Online-to-Offline Attribution: Three Vital Questions to Ask

Traditional advertising has long been a core part of the brand advertising arsenal, because it has been perceived as reaching the largest audience. But the landscape of advertising has shifted heavily toward digital, with mobile’s rapid growth changing the status quo faster than any medium had before. 

To underscore this point, Digital Advertising increased by 19.1% in 2019 to $129.3 billion, while traditional advertising fell 19% to $109.5 billion, meaning that digital will account for 54.2% of total spending. While this shift continues to ramp up, the vast majority of purchases continue to be made at physical locations: an eMarketer study estimated that brick-and-mortar sales accounted for the majority of retail sales in Q4 2019 at 86.6%. 

With so much riding on the success of digital ad spend, and such little visibility into the sales happening at traditional brick-and-mortar locations, it’s more important than ever to understand the impact of advertising efforts in driving visits and measuring return on advertising spend (ROAS).

Read more: Beyond the Data Label: The Next Phase of Data Assessment

Savvy advertisers are now more focused on reaching the right set of eyes, versus spraying and praying, and as such they’re pushing the envelope of ROI measurement and spend accountability. That measurement takes a few forms, but whether it’s a foot traffic or sales lift study, the sources and methods used to measure are the most crucial part of understanding true behavioral changes and accurately assessing performance. 

With that in mind, here are three vital questions to ask when measuring the impact of Digital Advertising on offline behavior: 

Where Does the Data Come From?

At the heart of every online-to-offline attribution, the solution is location data. This is the lifeblood that enables measurement of real-world behavior, like store visitation, frequency, loyalty and dwell time. 

Accurately measuring a visit means understanding the sources, accuracy, and scale available to properly determine a visit. With so many vendors in the location data space, this can be confusing. But when you remove the branding and the marketing speak, there are three core sources of location data in the market right now: bid stream, beacon and software development kit (SDK). 

Bid stream data is known for being quite abundant, but low in quality, requiring scrubbing for random location placement (i.e. defaulting to the middle of the country). Beacon data is known for being very accurate, but challenging to acquire at scale, because it requires physical placement and battery/hardware maintenance. SDKs are integrated into mobile applications, 

which provides scale, while also tapping into GPS, WiFi, and Bluetooth for high-accuracy location determinations. For these reasons, SDK-based data is regarded as the most effective in attribution programs. 

It is critical that marketers first vet sources and understand the best uses and limitations of each, and have a clear understanding of where data comes from, before embarking on an attribution strategy. 

What Is the Methodology?

While sources and data quality represent the foundation, the methodology is where the rubber meets the road. Understanding the process and analysis methods employed to ensure statistically relevant results is yet another essential component of selecting the right offline attribution tool. 

A solid methodology includes a control vs. exposed approach and the process surrounding behavioral changes pre vs. post ad exposure. Marketers should also familiarize themselves with how statistical significance has been determined. Without a high degree of confidence in this area, it becomes indeterminable whether the observed behavior, be it visit lift or sales lift, was actually tied to the campaign in question. 

Independent, third-party reviews including audits and certifications can be an important and critical litmus test in this area. Typically, these auditors provide an independent examination of the procedures and approaches including methodologies, designs, and controls to assure the services committed to would achieve the results claimed by an advertising vendor. 

What KPIs Are You Measuring?

Incremental lift is the most critical KPI to consider when evaluating the success of a campaign. Lift is a measure of how well devices in a campaign performed when compared to a control group of devices, not in the campaign. 

A well-formed control group will help account for seasonal effects, different socioeconomic characteristics between groups, usage patterns, and more. The control group should be constructed by accounting for variables such as time of day, dwell time, proximity to stores being evaluated, and historical behaviors (among others) with the goal of designing a control group as similar to the exposed campaign devices as possible. 

While foot traffic lift is a common deliverable of most platforms, the rise of sales lifts through the inclusion of transactional data has introduced new metrics (i.e. sales lift), allowing advertisers to better understand the impact of campaign efforts on driving sales. With sales lift, advertisers are able to close the loop from impression to sale, providing a more comprehensive calculation around a campaign’s impact and ROAS. 

Optimizing for the Future 

While understanding campaign impact is essential, the actionable intelligence to optimize future campaigns can help marketers improve their return on Marketing spend moving forward. The rise of identity graphics combined with third-party data including demographics, psychographics, and transactional data provides advertisers the ability to dig deeper into not only who responders are, but what motivates them and what they are looking to buy in the future. 

When combined with Data Science and Machine Learning, these insights can be transformed into actionable intelligence allowing marketers to more efficiently remarket to individuals and motivate them with dynamic content that speaks more closely to what drives their decisions and preferences.

Understanding factors like dwell time, frequency of visit, brand preferences, competitive visitation, demographics, and psychographics can be used to create profiles or refined segments that, when combined with dynamic creative investments, will drive performance over time. 

Read more: What Matters With Intent Data