The following was previously published in an earlier edition of Marketing Insider.
Our once-stalwart marketing model is breaking. To understand why, look no further than
Facebook’s expected $10 billion ad sale revenue hit from Apple’s changes to give users more control of their privacy.
It’s not just Apple whose privacy changes have caused
advertisers to halt marketing activity on iPhones and iPads. User backlash over privacy has pushed Google to commit to phasing out tracking cookies.
These changes make it ever harder for
marketers to answer the fundamental question: What value am I getting for my spend? As a possible recession and interest rates over 7% loom, the pressure to demonstrate profitability and value is
intense.
The solution is staring marketers in the face.
Incrementality testing, underutilized for years, is a great way to gain reliable information on value -- without privacy
infringements. It enables marketers to segment their channels to get a granular view of advertising effectiveness and make spending decisions accordingly.
When I worked at Lyft, our
incrementality testing revealed that competitors were overlooking the high-value strategy of offering free airport WiFi to customers who downloaded the app. The testing also revealed that Instagram
was a highly effective platform on which to run ads. Attribution models would’ve missed this insight, because most users were downloading the Lyft app separately, rather than clicking through to
it on Instagram.
So why haven't more marketers embraced incrementality testing and reduced their dependence on attribution models? Three reasons:
First, it requires deep expertise in
data science and analytics, something that few marketing leaders have.
Second, platforms like Google have little incentive to provide tools that give advertisers a better way to measure
impact. They know most customers are overvaluing their marketing returns and would cut spending if they knew the truth.
Third, incrementality testing is a hard sell for providers, compared to
the traditional attribution model. An attribution product is a plug-and-play tool that promises to give CMOs quick results. By contrast, setting up incrementality testing requires custom work and data
analysis capability that might take six to nine months to bear fruit.
To overcome these challenges, you need to ensure the CFO and CEO have bought in to the idea of incrementality
testing. If they don’t understand the value and aren’t prepared to fund the necessary investment, your proposal will be dead on arrival.
It’s also important to know how
you want to target testing to get the most value. In general, the biggest potential gains come from examining the strategies that account for the most spending.
Strong data science gives
incrementality testing its magic, requiring firms to invest in building a data analytics team or hiring a consultant with analytical modeling experience.
Companies that do this best will test
regularly -- at least quarterly -- to ensure they’re improving on their measurement process.
Using incrementality testing to assess the efficiency of marketing dollars should be top of
mind at this moment. I’ve seen a lot of buried bodies out there. One CMO I know was fired after the CEO found they’d tried to hide underperformance. If you don’t find the problems,
someone else will.