A solid 67% of consumers conduct product research online before stepping foot into a store, according to research from Retail Dive.
So, as marketers tweak campaigns and reallocate spend across channels, it’s imperative that they take the full picture into account. Attributing every marketing dollar to online revenue while overlooking in-store sales will result in inaccurate reporting and uninformed ad spend.
To calculate your omnichannel ROAS, you should be thinking along the lines of (Online Revenue + In-Store Revenue) / Ad Spend
But, before you can get there, you need to accurately attribute your online revenue and in-store revenue back to specific campaigns. Conor Ryan, Co-Founder and SVP North America here at StitcherAds, has given this a lot of thought in regards to Facebook marketing. He conducted a significant amount of research last quarter to develop an omnichannel attribution model that retail marketers can start to employ.
Take a look.
Tracking your online revenue can be pretty straightforward, but to track in-store sales, you need to be tapping into Facebook Offline Conversions. This API will allow you to upload in-store transactions into Facebook and match them to consumers’ interactions with your ads.
To calculate the Online Revenue part of the equation, Conor suggests using your existing model.
Here’s an example:
1 Day Click + 50% 1 Day View
This means we’ll attribute all the revenue from people who clicked the ad and made an online purchase within 24 hours + half the revenue from people who viewed the ad and purchased within 24 hours.
Now, let’s move to the In-Store Revenue component. Conor created this model to attribute offline sales to Facebook ads:
1 Day Click + 50% 1 Day View + 50% Day 2-7 Click + 20% Day 2-7 View
Let’s break that down. Just like the online component, we’ll attribute all the revenue from people who clicked the ad and made an in-store purchase within 24 hours + half the revenue from people who viewed the ad and purchased within 24 hours.
It’s important to remember that not all customers are able to get to a brick-and-mortar store within 24 hours. To account for that, we’ll attribute 50% of in-store purchases over the next 2-7 days from people who clicked the ad and 20% of purchases from people who viewed the ad during that same timeframe. For example, if someone happened to see an ad on Tuesday for a $100 wallet and went in-store on Friday to buy it, we attribute $20 to the campaign and $80 to other media.
This attribution model would work well for retailers in certain spaces, e.g. apparel, beauty, sporting goods and discount stores. Retailers that sell items at higher price points (e.g. furniture stores or luxury goods) will need to factor in the customer consideration period into their offline attribution models. This might present a challenge considering Facebook’s attribution capabilities. The platform can take in-store data that’s up to 720 days old, but marketers can only attribute revenue up to 28 days after an impression or click occurred. (It’s also not possible to attribute revenue to activity that took place before an Offline Event Set was created.)
For retailers that sell products that range significantly in price within the same store (e.g. a $10 mug and a $2,000 espresso machine), Conor suggests setting up different offline attribution models. Depending on the price point and variation in consideration period, shoppers might shift how they buy with the same retailer.
Keep in mind that the modern shopping journey doesn’t just spring from online to offline – it spans devices and platforms too.
For a long time, the last-click attribution model served marketers well, but that’s not going to cut it in 2018. That brings us to multi-touch attribution, which can be a beast to figure out. But, because it gives credit to each marketing channel, it's one of the best ways to determine how to allocate future spend.
As the customer journey continues to evolve, marketers need to consistently look at their attribution models and update them to reflect consumer behavior. Without a willingness to adapt, there’s a high risk of losing the omnichannel game and missing out on significant revenue.