WORKWEEK

Proprietary reports are a decently common content type. And they should be, really, for any brand wanting to grow its organic search traffic, PR mentions, and overall visibility as a thought leader in space. 

What are proprietary data reports?

They are reports that analyze a large sum of data, summarize the findings and the trends, and release that information to your reading audience––for free, and often, behind a gate for lead generation. 

You can collect this data in a number of ways. Some companies leverage tools like Qualtrics or SurveyMonkey or Attest. 

With these tools, you can build surveys and even target specific audiences to send the survey too (this is the part that costs the $$$. The more consumer-focused your survey, the cheaper it will be per answer. The more business-focused your survey, the more expensive it is per answer). 

Then, once you have all of your answers, these tools give you reporting features that allow you to see variations in answers in charts, break that down by segments, change the color of the charts so you don’t even have to necessarily redesign them (SurveyMonkey is particularly good at this!). 

Other times, brands leverage their internal data for proprietary reports. Say you are a company like Shopify, and you have access to the anonymized data of >300,000 online businesses. You don’t need to go run a survey. You can ask your data science team to pull reports for you, or you can simply survey your own audience (though this works to varying degrees since you want a decent survey size to respond, and not a ton of your users will fill out a survey). 

Either way, collecting this data isn’t even the first step of this process. I wrote an older newsletter issue that talks through everything you need to do to get to this point, and then, of course, how to bring something like this to market. 

But today, I want to talk about a really crucial step between the gathering of the data and the writing of the report about the data, and that’s the analysis of the data. 

Here is how this process usually goes for me and my teams at least:

1. You don’t know what you have until you lay it all out. 

First things first, you have to go through every single question in the survey, review it, and analyze it. I typically download the data and drop it into a Google doc and even rearrange it a bit––oftentimes applying different formulas to see the full picture of a question. 

For instance, is social media marketing the main channel in which marketers are increasing spend this year, or does it only look like that because the data is skewed in some way? How can I remove that bias to see a clearer picture? 

I go through every single question in this way, including adding segmented views overtop of the ones that I find interesting. 

For instance, maybe in general brands say they are investing more resources in email automation this year, but if I apply a segment, like “brands making over $20M,” suddenly I see that those brands are actually focusing on email personalization and campaigns, instead. 

How do you know where to apply segments, or where to dive deeper in general? Well, you’re looking for anomalies or anything that is “too good to be true.” To do this, you need to have a strong understanding of your market. 

Ask yourself as you go through the data:

• Is this lining up with my expectations? 
• Why or why not? 
• How is it different? 
• How could I more clearly see those differences? 
• If it isn’t different than my expectations, is this data “too good to be true,” and am I therefore missing a way in which the answers are skewed? 

2. Make correlations, and tie them back to larger trends or activity.

In middle school at some point, I was taught to start research reports from a larger vantage point. Start big, and then detail your way down to how a specific thing you are researching impacts that larger whole. This typically makes introductions far more interesting, since people aren’t always likely to be interested in the pigmentation of hair follicles, for instance, but are interested in how UV radiation impacted skin pigmentation across the globe––and therefore created what some today call “races,” though it’s not that simple (There’s a really great Ologies episode on this!).

For data reports, you need to work backward toward that larger picture. After all, you have the details. You have the dots. Now, you need to connect them. 

So, now that you’ve gathered all of your data in the first step, look through each segmentation that applies, and then begin to jot down your overall view, opinion, and assumptions based on the data within singular charts, but also how it applies across various survey answers. 

How do different answers impact other ones? Do they increase or decrease the relevancy? Do they tell a stronger market narrative, or do they contradict the main market narrative? 

For instance, in Klaviyo data we pulled back in February 2023, it was clear that brands were planning to continue investing heavily in Facebook advertising, despite all the drama and hubbub about Apple’s iOS privacy update making Facebook less efficient. 

Now, brands investing in Facebook advertising doesn’t mean that it wasn’t made more efficient––but it does mean that the hubbub about brands “abandoning Facebook” likely wasn’t true. So what was? 

Well, given the economic uncertainty of 2023, mass layoffs, rising supply chain costs impacting margins, and more, brands didn’t have the luxury to cut off Facebook advertising entirely. They needed to get in front of new audiences for acquisition, and Facebook remains one of the best platforms for doing that, Apple’s privacy update notwithstanding. 

It signaled, then, that Facebook had indeed become a table stakes acquisition channel for brands, and that even in the wake of economic uncertainty and less effectiveness, the channel was still producing return for the brands who could figure it out––and a lot of them were investing to do just that, or increasing spend. 

That was the storyline we first came up with in our analysis document––and it was in many ways a reframe of the larger industry narrative around the time. That’s great! Because we had data to back it up, and we tied the dots together across a variety of larger industry trends and activities to build the story, too. 

That’s what you’re looking to do. 

3. Validate with external sources. 

Now, you can either write your report before this step or do this step before you write. I’ve had success with both. Either way, you do want to validate your take with external voices like your customers, your partners, industry experts and influencers, or even your internal subject matter experts. 

With these folks, you want to create a document that tells the story of the report, including the data, and ask folks (privately, of course, and often under embargo!), to offer insights or advice for readers based on what the data and analysis is saying. 

This helps to make your findings actionable, and you can often also work with these same folks to write blog posts or help promote the piece as soon as it goes live, too––creating that feeling of a larger industry report that has gone “viral.” 

If partners cannot come up with insights or advice on a section though, it might mean you need to look into it. Or more, you can ask partners what they think the data means. The more insight you can get, the better––especially if you are working with partners who are closer to the numbers and the larger industry trends and narratives than you are. 

You want to provide your audience with actionability, but you also want to make sure that you aren’t drawing lines between dots where no correlation exists. 

And that’s it. 

All of which to say, is a lot. In the journalism world, this kind of work was considered data journalism. 

• You had to get good at Excel or Google Sheets. 
• You had to understand the larger industry trends and what were expected outcomes versus anomalies 
• You had to dig into the anomalies and understand if expected outcomes were due to skewed data 
• And then you had to tell a larger story around those findings, connecting the data back to the larger narrative and, ideally, adding to it. 

I like to run content marketing programs in which we do proprietary reports once a quarter, at least. They really do provide such fantastic runway for SEO, PR, social media, paid social, and really all go-to-market functions. 

But you have to do them right. And you have to get the analysis right. And that, for the most part, falls on the content marketing team. 

This is where you have to learn data journalism skills, and flex them confidently.