Digital advertising can be a total cesspool. In fact, the last decade or so has been a study in how advertisers can create advertising that makes the general public hate advertising even more than they already did, which is no small feat. Last post, I talked about how advertising online isn’t vastly different from advertising in other spaces, so how did things get so bad?
The first thing to understand about advertising in the digital space is just how many opportunities there are to advertise, which makes modern media planning a potential nightmare. Of course, every major evolution in publishing has been accompanied by a more complex landscape in which to purchase advertising space. For example, buying advertising space in magazines used to be much simpler: there were only a handful of magazines that were read all around the country. Publishers discovered that there was money to be made in appealing to niche markets, and in a relatively short period of time, the number of magazines looking for advertising ballooned to several hundred. This happened again in the world of television. When television was restricted to a handful of major networks that broadcast over the airwaves, the available number of ad units up for sale was relatively small. When cable exploded in popularity, the number of stations offering ad units jumped into the hundreds. In both cases, advertisers adapted to the changes by researching these new outlets and adding them to their tool kit when considering ad buys. The problem with applying this tactic to the Internet is that the number of webpages looking to sell ad space didn’t balloon up from a handful to a couple hundred, it exploded into the millions.
In order to cope with the vast number of sites on which to place advertising, specialized media companies, called ad networks, came along to act as middlemen between advertising agencies and web publishers. So how do these networks work? On the publishing end, the website owner places standardized spaces on their site, usually with fixed dimensions, where ads will be shown to the viewer. When everything is working correctly, each time a visitor loads the site those ad spaces are filled with ads that are drawn from the ad network’s database of available advertisements. This is great for the publisher because they can simply opt into one or more ad networks and let the network handle their ad placements. This saves them the trouble of hunting down their own sponsors. It is also great for the agencies because their media buyers buy advertising space from a handful of convenient ad networks instead of millions of individual websites.
There are variations on the basic model, which I will address later in this post, but basically agency media buyers submit a set of digital ads along with a set budget for the ad campaign and when the agency’s budget is exhausted, the submitted ads are no longer served. On the other end of things, the ad network serves those ads to their partner websites, charging the agency for each view, and a percentage of that income goes to the partner site for hosting the ads. To show how this all fits together, let’s imagine a fake soda company, Fako Soda, and a fake ad network, BestAds. An agency representing Fako Soda wants to advertise online. They develop a series of banner ads that promote their brand and then contact the BestAds network to place those ads on websites. The agency has a digital media budget of $1 million, which they plan to spread across several networks and other outlets, so they tell BestAds that their budget for the BestAds network is $100,000. BestAds then places the agency’s banner ads in their database and sets them to be served on their partner websites, along with the other ads other agencies have purchased on the BestAds’s network. The Fako Soda agency selected a pay per impression model, so every time an ad is served to a website, and hopefully seen by that visitor, the agency’s budget with BestAds is charged a fixed cost. This value, in industry terms, is known as CPM, which stands for cost per mil (Latin for one thousand, same root as millimeter). In our example, BestAds pays $3 per 1000 views, or $0.003 per visitor, so when BestAds has placed the ad roughly 33 million times, the money the agency paid would run out and the campaign ends. While this example provides a good basis for understanding how ad networks work, the system gets more complex as agencies are given more flexibility in how they buy their placements. For example, agencies pay a premium for above the fold horizontal banners. They can also put constraints on how long a campaign runs regardless of budget, such as limiting the ads to run only during the months where they are relevant.
So far, ad networks don’t seem like such a raw deal. They provide a valuable service to both publishers and advertisers, they take their cut for services rendered, and everyone is happy…right? Well, not exactly. Ad networks are in the business of placing ads. They don’t necessarily care where those ads originate from, what their quality might be, or what kind of business is being advertised. As a result, ad spaces are served with ads that people find offensive: they are pornographic, they use annoying flash animations to ‘catch the eye,’ they appear to be scams, they try and trick you in some way, or they employ any number of other slimy tactics. This kind of advertising ends up being pretty terrible for mainstream advertisers, the publishers, and the visitors. For the mainstream advertisers, it taints their ads, and the brands they represent, to be served next to disreputable advertising. For the publishers, it can make their site look trashy by filling the non-content parts of the site with low rent advertising. For the visitors, it ruins the browsing experience when their attempts to read an article, or check out some funny pictures, are interrupted by advertising eyesores.
