In today’s complicated digital advertising and advertising-related technology landscape where the focus has shifted to audience-based buying, there is limited attention paid to the quality of the advertising environment. One of my colleagues has suggested previously that this is in part due to the overreliance on machine learning and algorithms in the programmatic purchase of inventory. “The main reason DSPs are not good at valuing the quality of inventory is that they over rely on the algorithm to make effective decisions,” according to Nate Woodman. I agree with Nate on this point but also believe that perspective is at issue.
If you are reading this, my assumption is that you are familiar with the mechanics of programmatic display inventory buying via DSP. You are well aware that most DSPs that employ machine learning algorithms to “explore” and “discover” inventory use a pixel fire as an indicator of success to tell the machine that the inventory acquired has been “successful.” Clearly, in an over-supplied environment like display, we need the use of technology to find valuable inventory and to create efficiency in the valuation and purchase processes. Some clients can actually derive real value from inventory purchased this way and would likely deem inventory that fires a conversion pixel to be of high quality. I would argue that this has a higher likelihood of being accurate for clients whose primary sales channel is online.
To simplify what I have just written, if you are using DSPs to exploit your online commerce channel and measuring success on a last-action basis, letting a DSP run “unrestrained” to maximize conversion volume may be a successful formula. Unfortunately, this formula and definition of quality starts to lose esteem if you are trying to do much more than that online.
Let’s start by considering a logical notion: any brand that is trying to influence a consumer via advertising (of any marketing-related concept, be it awareness, consideration, purchase, or loyalty) wants to ensure that its advertising is at least seen (or heard) by the consumer it is trying to influence. Sadly, in today’s exchange environment, there is a strong likelihood that this does not happen. Why? Because most likely the inventory they are purchasing is below the fold, and often very far below the fold. Bizarrely, it is also likely that they are paying a premium for the ad that is less likely to be seen, because a machine may be placing a value on the fact that by the very act of the page rendering and a pixel firing, it has found “quality” inventory regardless of whether the ad has been seen by the intended recipient. The premium I have suggested comes from the fact that many “machines” may be seeking that same last action pixel creating competition for below-the-fold inventory.
Let me offer another likely issue with a “pixel only”-based quality perspective: any brand that is trying to influence a consumer via advertising is likely to be more successful in uncluttered and uncompetitive environments. Put another way, the lower the quantity of competing ads on a page, the greater the likelihood that the advertising will influence the user. Unfortunately, the average DSP fails to take environmental quality (such as the number of ads on a page) into account, and fails to give an advertiser the maximum opportunity to influence the client.
Advertisers need to define the permissible quality of the advertising environments on which they choose to promote their brands. To reach this definition, advertisers need to fully understand the effect that their measurement methodology has on the technology they are using to value, purchase, and deliver advertising inventory. It has been my experience when speaking to marketers on this topic – especially high-ranking marketers – that most agree that environmental and structural quality are important. Yet, when it comes time to transact, they push back against changing the way they measure display advertising and revert to some form of last action methodology. Matters only get worse when it comes time to optimize inventory planning and purchasing that should be based on environmental quality but measured with some illusory methodology of attribution. Nothing in our ecosystem improves if clients don’t start to better understand measurement and quality.
In a marketplace where evaluating billions of impressions daily requires the use of technology, a simple tenet should govern the purchase of inventory; “you get what you pay for.” Clients should place a premium on providers that can better illustrate the actual quality of the advertising environment on multiple dimensions. Advertisers should start the evaluation of any plan, provider, or publisher by asking the simple question, “Does my advertising have a chance to influence my intended target?” If you don’t know where the ad is likely to be on the page or how competitive the environment is, how can you know the quality or the value?
Header bidding is a programmatic technique that allows publishers to offer their inventory through multiple ad exchanges before they serve up ads from their ad server.
As Facebook keeps changing its news feed algorithm, one constant factor is the domination of video content and so brands keep experimenting with ... read more
Advertisers could be doing more to understand, measure and evaluate the effectiveness of their display advertising campaigns. In this whitepaper, Quantcast explains how in four easy steps.