Conversations with online ad salesmen can take predictable paths. I’ve always worked in an environment where we measure marketing by how much it costs to get a new sale or subscriber, so I evaluate marketing spend on a cost per acquisition basis, and would prefer to purchase ads on that same basis. Ad salesmen usually want to sell ads on a cost per click, or even a cost per display basis. I tell them a “display” doesn’t mean anything. Just because the server sent the ad doesn’t mean the browser displayed it or the visitor saw it. <\/p>\n
When it is possible to talk an ad salesmen into a cost per conversion campaign, they want to count “view through conversions.” That means the web visitor saw the ad, then converted later. This makes sense because hardly anybody clicks on ads these days, so “cost per click” underestimates the effect of the ad. We all know there is some effect<\/i> from just seeing an ad — otherwise nobody would invest in road-side signs, or print magazine ads. <\/p>\n
Also, it’s not really a “view through” conversion but a “display through” conversion. The ad salesman can’t promise that the visitor saw the ad<\/i>. He can’t even promise that the ad showed up on a part of the page that was viewable in his browser. (More on that below.) <\/p>\n
Counting a “view through” is fair, so that you don’t underestimate the effect of the ad, but depending on how you do it you can easily overestimate the effect of the ad. You do this, first, by counting “views” the user never saw, but second, by the length of time you allow between the metric you’re using (display, view, click) and the conversion event. If somebody sees an ad and buys 5 seconds later, you’re pretty safe attributing that ad to that sale. But what if the visitor saw the ad four days ago, or a month ago? The cause and effect chain gets a little shakier. <\/p>\n
If you’re buying ads on a “view through” basis you need to watch that time lag very carefully. The ad salesman will try to stretch it out. Keep an eye on what standard they’re using to “count” a conversion, and make sure you have a way of seeing the gap between the event and the conversion. <\/p>\n
One part of the tangled mess of online ads is getting a little less tangled. New standards on viewability<\/a> require people who serve ads to distinguish between the ones that are viewable and the ones that are not. (For example, an ad that is “displayed” down the page where the visitor never scrolled.) There are still some details to work out, but it’s a good development, and it makes the math of ad buying a little more direct. <\/p>\n Marketers would like to be able to track sales to a specific cost so they can calculate the ROI, but as I mentioned in the discussion above, it’s never as simple as that. There’s always a certain amount of guessing and fudging involved. Allowing ad purchasers to only pay for ads that are “viewed” (i.e., on the viewable area of the screen) is a step in the right direction. <\/p>\n I’ve covered this before and I don’t mean to harp on it, but anyone who says “print is dead” should be politely escorted to a nearby sink and invited to soak his head until the fever passes. <\/p>\nDigital magazine launches print product to expand its reach<\/h2>\n