "Never show me this ad again or I swear I will lose it and someone will have to call security." And you just click that link and bam, you'll never see that ad again. The number of people who turn off a particular ad could be a way of truly knowing how counter productive certain ads are.

Choice of ad topics and categories:

What if you could select which kinds of ads you want to see, and which kinds of ads you don't want to see? For example what if you could explicately set your ad preferences so that you're are more about networking, movies, gadgets, and events but you don't want to see ads for alcohol, web design, or luxury items... and these ad preferences would apply to you within whole OSDN network of web sites. Would we use your information to for demographic studies? Yes absolutely, we'd tell advertisers that we have X number of people over here who explicately told us that they'd prefer to see ads about their kind of product. The overall effect we won't waste our effort chasing after advertsiers that have nothing of interest to our community and we won't waste your bandwidth downloading ads you don't want.

What about ad system karma?

I'm thinking there could be a point-based reward system that gives you credit for everything you do that helps our advertising business. As you accumulate karma points in our ad system you could redeem them gain access to an extended set of features in the ad system itself...

To increase your ad system karma you could (Hypothetical examples)

• 1 point for every time you load a paid ad
• 0 points for clicking on an ad (I don't want to encourage excessive ad clicking)
• 50 points for loading bigger ads
• 100 points for loading a pop-up ad
• 500 points for filling out an advertiser's survey
• 100 points for loading a Flash ad
• 300 points for posting a meaningful critique on an ad
• 200 points for alerting us if an ad is broken
• 500 points for helping us test an ad before it goes live

(Just assume for the sake of this disussion that this point system is mostly immune to people running bots to accumulate points. We're still in hypothetical land here.)

Redeem your points to gain access to such features as (Hypothetical examples)

• Turn off all ads
• Upload your own ads
• Get stats on the ads you uploaded
• Specify which sites you want your ads to run on
• Whetever else anyone can think of...

Note that this entire karma point system is just my own personal ideas and not officially sanctioned by anyone else working here. I figured I'd bounce this off you all out there in the audience and see how it plays with you all.

How would ad system karma affect web site user karma?

It wouldn't. The ad system is totally disconnected from any web site user database. Our ad system runs ads on many web sites, so even if we felt compelled to tie it into the user accounts of any web site it would be a lot of work, too much work, and I don't see any reason to even attempt it.

So the ad system would have its own user accounts independant and unrelated to web site accounts. Does that complicate things? No, the ad system user account is low maintenance, transparent, maybe as simple as cookie, nothing too visible, not in your face all the time nagging you to come play. The ad system preferences web page could be one click away, simple web form, nothing too fancy.

Hey I don't like you spying on me! I'm going to wear a metal bowl on my head and warn the others that you're all sneaky opportunist-type people. You are one of them.

That's OK. I have my metal bowl on too. As far as these ad system ideas go, you wouldn't need to have an ad system user account if you want to be anonymous and outside the loop as far as the ad delivery goes that's fine. This user account would be something you'd actively choose to create, and if you don't bother doing so then fine, you're anonymous, unknown, you'll see the normal general rotation of banner ads, and maybe later hopefully you'll find that out food tastes better when you try out some of these features and take advantage of the bonuses.

We're a community, damnit! We're not your ad-clicking sheep! If you can't sell ads then that's your problem! One day this web site will be free of your commercial opportunist tryannical business, all the trolls will leave, this site will be cool again, and then food will taste better!

These web sites have grown way beyond the realm of affordable to operate by volunteers and donors. If OSDN and/or VA collapsed someday then the OSDN web sites would not be simply released back into the wild but rather be liquidated as assets to the highest bidder, and you can bet the new owners would gladly run these sites into the ground for every last penny they can quickly earn from them. So at least you can be glad the original founders of these web sites still work here and they care a lot about how this web site works for you, the community. And if we're not able to turn a profit here despite our best efforts, whoever ends up grabbing our helm here will most likely toss this whole crew overboard, and I can assure you that the new crew will care far less about "community" then we ever did. But that's not your problem anyway because there are plenty of other web sites out there like this one, and if you log off now you may even discover that there is whole world of amazing life outside the Internet, I don't know much about that myself so I can't descibe it to you but I've downloaded pictures of it. So is this as good as it gets for these web sites? No, we can do better here, and last week resolved to be a lot more focused. We're determined not to give Jesse Berst and his ilk any reason to gloat.

So I can't think of what else I was going to pitch here. So please if you have feedback on any of the ideas pitched above then post them here.

Kurt Gray, OSDN, ad system engineer