[Original Research] The 2020 Conversion Benchmark Report
Oct 2019—May 2020
As the wise Andy Crestodina says, original research can be a powerful way to “provide the stats that are missing in your industry”. Whether you publish results from a simple survey—or a lengthy report—proprietary data can help you become an authority in your space.
If done well, the legitimacy of your data, your methodology, and the lengths you take to report accurately can signal that your brand’s an especially trustworthy source. By answering a question your audience really values, research can even become an annual win for you—something you can create anticipated demand for.
Oh, and there are those sweet, sweet backlinks, too…
All this to say, Andy’s right. Original research is rad.
Though, as my content team discovered in 2019, it’s not without its challenges.
Banking on a sequel (but like, the good kind.)
“Yo, we should totally remake that report from 2017!”
The original 2017 Conversion Benchmark Report
Similar to Disney a la ~2014, at Unbounce we realized older, classic content from “the vault” still performed well. And so, in 2019 we got the idea to zhuzh up something older, or fund the live-action remake, if you will.
3 years prior, we’d had success with a Conversion Benchmark report that answered the question, “how should my landing pages be converting?”(one of the biggest unsolved mysteries our audience asked us about). We’d used machine learning to analyze thousands of customer pages to reveal what was a good, poor, or average result for lead-gen pages.
However, on scoping an updated report for 2020, I discovered we’d need to start from scratch.
Our proprietary machine learning models had become even more advanced over three years (not a bad problem to have), and
The data sets from 2017 were now unavailable.
If we wanted an annual report, we’d have to go all in.
This was great because we could evaluate all aspects of the content strategy. We weren’t just making a new Star Wars. We’d be making the Mandelorian with baby Yoda on a streaming service. (is this remake analogy falling apart…?)
New data, ‘yer in!
PDF format? You’Re out.
After brainstorming what we’d replicate about the original and what we’d change, here’s what we knew the new report had to accomplish:
The new report’s production had to enable yearly replication: we wanted to be able to push a button and have the data sets for our select16 industries shoot out, ready for next year for analysis. This meant a good deal of upfront work with our data team, but it meant future-proofing our content vs. a one-time win.
We had to get this sucker out of a PDF and onto a webpage. Having the report live on a webpage ensured we’d benefit from the domain authority and backlinks, as well as be able to track which industries in the report were receiving the most attention (this could reveal good audience segments for us in future). All the same, we wanted to gate this content as it was incredibly valuable (and time consuming for the company to generate).
The report had to surface not just data, but insights. Our machine learning models were surfacing up some cool patterns in real customer data, and instead of having each industry feel like a bunch of repeat info on reading ease, sentiment, and word count—we wanted to share even greater, nuanced learnings on how to increase conversions.
The report had to be in our updated voice and tone. We wanted the latest release to feel more like infotainment vs. scientific-feeling data and analysis. How might we present things in a really approachable, actionable way?
Whew, buckle up right? This was going to be a tall order from data, design, and content.
It’d be one thing if the content team could do this mission solo, but instead…
Interdepartmental resources…assemble!
In Oct 2019 I made a formal pitch to secure the help we’d need to make the report happen. We needed a dedicated data analyst, and some wicked-schmart R&D team members (they were training the machine learning models that would sort the data sets, and unearth the insights).
My team’s Colin Loughran worked tirelessly to write the report (rounding out the insights the AI helped uncover), and I was editor (aligning the report to our strategic vision and keeping the interdepartmental production progressing). Designer Ceci Martinez created the stunning visual treatment for the report, Senior Art Director James Thomson provided creative direction, and data analyst Nicole Wright ensured this report was able to exceed all of our expectations by surfacing the insights in the first place. Rachel Scott’s team created the distribution plan.
Inside the Report
Access the full report here, or check out the short clips below.
Launch and beyond 🚀
Launched on Product Hunt May 28th (with nearly 500 upvotes, and a slick video trailer produced by Alana Thorburn Watt), we were thrilled to see so much support from the marketing community:
Results (How’d we do?)
We successfully leveraged our company’s investment in machine learning to analyze the performance of 16 industries, and 34 thousand Unbounce customer landing pages. This delivered huge value to our audience who could now see up-to-date benchmarks for their industry (and the content of the report teaches folks how to be successful if/once they start using the Unbounce product).
Our new report would now be replicable year after year, and because we invested in the digital format, the content team could now release new gated original research in the future more quickly (no more three year gaps before launching more value-based content!)
In just 4 months after launch we’d secured 8229 downloads of the report and 5219 new leads.
From May 28-Oct 17,2020 (about 5 months), 206 new trials of Unbounce could be directly attributed to the report. Contributing a total of $13,293 USD in MRR*
*This number doesn’t take into account the lifetime value of these new trials. Just the first 1x payment so far over a short period. The total revenue associated with the report will be even higher now :)
The Report became the third most visited page on our site, amassing tens of thousands of views shortly after launch.
Not bad for a sequel.