Measuring the success of online display advertising is a well-known industry challenge. The industry tends to focus on measuring direct response online ads by conversion metrics, but has failed to come up with a consistent means to measure the effectiveness of brand marketing. Writing exclusively for ExchangeWire, Martin Pavey (pictured below), UK country director, Flashtalking, outlines the steps required to create a cleaner dataset for attribution models.
When putting together your online marketing strategy, I’d be stating the obvious if I reminded you to ensure your strategy is different for upper-funnel versus lower-funnel users. But you may not think I am stating the obvious if I reminded you to use a different means of measurement for your upper-funnel versus lower-funnel ad campaigns.
Different display ad techniques are used at each stage of the purchase funnel. Some are used for brand marketing, which intend to increase awareness and consideration, others are used for direct response to drive the consumer through to conversion. These ads cannot be measured in the same way.
Banners ads don’t work, according to Clare O’Brien, senior industry programs manager at the IAB. Her reason being that: “Banners have to be served 1,250 times before someone clicks on one.” I beg to differ. Banner ads do work; but what Clare has failed to notice is that banner ads are being measured wrongly, making them appear not to work. The creative executions we see these days also needs some love, but that’s for a different day.
Martin Pavey, UK Country Director, Flashtalking
Online display ads are being measured by number of clicks – but as Clare quite rightly pointed out, an extremely small percentage of consumers actually click on brand marketing display ads. Why do the number of clicks matter? If these ads are designed to drive awareness, then surely their success should be measured on the amount of awareness they are driving and not by the amount of clicks they’ve garnered. If we are wrongly measuring the success of our ads, then we don’t have the data sets needed to perform attribution for optimising digital marketing spend.
"But how on earth do you measure awareness?", I hear you asking.
The ad tech ecosystem is made up of a number of specialist agencies who are working hard to answer this question and who can tell you which ads are actually reaching our devices and audiences. There are a number of factors to consider when measuring awareness: is the ad actually viewable? Is the ad viewable for long enough for the user to actually notice it? Has it got the attention of the consumer? How much attention has the consumer given it? Is the consumer who viewed the ad on one device the same as the person you think it is who is converting on another device?
Cookies greatly distort our view of reach and frequency against devices, let alone against people. Log files include fraudulent impressions and clicks; and ads that never had a chance to be viewed are often included in the data. All these industry challenges lead to inaccurate datasets; and you simply cannot build decent attribution models from bad data.
In order to create a cleaner dataset for attribution, the following steps should be considered:
- Integrate a cookieless tracking solution to capture the full path of impressions, clicks, visits, and conversions for each user.
- Incorporate device graph data to capture consumers’ experience and actions across their various devices.
- Integrate user-level verification data to remove fraudulent impressions and impressions that were not viewed.
- Remove impressions and clicks that are delivered after the conversion event, and those that represent unreasonable frequency against individuals.
- Unify and join the data for each user to assemble the timestamped history of engagement for each converter and non-converter.
- Split the log file for display into two sets of data: one for upper funnel and one for lower funnel.
If you take these steps, your data will be a dramatically better representation of the marketing that consumers are actually experiencing, as opposed to what they should be experiencing. That better representation will generate more accurate and actionable models for optimising media investments across platforms, channels, and media partners.