The problems don’t end there, however. Even when dealing with networks that make their value-added a vetting process to weed out undesirable advertising, many still do not do a good job of vetting the partner sites. A recent piece in Advertising Week showed that even respected ad networks which promise to avoid a multitude of advertising woes, such as ads being placed too close to one another, too many ads being placed on a single page, and ads being placed next to questionable content, often fail to deliver on their promises. These networks may have a system where questionable ads and/or content can be reported but that doesn’t necessarily lead to them being blocked. Or they may have a real-time bidding process that forces buyers to make quick decisions about seemingly valuable advertising real estate, leaving them not enough time to fully research every site in a given ad network’s partnership network. As a result, mainstream advertisers representing blue chip brands have to be very careful when picking an ad network in order to avoid having their ads served alongside trashy ads or on sites with questionable content. And even if they pick a good network, because publishers are generally free to sign up with multiple ad networks, their ads can end up on the same page as ads for trashy sites served by a different network.
Another problem facing digital advertising is how little publishers end up getting paid, whether they buy into a network or independently acquire sponsorship. Assuming the publisher is using a per impression payment system, each ad unit a publisher places on their site has its own CPM, which can fluctuate based on how valuable the real estate is it holds on the page. Publishers get paid for every ad that gets placed on their site, so the more units they have to sell on each page, the more they get paid for each page impression.
The value given to each unique page view is called the eCPM, or effective CPM, and represents the amount the publisher will get paid when every ad unit on a page loads. For example, let’s say you have just started a website to house a new web series you are producing. This site posts a new episode every week, so most people will visit the site once a week to watch each new episode. Since the new episode is hosted on the home page, and archive viewership is rare, the majority of your income will come from front-page ads. You place three ad units on the front page: a horizontal banner above the main content, a vertical banner along the left side of the main content, and a horizontal banner below the main content as a footer. It is not uncommon for a single ad unit to have an approximate CPM of $3, or, in other words, be worth about $3 per thousand impressions, so for our example let’s pretend the top ad space has a CPM of $4, the side gets $3.50 and the bottom gets $2.50. This would give the home page an eCPM of $10. Now let’s pretend the series becomes an overnight success, and begins consistently attracting about 1 million viewers every update. With an eCPM of $10, this would put about $10,000 dollars in your corporate accounts every week.
This might be a lot if you could produce the content by yourself, but to maintain the show you have production costs (writers, cameramen, locations, equipment), and to maintain the website you have upkeep costs (coders, bandwidth, storage/archive), not to mention overhead costs (office, employee benefits). Even when you move the variables around things don’t change much. The mere fact that a site attracts millions of visitors makes it more expensive to run. Sites like IGN, Bleacher Report, or Penny Arcade that are able to attract the big numbers require more bandwidth, more staff to keep things running smoothly, and more overhead.
To further illustrate this point, let’s take a deeper look at one of the big sites I mentioned before: Penny Arcade. The Penny Arcade eCPM for a year’s worth of front-page ad impressions is around $525,000. We know this because it was effectively disclosed as part of their Kickstarter campaign. Also, according to 2010 numbers, they have roughly 3.4 million readers. This is three times the viewers from our example, for a website that updates three times more often than our example, that annually makes only marginally more than our example. In other words, despite being one of the most widely read webcomics in the world, Penny Arcade still has an eCPM for their front page that is lower than our example. To put this in perspective, a one-hour documentary sold to a cable network with a likely audience of roughly a million in its first airing, like those broadcast by Discovery or National Geographic, are typically purchased for a flat rate of between $125,000 to $175,000. That’s already more than ten times as much as the $10,000 you would get for a million views online! And of course those cable broadcasters intend to make a profit from those documentaries, so you can be sure they expect to make more than $175,000 in ad revenue from the single episode. In short, content is expensive to make, but can be profitable to publish… just not on the Internet.
