Where SparkToro's Data Comes From

Where SparkToro's audience data actually comes from: LinkedIn for demographics, a Datos clickstream panel for behavior, plus crawled podcasts, YouTube, keywords, and subreddits, all aggregated and anonymized.

How does SparkToro work? When you see data like this showing job and salary information about logistics decision makers who are serving East Coast ports, how is it that we know their salary, their years of experience, what employer industries they're in, what job roles they have? How do we know which social networks are more and less popular with them?

Great question. I'm going to show you in just a few minutes exactly what's under the hood of SparkToro so that when you see any report, you will know what's going on. First thing to note, our demographic data comes from LinkedIn.

So it's a little bit LinkedIn shaped. If you notice that we don't have data on folks who are under eighteen, for example, or very few folks who are in the nineteen to twenty two range, that is because they often don't have LinkedIn profiles. Of course, you can't have a LinkedIn profile if you're under eighteen and we don't collect or store or use any data about anyone who's under age, in the United States or UK or Canada, any of the places that we serve.

The data that you see here, the demographic data, job and salary, age range, gender, all of that comes from LinkedIn's sources. So this is from real people who've said, this is who I am and describe themselves and their characteristics. The data below that, the social networks that they visit, the Search and AI tools that they use, the websites that they visit.

These all come from Clickstream panel data. We use a company called Datos who's got a panel of tens of millions of devices around the world and more than ten million devices specifically in the United States. That's mobile and desktop devices.

And what it tells us is every URL that these people visit. And so essentially what we buy is the aggregated and anonymized. None of this is shown with individuals behind it.

Anonymized data about where people go on the web. So combining this data with our LinkedIn profile, we're able to say, logistics folks who serve East Coast ports, they use Instagram and Reddit and Facebook less than the average American by this amount. They're almost exactly on par with Americans for LinkedIn and Substack, which doesn't particularly surprise me.

Just surprise me. Search and AI tools, same story. We're able to say, they're adopting AI, but a little bit less than you might expect from other professionals with the exception of Claude, which is quite popular with this group right now.

And they use Google a little bit less than the average American. We know that they visit Maritime Executive and Supply Chain Dive and InboundLogistics dot com with these relative affinities. We know that these hidden gems are popular with them.

All of that comes from clickstream data. And then we get into YouTube channels and podcasts and AI prompt topics and high affinity keywords. All of this data is through the combination of sources.

So we actually go out and crawl podcasts. We crawl YouTube channels. We get search, keyword data and AI prompt data.

We crawl social networks and subreddits. And then we combine that, the information that we retrieve from those with the information that we get from Datos in our clickstream panel and our LinkedIn profile data to build the reports and the information that you see here. So when we know that lots of these people are searching for container shipping US or that they're prompting on topics like supply chain disruption solutions, or that they subscribe to the all things supply chain subreddit.

That is because we have crawled that data and combined it with these other sources. That's the basis for everything that you see inside SparkToro. We're always auditing and always improving these data sources.

So if you find anything in here, email us like many folks do and say, is this right? We will often take a look nine times out of ten. It is right.

One time out of ten, we'll fix something behind the scenes and make it right. This product is quite impressive and quite unique. I am actually unaware of any direct competitor to SparkToro at this time, but now you know how it works.