At Madhive, we work every day with clients to find their prized, needle-in-a-haystack audiences, and we know that it can seem like a daunting task — especially for marketers that are new to CTV advertising.
So we wanted to put our tech to the test. We launched our first CTV campaign for B2B audiences (which are pretty difficult needles to find in a haystack!) so we could make what we do and what clients can learn more transparent.
Our campaign has three phases:
From the get go, it was clear that to reach our audience on CTV, we needed to cast a wide net. While we have run hundreds of thousands of campaigns for clients, this is a first for our Madhive business, after all! Audience data was essential to shaping success for our campaign.
We’re now deep into the test, and we’re ready to share what we’ve learned so far.
In the world of B2B advertising, your target audience tends to be extremely specific. We’re offering specialized services, so we wanted to target people who could benefit from CTV advertising solutions and have the power to influence within their business.
We knew that precision would be key, but we also knew we had very little data from which to start. So we built a broad audience strategy to make sure we were reaching viewers that were relevant to our message.
Naturally, we started by looking at our own first-party data. We took a list of known contacts, anonymized them, and matched them to IP addresses so we could deliver our ad to these households. Of course, this is all conducted in a privacy-safe manner.
That was a great start, but we didn’t have the scale we wanted for this higher-intent audience bucket. So we decided to seed the campaign on Linkedin before the CTV ad went live, to build awareness and increase traffic to Madhive.com.
People that engaged with these posts were then added to our 1P bucket, and we retargeted them with the CTV ad. This turned out to be one of the best planning decisions we made.
Once we had our 1P audiences planned, it was time to add a bit of spice to the mix with third-party data. (Thanks, Lou, for the spice rack analogy.)
Using what we know about our current clients and CTV decision makers, we aimed to extend our reach to new audiences. To do that, we relied on third-party data partners.
So how granular did we get? To start, we focused on job titles:
We also segmented by industry and job function:
Finally, we layered in income targets as well, to ensure we were more likely to hit decision-making audiences.
Regardless of what channel you use, most B2B customer journeys require multiple touches before conversion — far more than your typical B2C ad. Knowing that, we had two goals in mind for our campaign:
For the purposes of this experiment, we defined conversion as a visit to our website after viewing our CTV ad, and we measured reach through impressions.
Not surprisingly, our 1P bucket performed the best for conversion, with a CVR that was 300x higher than third-party targeting. A whopping 62% of those site visits, however, came specifically from our retargeting audience.
So what did we learn? While CTV is certainly effective for brand awareness and reaching new audiences, it’s also a tool for lower-funnel strategy when used with the correct targeting.
It’s also worth noting that we saw a healthy number of conversions from our AI audience modeling, driving 13% of our site traffic. We’ll dive deeper into how we’re using AI in a later blog post.
We were able to deliver our message to a staggering 870,000 people – and counting – by leveraging the megaphone that is 3P data.
A whopping 86% of impressions were delivered to 3P audiences, with the Sales & Marketing segment leading the pack.
*note that numbers will exceed 100, as some IPs may be in more than one 3P bucket*
Now that we’ve learned from a broad data strategy, we’re moving into the hyper-targeting phase. We’re optimizing for what has worked so far, and narrowing in on our best performing audiences.
To that end, we're limiting our 3P buckets to just two groups: Media & Internet Professionals and C Suite Executives. We're also layering on additional targeting tests, including geography, industry, and programming types. Our goal is to put hyperlocal CTV to the test and see how precise we can really get.
As for site retargeting and AI, we’re keeping those audiences live since they’ve been successful in driving website traffic for conversion. Like our 3P bucket, we wanted to try layering additional targeting on top of site retargeting, but the scale just isn't there. So we’re going all in on hyperlocal testing with 3P audiences.
Phase two is in progress now! Want to be the first to know how it performs?