Popular sites with dedicated fan bases don’t have to rely exclusively on advertising for income. Penny Arcade, for example, sells a large variety of merchandise that ties into the characters, art design, and themes of the site. Smaller sites, on the other hand, are unlikely to find this a profitable model. Instead, most focus on simply trying to raise ad revenue. In any site where viewership has stagnated or hit a plateau, the solution is to simply increase the number of ad units on any given page. It makes sense: if the CPM of each ad unit is relatively similar, then the easiest way to raise eCPM is to increase the number of ad units. A side effect, however, is that the site becomes cluttered with advertisements, sometimes to the point where the actual content of the site becomes lost in the noise. Several ad networks have tried to combat this by putting limits to the number of ad units you can place with their network on any given page, but clever webmasters got around this by signing up with multiple ad networks and placing the maximum number of ad units each network would allow on each and every page. As much as I hate this behavior as a consumer, I can’t really blame the webmasters for doing it. If running the site is going to be someone’s career, or even support a small company, there has to be revenue to cover the site’s overhead, including living expenses.
The low payment structure isn’t the only problem with digital advertising. A whole host of other problems stem from the different pricing models available. Up until this point I have spoken mostly about the CPM model, as it is a commonly used and easy to understand payment system. It is not, however, the only way.
Advertising has always been in that bizarre space between an art and a science. The making of advertising is definitely an art. It is undeniably a creative exercise, and when done correctly the result is not less humorous, emotionally engaging, or thought-provoking than a good book, play, movie, or painting. On the other hand, beyond being humorous, emotionally engaging, or thought-provoking, it is art with a purpose: to move product. Ultimately, companies spend money on advertising because they want to grow their market-share (priority one), or at least maintain their market-share against aggressive competition (priority two). Companies are also risk averse; they have a fiduciary responsibility to spend their money wisely to improve stockholder value. And advertising isn’t cheap, so the risk with advertising is typically a substantial amount of money. This puts advertising in a place similar to Hollywood movies: an expensive creative process that is constantly under pressure to guarantee results. The way this has manifested in advertising is through a series of pseudo-scientific experiments, or analytics, which run the gambit from the completely absurd to the incredibly insightful. The idea of analytics is to track the effectiveness of an advertising campaign separate from a noticeable uptick in sales.
In the TV era, to track the effectiveness of advertising, agencies would commission focus groups and nationwide polls to try and measure recall, or whether viewers of a commercial remember anything about it after watching it. This became a controversial way to analyze ad performance, as many detractors pointed out that making an ad memorable and making an ad that affects sales are not necessarily one and the same. The classic example being any ad with attractive, scantily clad women will rate high in recall but low in sales, since people remember the sexuality of the ad and not the brand that was behind it. In the digital space, this style of analytics has evolved into two different advertising pricing models, cost per click (CPC) and cost per action (CPA). Cost per click is, unsurprisingly, when the agency pays the publisher, through an ad network when applicable, every time a visitor clicks on an ad. Cost per action is when the agency pays the publisher every time a visitor clicks an ad and then takes action on the landing page the ad sends them to. The theory is that tracking people’s behavior in relation to the ad, and making the payment system tied to that behavior, will make the advertising spending more efficient. Advertising is intended to drive people to purchase things, and in CPC and CPA advertising money is only spent on those people who were actually motivated towards buying. Except, advertising is far more complex than that.
CPC and CPA are both variants of direct-response advertising, with a different pricing model afforded by the wild west of the digital frontier. Direct-response has always been a feature of advertising, in one capacity or another, but it has never succeeded as the main event. In the TV world, the epitome of direct-response advertising is infomercials. Infomercials are very effective at driving sales, but viewers hate them. People don’t like the hard sell, even if it works. That’s why infomercials are relegated to selling various household doodads, while major brands stick with creative advertising. CPC and CPA have distorting effects on the web landscape as well. Because the money that publishers get is tied to getting clicks, people have tried using tricks to get visitors to click on links. They run ads that pretend to be flash games, they run ads that look like legitimate links to content but aren’t, they run ads that offer outrageous and non-existent deals on premium products, and so on. Advertising can and should be about more than being slaves to analytics, it should be about making people understand the heart of a brand, whether you call that the unique selling proposition, the positioning, or something else. As Bernbach so smartly put it: “Nobody counts the number of ads you run; they just remember the impression you make.”
