How Uberflip ran a data-driven experiment in content activation w/ Jeff Kew, Andriy Koval, and Waqas Khan [The Flipside]
Welcome to The Flipside, where we follow the adventures of a shiny new marketing team as they navigate the ups and downs of using Uberflip to drive our go-to-market. From triumphs to tribulations, we're sharing the whole story.
Very little great marketing gets done in a silo. Even small-time tactics require a team effort to build something that delivers results. And when people in different functions work together informed by broader strategy, that’s when the magic happens.
In this week’s The Flipside, we’re showcasing one such superhero team here at Uberflip: a data-driven experiment by marketers that may result in additional functionality for customers.
I sat down with all three Uberflippers to discuss their efforts to integrate Uberflip CTAs into our website as effectively as possible: Jeff Kew, a seasoned veteran of marketing ops space; Andriy Koval, our front-end web developer; and Waqas Khan, a demand gen marketer working to refine our paid ad strategy.
Our discussion revealed the importance of adopting a collaborative, experimental mindset whenever you’re innovating—and why you sometimes need to hack together an imperfect solution before you can find a clear path forward.
Colin: Hi everyone. Before we jump into the CTA project, why don’t you each speak to your roles here at Uberflip? Jeff, why don’t we start with you?
Jeff: Thanks Colin, my role is the Director of Marketing Operations. I manage the technology platforms we use to support our go-to-market efforts.
Andriy: I’m Andriy Koval, a web developer on the marketing team here at Uberflip, working closely with Jeff on integrating all our tech into our platforms.
Waqas: My role is Demand Generation Specialist. I'm responsible for running campaigns—right now, specifically for demo requests, increasing engagement and awareness. I'm also trying to understand different segments—like MQLs and SQLs—and how to move these segments through the funnel.
C: Jeff, can you give me a quick overview—an elevator pitch—for this project?
J: Sure. We wanted to better leverage Uberflip CTAs in the form conversion process.
Essentially, changing the flow of capturing form data using a Marketo form to capturing the data using our native Uberflip forms. This allows us to instantly track the converted visitor across our hub and personalize their experience faster, while also pushing all of the data to Marketo using our native integration.
C: Interesting. So what inspired you to go down this road?
A: This started in a conversation between Jeff and me about how we use forms on our website.
Historically, we've been using Marketo in a few ways. We embed Marketo forms throughout our website, for instance, and we used it as a native integration with Marketo on our Uberflip hub.
But we noticed, though, that that setup could work more efficiently for personalization.
J: Marketo forms on a Marketo landing page work extremely well. Marketo is a very robust platform, and the forms are very robust too. But we knew there was more we could achieve by combining Uberflip CTAs with Marketo than we weren’t getting by using Marketo forms on their own.
Previously, when visitors filled out a Marketo form, we would have to wait until someone later filled out an Uberflip CTA associated with a stream or item before we could build a more detailed tracking history, including first touch, last touch, and UTM data pushed into Uberflip Analytics.
If they filled out an Uberflip CTA first, though, the data flows into Marketo automatically through the native integration.
C: And that’s where you started?
J: The ability to add an Uberflip CTA to a page on your website was developed, but needed some tweaking to capture the kind of attribution that we use. It wasn't a priority yet on the product roadmap. For me, it was like, well, why not? From a marketing ops perspective, this makes the most sense for us and our customers.
Fortunately, Andriy, owning web ops, has that dev experience and could build and test it.
A: Yeah, that’s how things started. I was, like, yeah, okay let me give it a try. I’ll try to iframe it. I’ll try this, I’ll try that. We tried a few different things actually. Really experimenting to understand how we can solve this problem for Jeff.
C: Waqas, I know you’ve been a contributor to this project too but you have a slightly different objective in mind. Do you want to give us your perspective?
W: At the start, my separate goal was to test the performance of using Uberflip items as landing pages compared to our WordPress landing pages on the website, to find out which are better optimized for conversion rates in different contexts.
We ran a few tests running ads promoting a downloadable ebook, though, and we were seeing a much higher bounce rate sending people directly to an Uberflip item compared to a WordPress landing page. J: Interestingly enough, before our time at Uberflip, there was a study done to show that if a visitor went to a hub item—they landed on a stream, they clicked on an item—the conversion, once they were in, was higher than a traditional landing page or squeeze page. The study showed that people converted 17% higher on an overlaid CTA compared to a landing page.
But when Waqas repeated the test, we saw the opposite results. That makes sense—it’s a different world, audiences are different, behaviors change, and the market changes.
W: So really, my involvement started with me trying to make a data-informed decision on which approach to use for our demand campaigns: a squeeze page or an Uberflip stream.
I knew if we can embed an Uberflip CTA on a WordPress squeeze page, we could also take advantage of the benefits specific to Uberflip that Jeff’s already mentioned—while still being able to visually optimize it for conversions as our WordPress page layout does.
