Yair Halevi Wed Sep 19, 2012

A majority of marketers still use last-click attribution to value performance-oriented campaigns. Yet, at the same time, marketers realize it is ineffective at gauging the true influence campaigns have on the consumer’s path to conversion.

Is this you? Is this your performance marketing counterpart? For performance marketers who solely rely on last-click attribution to effectively measure online advertising campaigns, we’ve got four metaphors for you…

Last Click Attribution Infographic

Let me break it down for you… 

Last-Click Attribution Is Wrong#1. A tiny mouse riding a massive elephant, stating, “Wow, look how much dust WE are creating!”

Last-click (and first-click) model completely under-value marketing efforts influencing the middle of the sales funnel (e.g., research and consideration), just as this mouse is under-valuing the elephant’s size and strength compared to his own.

Research implies measurement models that assume a single touchpoint influence are not correct at estimating value, and sorry Mr. Mouse, neither are you.
Last-Click Undervalues Infographic
#2. Choosing a product according to its packaging, rather than the product quality.
Choosing a product based on its packaging and price is akin to measuring a campaign by a single touchpoint. Single-click attribution does not accurately capture the real influence a campaign has on a customer conversion (e.g., a purchase), just as a product’s packaging does not accurately reflect the true quality and benefits of the product.

Last Click Infographic Online Video Campaign

#3. Building a house and only paying the painter, because he was the last person to work on it.
If only this was a real-life scenario! What an incredible investment: paying a $2,000 paint job for a 2,500 square-foot house with a $350,000 price tag?

But this would never happen, so why is it happening with online marketing measurement models? As customers journey down the sales funnel, they engage with numerous advertisements and marketing messages. If these multiple touchpoints are not all measured, then marketers are not receiving an accurate ROI of their online ad campaigns.
Last Click Infographics

#4. A graduating senior thanking her last teacher for her entire high school education and success.
Just like the above metaphors, this graduating student is under-valuing her 4-year academic experience by acknowledging one teacher. Need I say more?

The message of these metaphors is all the same: any click-based attribution model does not accurately capture the real influence of an online ad campaign, and therefore, under-values brand impact. Especially with SmartVideo advertising, a more personal and intimate medium that captures user attention and engagement for longer periods, video has more potential to influence a buy decision at distance, with or without a click.

Ok, I get it. But I still use marketing attribution. Now what? 

We understand that marketers employ multiple marketing tactics to drive traffic and acquisition, and that all tactics contribute to overall performance.SundaySky is a firm believer that there is one source for each order converted, and we do support the notion of having attribution to every marketing tactic. A single conversion should only be counted once and attribution needs to add up to 100%. We also recognize omni-channel strategies, meaning tactics contribute to online and offline impact.

We recommend using control-group analysis together with attribution modeling. Attribution models are still important for analyzing behavioral segments, optimizing the campaign via creative selection and smart targeting, and understanding cross-campaign interactions using control-group testing. However, using control-group testing can take much of the guesswork out of attribution modeling. If you employ both tools, you can use the results of control-group analysis to tune and validate your attribution models.

What do you think? Do you agree with our above examples? We’d love to hear from performance and brand marketers alike on this topic.


Leave a Reply

*
*