Even if all of the issues with payment models get solved, there are still more issues with the digital advertising market. The digital space, probably for many of the reasons I’ve previously mentioned, is still not a very desirable place to advertise. Often ad networks run a shortage of ads against an abundance of partner sites and so ad units go unfilled because no one has paid to fill them. This is referred to as having low fill-rates. This is especially bad when new technologies, new broadcasters, or new networks enter the picture. Another issue related to low fill-rates is repetitive advertising. This happens when there are enough ads to go around, but only a small number of agencies are currently buying ad space for a small number of unique ads. This causes the ad units to be filled with the same ads for the same products over and over again. For example, if Toyota is the only brand currently advertising heavily on Hulu, then I might see the exact same Camry ad every single ad break of every show. It’s well established that such repetition is death for advertisements. Advertisements are designed to be seen again and again, but with enough of a break in between that they don’t become grating and irritating. It’s like hearing the same joke again just a few minutes after hearing it the first time. It’s never as good. Then imagine hearing it every eight to ten minutes for an hour straight.
This all builds to one of the biggest problems in digital advertising: people ignore advertising. People have never liked advertising. They mostly put up with it because it’s part of life, and many enjoyable things are subsidized by advertising to the point of being cheap or even free. Sports teams are sponsored, newspapers are incredibly cheap, and broadcast TV only requires an antenna. The Internet, however, is a different story. Digital advertising is crippled by what the industry calls “ad blindness.” Basically, people have come to hate digital advertising so much that they have trained themselves to completely ignore it. The traditional banner ad spots have steadily declined in effectiveness as people mentally tuned them out. One of the reasons that CPC and CPA are preferred over CPM by advertisers is because a click ought to demonstrate that a person saw the ad and clicked on it by choice, whereas CPM could be filling an ad space no one sees thanks to ad blindness. But ad blindness doesn’t end with passive disengagement. The more capable web browsers have third party extensions that you can install that block ads at the source: server-side at the ad network. These block not only banner ads, but also video ads on YouTube, Twitch, Blip, and so on. This is great in the short term for the visitor, they get a blissfully ad-free online experience, but for everyone else it’s potentially disastrous. The publishers don’t get paid because the ads never get served and the advertisers lose eyeballs as TV and traditional media die and are replaced by new media. Is the damage already done? Can digital advertising be salvaged?
The good news is that ad blockers are still only used by a small percentage of the population. The bad news is that that small percentage is made up mostly of young people. So while things are struggling now, it’s easy to see that as time passes the percentage of the population that uses ad blockers will grow. As the younger, tech savvy generation grows older, and the next generation grows up and starts using the web, ad blockers will become more popular. Assuming this is true, and the current methods for serving ads have a limited lifespan, efforts need to be made to change the paradigms under which advertising is handled in the digital space.
When it comes to developing new paradigms, I certainly don’t have all the answers, but we can look at some things advertisers and publishers are already trying. One of the approaches being tried, and a tactic I have written about before, is content marketing. As Howard Gossage, the godfather of western advertising, liked to say: “The real fact of the matter is that nobody reads ads. People read what interests them, and sometimes it’s an ad.” I don’t want to repeat myself about content marketing too much, so I’ll focus on a few examples of what people are trying.
Starting with the publisher end of things, BuzzFeed is a stellar example of using content marketing to create a more visitor friendly browsing experience. Instead of having obvious advertising all around their site, they place sponsored content in amongst the editorial content that is virtually indistinguishable from said editorial content. While on its face this may seem like another one of those cheap advertising tricks to get views, it actually works because the advertisements don’t feel like ads; they feel like regular content that just happens to be promoting a certain product. In the same way that subtle product placement doesn’t kill the flow of a TV show while offering a branding or sales message, sponsored content that blends seamlessly lessens the visitor’s grievance over the presence of ads by making the ads not feel like ads. On the ad network and ad agency end, a company like Sharethrough represents a solid effort to make this kind of sponsored content scalable. Sharethrough is built on the idea that people don’t hate ads, they just hate bad ones. People go out of their way to watch Super Bowl ads because they know that they are top quality. Think about that. Sharethrough’s mission, then, is to get this kind of quality advertising out there in an unobtrusive way and then track how many times the content is viewed, how often it is shared over social media, and how effective this content is at driving people to a company’s web presence. This creates a great ecosystem where publishers and advertisers can use the convenience of an ad network, while keeping the ad content visitor friendly.
