Measuring cross-device uplift during Black Friday 2015

Big retailers need to ensure that their marketing spend across a range of different channels (search, social, retargeting, etc) is accurately weighted throughout the year. This is especially true during key ecommerce periods such as Black Friday.

Online traffic was 60% higher on Black Friday this year than on normal days, so it was essential for brands to know where to spend their advertising budget to maximise revenue in this highly competitive marketplace.

But knowing exactly where to place that budget is very difficult, and can only be done with accurate cross-channel, cross-device attribution.

We work with leading cross-device solutions to allow us to measure the uplift in sales that different channels created. For example, we are able to discover whether seeing an advert on mobile social brought about a sale on desktop.

With a leading ecommerce client, we measured the difference across November when conversions were attributed to the last cross-device click, instead of just the last click. Last cross-device click refers to tracking a user’s last click across all devices, rather than presuming a user's clicks on a number of different devices are from different people.

Results

We found:

• Paid search and direct traffic (symbolised in the chart by the two thicker lines) are the leading channels for both 'last click' and 'last cross-device click conversions'. This (of course) increases a lot before, after, and during Black Friday Weekend.
• Interestingly, paid search is consistently undervalued when using traditional attribution (shown by the uplift when using cross-device measurement).
• Direct traffic is consistently overvalued (this makes sense, as with traditional measurement we think users are just visiting directly when in fact they may have received messaging on another device)

Cross-device measurement also reveals that the disparity between traditional attribution and cross-device attribution is exacerbated as we move towards high traffic, high marketing periods such as Black Friday Weekend. This year, Comparison Shopping Engines would have been undervalued when using traditional attribution, and Email and Organic Social would have been overvalued.

This makes sense - sales periods attract savvy shoppers that would visit and be influenced by price comparison sites, and maybe upper funnel activity (email, social) is more prone to getting lost in the noise when happening to users across many channels.

With the number of devices and sources of media now available,  advertisers need help to cut through the complexity and apportion budgets appropriately to drive incremental ROI.

Carter Crouch, Business Analytics Consultant, Infectious Media