The Marketing Technologist.

We talk about analytics, code, data science and everything related to marketing technology. Backed by the tech geeks of Greenhouse.

Research: Mobile advertising increases both online and offline sales

Finally, there’s proof for what we’ve known all along: mobile advertising has a positive effect on sales through other channels, such as other devices or an offline shop. Finally, because anno 2016, mobile advertising is still struggling to claim its rightful place in the marketing world. Sure, budgets for mobile ads are on the rise, but if we need a budget cut, mobile ads are the first to go.

This makes perfect sense, of course. A result hardly counts unless you can measure it, and the ROI of a mobile ad isn’t easy to calculate. Many marketers prefer to take the easy road, and choose the method with the most obvious result. Now you might wonder why the effect of mobile bannering hasn’t been researched before. After all, in this era of the smartphone, it’s unthinkable that mobile ads don’t have an impact. The biggest hurdle? Data. To research this impact, you need to target a large, diverse group of mobile users, and find a way to compare the online and offline sales results of this campaign with a control group. That may sound like a piece of cake, but there’s a reason we’re the first in the world to research the impact of mobile ads this way.

Golden opportunity with the Staatsloterij

The perfect partner for this experiment wasn’t easy to find. Luckily, one of our clients was just as curious about the ROI of mobile advertising as we are: the national lottery (Staatsloterij). A golden opportunity, because Staatsloterij has a very heterogeneous client base. Young, old, unemployed, highly educated, indigenous people, immigrants and everything in between. People from all segments of the population are represented. Best of all, these clients are distributed somewhat evenly across the country. Staatsloterij was very eager to help us with our experiment and provided us with 1.5 years of historical sales data.

The research approach

To test the impact of mobile ads, we selected 3 Dutch provinces: Drenthe, Overijssel and Zuid-Holland. Using IP-masts, we targeted people in just these provinces and served them a total of 3.5 million ad impressions on their mobile phones. Meanwhile, we paused all mobile ads in the rest of the country: our control group. Through the so-called fixed effects model we neutralised constants. I won’t go down the statistics rabbit hole, but this basically means we filtered the effect of other factors (besides the mobile ads) to achieve a clean result. This is where those 1.5 years of historical data came in handy. Apart from that, we carefully selected our test and control groups to diminish result-altering differences in areas such as age, education, or employment.

On to the results

Hopefully, at this point you’re still curious about the results of our experiment. And those were pretty outstanding. The offline sales increased with 2.8% on average! A great result! What's also interesting is we discovered that offline sales don’t have a so-called cannibalistic effect on online sales. On average, the amount of online sales through mobile and non-mobile channels increased with 7%! Our conclusion? The ROI of mobile ads is considerably higher than previously calculated. This definitely calls for further research, for instance with other clients and in different markets. That’s why we hope that others follow our lead. Of course, we also hope that mobile ads will finally get the attention (and budgets) they deserve.