Retargeting for demand audiences and running ABM campaigns are also way easier with Uberflip. And with the 6sense integration, we can find out which companies are engaging with us. Then we can better work with the sales team and give them data to go after these accounts that are showing strong signals.
So I was excited by the work they were doing.
C: Andriy mentioned earlier that we were experimenting to find a solution. How’d that go?
J: Using an Uberflip CTA was the key, but that takes us to the technical problem we faced.
With a Marketo form, you can generate an embed script and drop it into the page at a specific location and it generates a form for you. It’s a very tight integration.
But using an Uberflip CTA didn’t do that out of the box. It only offered the ability to iframe the form on that landing page, which is sometimes not ideal.
C: Why not, Andriy?
A: There are pros and cons, obviously. I like to use the analogy of a house. An embedded form actually lives on your website, it’s in the house or your living room. So, any attribution you’ve added to the page is instantly associated with that form from your house.
But an iframe is like looking through the window at the house across a street, and the problem becomes, how do you get that information into your house?
An iframe implementation works in some use cases, and it’s faster to set up. In many instances, though, it can be harder to customize it—and it might be slightly harder for customers to use effectively. And we were worried that with how Jeff wanted to use it with Marketo, it might not work correctly.
But iframing was a step in the right direction to give it a try and see how this performs.
C: I understand the current implementation is an evolution of that first version.
J: The first version that Andriy implemented was very much a proof of concept to show that we can push information through an Uberflip CTA from the website to the hub. But it wasn't something that was going to be scalable in the sense that our customers couldn’t apply it because it still required a lot of customization from Andriy.
A: Yeah, so we had help from Yoav in taking it to the next step. [eds. note, Yoav Schwartz is Uberflip Co-Founder and CEO]. He had more access and knowledge of how the Uberflip hub works. So he was able to review when Andriy developed and build a better implementation, the one that we're testing right now.
J: In a way, what Yoav went and did—Andriy, made that possible. Yoav was able to see a lot of that work that Andriy had already done and packaged it into a piece of code that could easily be added as literally a toggle switch on or off in the Uberflip frontend.
Now the customer can actually go and take that script, and just drop it into their WordPress page. At this current stage, all you need to do is quickly edit the script and change the CTA ID, but otherwise, it’s plug-and-play.
But Andriy, I don’t think I’ve talked to you yet, but is there a way to build a module on that WordPress page where the marketer can just drop it in?
At this point in the interview, Jeff and Andriy began to discuss adding additional functionality to make Uberflip CTAs even easier to use with Uberflip. I’ve omitted the details, but it’s really obvious that when you get people like Jeff, Andriy, and Waqas together, they’re incredibly enthusiastic about experimenting to solve a problem.
C: We've called this project an experiment. Is it still in the experimental phase? Are we still seeing how it works against a traditional setup? Or are we all in on this?
J: We are. Waqas is running experiments right now. I set up the Marketo program for him to capture those leads on that landing page he created based on this new, more customer-friendly solution.
W: Right now, we’re running a lot of segmented campaigns that point to the WordPress page with a Marketo form. But also we’re testing an ad group that uses an iframed solution, sending people to a landing page that has an Uberflip CTA on it. So our goal is to run this campaign now and see if it's impacting conversion rates or if the conversion rates are generally the same. If it works, we can implement this solution across all our campaigns.
Once that's done, we’re going to run further tests to understand more about what converts.
What I want to do is run a campaign where we will promote a downloadable item, like an ebook. The test is going to be comparing a WordPress page with an Uberflip CTA versus sending them to an Uberflip item directly, with a show-beside or a pop-up CTA. And that way, we'll be able to understand how these perform—and that is going to be the real test.
C: What are the initial results looking like? Any learnings you’d like to share?
J: It’s still a work in progress. We've improved our tracking on the Uberflip hub side and maintained tracking on the Marketo side. As we expand our use of the hub CTA across all conversion pages, we'll approach 100% tracking of all conversions in Uberflip Analytics.
As of today, we have an Uberflip analytics report showing us the CTA conversions since January that are converting as high as 33%. Plus, we can now identify each use of the global CTA we have for demo requests.
W: It's going to make sharing our results in a future installment of The Flipside so much easier, that's for sure.
Improved tracking in action lets us see all conversions in Uberflip Analytics.
C: I can't wait, Waqas.
There’s an amazing story here.
The three of you started by trying to solve some challenges you were facing and better understand our prospects' behavior. Then you came up with a hacked-together proof of concept to help you solve the problem. And now it’s been refined into something our customers can use too.
A: Yeah, I just think it's cool. Like we're experimenting, we're building something new and really cool. It's always fun to experiment, you know, and even help the product team develop something that will work better in the future.
J: It's up to the product team to write something around this plugin, this new code so that it toggles on or off and is easy for customers to use. Then we'll want to document it and show customers how they would benefit the same way we did.
W: You made it sound like we’re some sort of heroes. So thank you for that. [laughs]