For businesses whose primary or sole sales channel are online, tracking the effectiveness of online advertising is paramount to their success. With a clearly defined focus, connecting the dots between online advertising and sales is business as usual for these companies. For the businesses that rely on offline sales channels, however, tracking the ROI of online advertising can be incredibly complex.
With every complex problem, however, comes enormous opportunity. Be one of the first, or the best, at proving digital ROI for brick and mortar businesses, and suddenly you can take a bite out of the Google-Facebook duopoly.
With Snapchat’s recent acquisition of Placed – an ad tech company that specializes in measuring offline sales attribution – for a reported $125 million there are clearly significant efforts being made to bridge the gap between the digital world and physical sales.
Developing an attribution model
For many offline businesses, the first step on the path towards proving digital ROI is to develop an attribution model. As the name suggests, an attribution model allows a business to track sales or conversions across multiple touchpoints. For example, when looking at the ways in which consumers decide to go see a specific movie, we can easily identify several valuable actions by the consumer:
- Searching on Google for the trailer for a movie
- Clicking on a banner advertising a movie
- Watching the trailer on YouTube
- Visiting the website for the movie
- Liking the movie on Facebook
- Looking up show times on Fandango or Movietickets.com
We can assess these interactions through a number of primary models:
‘Last interaction’ model – as the name suggests, the last interaction is given the credit for the sale.
‘Time decay’ model – the actions performed closest to the sale take the majority of the credit for the sale
‘Linear’ model – all actions along the path to purchase take equal credit
‘Position based’ model – the first and last actions take the majority of the credit, and the intermediary steps share the remainder
Clearly all of these actions are desirable for a modern movie studio, but not all actions are created equal. And that is where statisticians can play a huge role. While it’s relatively easy to offer an opinion on which actions are the most valuable (i.e., looking up a show time is a clear signal of intent, in my opinion), nothing beats developing an attribution model that produces reliable, replicable outcomes across different variables, in this case movie titles.
But even armed with a custom attribution model, we’re still missing the final link to the physical purchase. Frequently, consumers will engage with a movie title online, view a trailer, look up a show time, and then simply walk up to the box office on a Friday night and pay with their credit card. Of course, this doesn’t mean we should throw out our attribution model. It provides us with a valuable tool through which we can assign value to actions taken online. With this in mind, we need to look towards the coming solutions that will help us bridge the digital-physical divide.
There are a number of ways in which this problem is being addressed:
Ad-tech companies such as Placed are using location signals from mobile devices to identify consumers and then compare groups of consumers exposed to online advertising against control groups who have not been exposed to the brand’s advertising.
For example, in our movie scenario, one group is exposed to online advertising for the movie, while the control group is not. On the opening weekend, Placed is able to compare the two groups and identify if there is a lift in movie attendance on opening weekend that can reasonably be correlated to the online advertising exposure. Similarly, Snapchat’s ‘Snap to store’ geofilters allow them to attribute increased store visits back to advertising run by brands on Snapchat.
Brands with strong loyalty programs are having a slightly easier time of bridging the physical-digital divide. Take the Balance Rewards program at Walgreens for example. With more than 100 million members – many of whom have both phone numbers and email addresses linked to their accounts – it’s relatively easy for Walgreens to target these consumers with online ads.
These ads can be targeted based on a customers past purchase history because, of course, Walgreens has that information. And, following exposure to an ad, guess what unique identifier customers enter during checkout? Their phone number, which is directly linked to their Balance Rewards account. In the case of Walgreens, it’s easier for them to bridge the digital and physical world. Sadly, loyalty programs can’t help our movie studio friends because very few of us select a movie based on the studio that produced it.
Emerging technical solutions
As we head towards the end of 2017, our devices are becoming increasingly connected to the physical world around us. Many of us now pay by merely tapping our phones or watches at checkout.
With the release of iOS 11, Apple is finally opening up NFC and allowing app developers to add NFC capability to their applications – for example, tapping your phone to gain entry to a concert, or a movie. And even the humble QR code has the potential to make a return, with QR code reading being built right into the iPhone camera application. While none of these developments are on their own the silver bullet to connect the digital and physical worlds, they represent significant steps along that path.