Transparency in Performance Marketing – Data Series Part 4 of 4

Desculpe-nos, mas este texto está apenas disponível em Inglês Americano. For the sake of viewer convenience, the content is shown below in the alternative language. You may click the link to switch the active language.

jeremy_photoBy Jeremy Verbit, Senior Product Manager, HookLogic

Advertisers know performance data is powerful.  It is used to construct marketing strategies, optimize marketing campaigns, and justify marketing budgets.  It’s the only way you can know if you are reaching the right customer with the right message, in the right place at the right time in order to drive sales.  But connecting your ad to a sale isn’t straightforward.  Figuring out how much a sale can be attributed to a particular ad introduces art to the science of performance marketing.  We will take a look at some of those attribution methods.

Attribution Window

The most basic factor is the attribution window, which is the duration of time between a sale and the moment its purchaser was exposed to the ad being assessed for attribution.  It’s most commonly used as a maximum length of time for which an ad is considered to impact a consumer’s decision to buy.  Any ad that was shown within the attribution window is considered valid for attribution.  While a thirty-day attribution window is most common in the ad tech industry, setting an appropriate attribution window is industry dependent.  For example, HookLogic Exchange data shows that on average for a consumer electronics purchase, 19 different products are considered over 10 days, while for a packaged goods product it is 8 different products over 6 days.  In general, the more expensive the purchase, the more comparison shopping and the longer the consideration time for the consumer.

SKU Relation

Another factor for sponsored product listings is SKU relation.  If a SKU that is purchased is very similar to a SKU that was previously clicked on in an ad, it’s likely the consumer found the SKU they purchased due to your ad, and thus your ad deserves credit.  For example, I’m upgrading all my lightbulbs to bright, energy efficient LED bulbs.  I see a vast number of SKUs in this booming category and I click on an ad for a Philips bulb.  I did not purchase the advertised bulb, but I did purchase a different bright, energy efficient LED Phillips bulb.  Phillips, as the advertiser, would likely want that ad to take credit for the purchase even though I bought a different bulb. They benefitted from the product relationship. (I should note that some advertisers prefer a strict same SKU approach to attribution.)

Clicks vs. View-Through

One factor that’s been the topic of much study since the early 2000s, is the validity of view-through attribution.  The question the industry has been asking is “when a consumer views a SKU in an ad, does not click on it, then purchases that SKU within a given attribution window, is that a significant enough indicator that the ad persuaded the consumer to convert?”  Way back in 2004, DoubleClick, agency Arc Worldwide and Continental Airlines did a study that used test vs. control methodology on a campaign to determine if there indeed was a correlation.  They found that 67.5% of impressions were directly attributable.   Subsequent A/B tests in the industry indicate that indeed a significant percentage of the time ads that are viewed but not clicked do contribute to conversion. The impact varies based on the ad creative and industry factors.  Many advertisers prefer to have their performance indicators include the impact of view-through rather than leave the incremental lift uncounted.  Others prefer the historical click-only approach to attribution.

HookLogic Enables Your Custom View of Performance

Ultimately, performance measurement is a function of your business and your goals.  At HookLogic, we provide an Attributed Transaction Log report that shows each individual attributable transaction, its relation to the advertised SKU, and its advertised SKU engagement as either click or view.  This not only allows you the option to filter for your own custom view of performance, but also allows you to see each individual transaction that made up that performance so that you know what is contributing to your performance metrics.

This brings our Data Blog Series to a close.  We have reviewed click fraud, the estimated $7B it will cost advertisers in 2016, and how we use automated methods to detect and filter it from billing and reports in under 10 seconds.  We have discussed ad viewability (how more than half of ads are never viewed), and how we ask our publishers to adhere to the Media Rating Council’s viewability standards.  We have also reviewed transparency in performance marketing and how we break out attribution window, SKU relation, and click vs. view-through attribution at the individual transaction level for your custom view of performance.  I hope you found this series valuable.  Until next time. I welcome your comments and questions. Please share this on social if you liked it and feel free to reach out:

Tags: Attribution, Data Science, ecommerce, research, Tips, UK, US,