The challenge and promise of ‘digital everything’ attribution
Since the modern marketing age began, marketers have dreamed of being able to see exactly which campaigns resulted in sales, and how their various efforts influenced those purchase decisions. In this perfect world, they’d funnel their spend to high-performing channels based on their primary goal, whether that was sales, brand awareness or lead generation, and de-emphasize the rest.
Marketers are no strangers to attribution. The rise of digital marketing brought more visibility into campaign performance, but attribution it has fallen short of being able to truly close the loop. It remains ridiculously complex to stitch together all of the various interactions a person makes with a company’s campaign across multiple sites and ad platforms to correctly attribute a sale or conversion.
This complexity explodes when marketers attempt to credit their digital campaigns for increased foot traffic and ultimately, sales. Here’s why.
Sourcing the signal
Matching a digital campaign to an in-store purchase with any degree of certainty requires syncing point-of-sale data to a digital identifier, specifically a cookie or a mobile advertising identifier.
There is no one-to-one link at scale here, so many companies attempt to match customer records based upon hashed (anonymized) email addresses. With each attempt to match disparate data sources, the quality and scale of that panel decrease. Perhaps when paying with a mobile phone at checkout reaches widespread consumer adoption, this will change.
Today, most companies work to show online-to-offline attribution by measuring the change in foot traffic. This is done primarily through measurement of GPS lat/long data points and WiFi connections, with some data also coming from Bluetooth beacon interactions.
The challenge here is ensuring that your source data is both accurate and privacy compliant. 5G has been hyped to deliver greater accuracy, but only time will tell whether it will live up to its promise. The best providers in this space undergo large engineering efforts to screen out the low-quality data, whether it be low-accuracy, limited scale or incomplete. This leads to the second challenge.
Statistically significant panels
Acquiring high-quality data signals and matching those across devices and point-of-sale systems drastically reduce the size of the panel used to measure attribution. We’re talking tiny numbers here, and a marketer may only be able to measure single-digit percentages of their foot traffic to begin with.
To measure that audience requires an opted-in location sharing device, a highly accurate location signal (GPS data degrades indoors due to signal strength), accurate nationwide building footprints to confirm the visit and up-to-date business information to confirm the location is still in business.
Taking all of this into account shrinks that already small panel further. Attempting to match this panel to other customer records reduces it further still. Where we end up is with a small, but accurate, panel, showing some gain/loss in foot traffic, with extrapolations of inferences based upon that data to a nationwide scale. This isn’t true attribution; this is educated guesswork.
Weighting other marketing campaigns
Unfortunately for those seeking the attribution Holy Grail, marketers don’t deploy one digital campaign across one channel. They run campaigns across Google, social media, ad networks, print, TV and radio.
How does a marketer take this into account, capture and sync those touch points, and then assign the appropriate weighting to each? To say this is a difficult process is an understatement.
The path forward
Yet given all of these challenges, marketers rightly clamor for more attribution solutions and more performance-oriented campaigns. The sales teams pitching agencies and brands would love to provide an all-encompassing attribution solution so marketers could know where to invest additional resources and what needs to improve when things don’t go as planned.
The foundation for meaningful online-to-offline attribution solutions will be accurate, clean and privacy-compliant location data.
Building upon this, the technology will advance, the silos of data will merge, and the data will grow in scale and quantity.
Eventually, as campaigns run, marketers will be able to spot the patterns, follow their hunches, in addition to their data, and adjust accordingly. We’re just not quite there yet.
This post originally ran on MarTech Today.
To learn more about how Reveal Mobile approaches location-based audiences and analytics, click the button below.