Publishers could also take more control over the advertising process to make sure that viewers get the best possible experience on their site, while not losing income. If publishers took more control over their own ad units, and who is allowed to advertise there, it could help to shape up the digital landscape by keeping everyone honest. Let’s explore a few examples of how this works by looking at Penny Arcade, Team Liquid, and Reddit. Penny Arcade sells ad spaces themselves, instead of using an ad network, and rigorously vets their ads for content, quality, and the company being advertised. This lets them prevent all of the low rent ads that are often found on sites that target their demographic, mostly young male gamers, such as ads that use sexuality, or even the implicit offer of sex, to sell low-quality free-to-play games, dating sites, and various computer performance scams (downloads to increase CPU, protect against viruses, etc.). This dramatically improves the user experience, and engenders a lot of goodwill with their regular readership. That keeps them loving the brand and willing to support Penny Arcade’s other income sources, such as selling merchandise or the Penny Arcade Expos. Team Liquid, a premiere e-sports and community site for Starcraft and DOTA, uses an ad network, but underneath the ad units are images hardcoded into the site, with an image of a crying elephant that admonishes visitors to white list the site in their ad blockers. This image only appears if the ad that would normally be placed there by an ad network was blocked, server-side, by an ad blocker. The power here is that it puts a real face, albeit in a tongue in cheek way, on the fact that a community site like Team Liquid needs revenue to stay alive and that every blocked ad costs the site real money. Finally, Reddit places their own ads that are hardcoded into the site underneath the served ads. If the network serves up the ad, then that ad is placed and functionally overrides the Reddit ad. If the network ad is blocked, an ad for Reddit merchandise appears in its place.
Another thing that publishers have done is to seek out direct sponsorship. Rather than have a series of rotating ads that are served up by an ad network, they rely on a specific relationship with a company or companies who wish to support the content in exchange for constant acknowledgment by the content creators. Essentially, this is the web equivalent of the old “So and So Present…” radio and television shows, like GE’s General Electric Theater. This model has been picked up mostly by content producers in the audio and video realms. In the audio content world, podcasters faced the quirks of being listed on iTunes. iTunes, like YouTube for video, is too big to ignore if you are a serious podcaster. Unfortunately, iTunes has some rules about the podcasts it will carry. Amongst the most troublesome for content producers: they can’t charge anything, and the shows are downloaded as raw audio files. In general, when something is free you make up the difference with advertising support, but with only raw audio being downloaded you can’t attach variable ads, such as those from an ad network. The solution, then, is to put the ad in the audio file itself. Major podcasts like The Adam Carolla Show and Kevin Pollack’s Chat Show are both independently sponsored, using each of the segment breaks to put in a fifteen second plug for their sponsoring company and it’s products. YouTube producers have also experimented with similar systems. As I discussed in my post on monetizing new media, producers have built relationships with big companies to get their projects funded, like Phillip DeFranco getting Netflix to sponsor his movie club.
Sponsorship can be done as a hybrid approach as well. Remember how sites like Team Liquid and Reddit hardcoded images into their sites underneath the ad network? Some sponsored content producers are experimenting with that approach as well. Take livestreaming as an example, looking specifically at eSports broadcasting with Twitch. On Twitch a broadcaster can choose to run ads whenever they want in 30, 60 or 90 second blocks. But unlike with television, the underlying video never stops. Instead, the stream is muted and the ads are displayed in front of the broadcast like an overlay. The ads shown during this break are drawn from an ad server and therefore blocked by ad blockers. Major tournaments will run ads between games, and since there is no game footage to be displayed, they fill their stream with ads for their sponsor organizations. This way, a typical stream viewer will see the ads served by the ad network and a viewer using an ad blocker will see the ads for the sponsor organizations that are broadcast by the tournament in the underlying video stream. Since both the blocked and unblocked viewer are getting some form of brand message, the broadcasters are getting value out of every eyeball watching the event. This approach could be modified and applied to a variety of other publishing venues on the web, including blogs, community sites, online video series and more.
Whatever the solution or solutions, publishers and advertisers need to collectively work towards a mental shift in the perceived value of digital advertising across the board. Digital needs to shrug off its ugly past, embrace quality, and let the sites and networks that break the rules become blacklisted and ignored for the spam and scams that they are. When this happens, the per impression value of digital advertising should rise, the money leaving traditional media should bolster the new media alternatives, online publishers will become more fairly compensated, and everyone should prosper (even the users). Until then, I’ll leave you with a quote from Howard Gossage: “The buying of time or space is not the taking out of a hunting license on someone else’s private preserve but is the renting of a stage on which we may perform.”