Previously, we discussed how over-targeting potentially costs advertisers in terms of lost conversions and learning opportunities. We explored how concerns about over-targeting may be over-shadowing the potential data gold mine we get from testing the impact of Grouping, Referring and Taxonomy on conversions. This week we will look at the fourth consideration: Cost of Scaling.
What do advertisers gain from over-targeting? Best case is they hope to run an efficient campaign. But, those marketers may be confusing optimizing with scaling. Let's take a closer look at both: if you give me $1 and I return to you $1.05 you are making money on this exchange and you may seek to optimize the deal. Suppose, after optimizing, you give me $1 and now I give you $1.10. You keep optimizing and now for every $1 you give me I return $2. You have increased the efficiency by 100%! Impressive numbers until you realize that you only made one dollar for all of that work. Scaling is when you try to keep that same ratio but increase the original spend from $1 to $10 or to $100 or $1,000. In online marketing, usually the recipe for optimizing is not the same as for scaling. At some point, the marketer has to choose between optimization and scale. Is it worth more to you if I can back off the optimization from 100% to 90% but increase the scale by $100,000/month?
Simply doubling the ad spend may not double the conversion rate. Why not? Well, for one thing, at lower ad spends I can sell you the least expensive inventory which I have in great abundance. Or give away my best inventory at a ridiculously low rate just to win a new account. Both kinds of inventory are limited. If I am a remnant-inventory network, I buy inventory at various price points—some so cheap I can resell it on a CPC or CPA basis. But, as the advertiser increases the spend we run out of the least expensive inventory and now move into higher priced inventory which we can't scale at the same ratio…the "increased" inventory has 2 limitations –the CPM (media cost) is higher and the inventory is different. Different inventory converts at a different rate. Inventory which is a blend of various price points will tend to have more higher priced inventory baked into the blend as we attempt to scale. We're usually blending better performing inventory with worse-performing inventory and/or higher-lesser demanded inventory as well as higher and lower media costs. Branded publishers have the same kinds of issues. When the advertiser insists on over-targeting, that is, insisting on a specific combination of demo-targeting specificity with preferred placement on specific content, they are digging themselves into a hole. The more specific the targeting, the less inventory there will be. The targeting specificity should follow testing of grouping, referring data, taxonomy, optimization and scaling on each site or network—all of which collectively points the advertiser to the best blend of efficiency and scale of conversions; not the other way around. How do I know which specific target works best on each site and each network? Going to site or network to see how well they can reach an impression goal against a very specific target is missing the point of the game – make more money than you spend. Otherwise, the game is a beauty contest of who can reach more of a particular target regardless of the potential of increasing conversions by looking just outside that box.
You really want to help the advertiser (or help yourself you are a marketer)? Targeting should follow testing, which needs to be done on a site by site, network by network, spend limit by spend limit basis and not the other way around.The opinions and points of view expressed in this commentary are exclusively the views of the author and do not necessarily represent the views of MediaBizBloggers.com management or associated bloggers. MediaBizBloggers is an open thought leadership platform and readers may share their comments and opinions in response to all commentaries.