We're in a market crunch right now, and to many people that means they should focus on direct response (DR) media sales. It makes a lot of sense in some ways, but I wanted to take a few minutes to examine this.
In online advertising, there has always been an inherent tension between sales of advertising to DR and brand advertisers. Sales strategies at major publishers are in constant flux depending on the economy, and methods of approaching brand and DR buyers vary widely. But the reality is that nearly all advertisers (that spend any significant amount of money) are both brand and DR advertisers. While some advertisers lean in one direction or another, in reality most large advertisers spend their budget in both ways, using different agencies (or even internal departments) to manage different parts of their budgets. And while it's easier to justify DR budgets in a down economy, we should all be clear about what we're doing, why we're doing it -- and what the long-term impacts could be.
The problem with DR buying is in getting statistically viable data with which you can do an ROI analysis. This is, of course, what DR buying is all about -- making sure that every dollar spent has more than a dollar in return on investment (generally within the life of the campaign). In the case of inventory that has very high response rates, the math is quite easy to do. If your conversion rate is above 1 percent (meaning that 1 percent of impressions lead to a conversion), then it's an easy thing to calculate your ROI. But when your conversion rate is .0001 percent, then a vast number of impressions are needed in order to do any kind of analysis -- and even more impressions are needed to do a bit of optimization.
DR buying at low response rates still works, but the cost of the inventory needs to be very low. So catering to DR advertisers on inventory with high conversion rates -- like search -- is a no-brainer. But catering to DR advertisers on display advertising that has any value whatsoever to a brand advertiser can cause big problems -- significantly undermining viable pricing that any brand advertiser will pay.
We really need to make educated decisions about how much of our inventory to offer up at low cost to DR buyers, and how much we will hold the line on our pricing. There needs to be an informed decision made about the balance of ads -- essentially what I call "The Advertising Experience" to which we subject users. Most DR ad creative tends to be -- how shall we say this politely? -- lower quality than brand ads from the perspective of advertising experience.
This has its roots in the fact that DR campaigns typically only measure results tied to the life of the campaign, meaning that their KPIs don't capture negative brand association. There are only so many dancing maids or sexy personals ads I want to see in a given day. There's a tipping point at which I (as a user) don't want to see bad ads any more -- and simply want you to stop showing them to me. But all too often, I'm seeing staggeringly bad ads all day long.
Brand advertisers are not stupid. They visit sites that their ads are running on, and they hit refresh. They are not happy when they see the dancing maids interleaved with their very nice creative. So what does a salesperson say when an advertiser complains about paying big bucks for inventory that has very poor quality DR ads running on the same placements?
Let's talk about television for a minute. Even in local TV -- where a vast amount of inventory goes unsold -- the TV programmers have done a great job of limiting the amount of inventory that is available for DR buys. Typically they'll limit this to 10 percent (or some fixed amount) of total inventory sold. Unsold inventory is used to run promotions for their own station or network shows, or public service announcements. Limiting the amount of remnant inventory that they sell has had an interesting impact on average CPMs on TV. Programmers keep the low quality impressions seen by users tied to low value times of day and lower value content. If they didn't, we'd be seeing infomercials and other DR campaigns all the time -- and this could have a negative impact on our perception of the TV network.
Unlike online display ads, where we have mastered the sale of remnant inventory to resellers like ad networks, in TV unsold inventory goes unsold. This has kept the average CPM on TV inventory relatively static between $5-9 (depending on who you believe) for a very long time. TV programmers have created some stability around the advertising experience while keeping the supply and demand in some kind of equilibrium.
But with online display ads, we have dynamic ad serving (every time a page is viewed, a different ad can be seen), and every experience of a page has opportunity for a terrible creative to be shown. Additionally, we have no controls in place to protect the average CPMs of ads -- we've seen downward pressure on pricing for years because there is such a huge supply of inventory available at dirt cheap prices through remnant ad networks. While advertisers can't pick and choose where to run ads on this inventory, for many brand buyers it simply is too good a deal to pass up (especially when their budgets are crunched).
Quality display inventory has a certain intrinsic value due to its ability to broadly influence the mind of the population -- the collective consciousness. We have to be very careful about bowing to pressure and letting high-value inventory be negatively affected by the down economy.
Eric Picard is the advertising technology advisor to the Advertising Platform Engineering team at Microsoft.
