How does SparkToro know which accounts, websites, podcasts, and channels are popular with a given audience?

SparkToro combines data from four unique sources:

  1. Clickstream panel data (from our partners at Datos), which powers our traffic estimates and cross-domain visitation results (i.e. people who do X on the web also do Y)
  2. Social profile data from LinkedIn (via our partners at Leadfuze), and crawled data from additional networks (e.g. Facebook, Instagram, Twitter)
  3. Keyword search data from Google (via our partners at MarketMuse), which powers the volume estimates, CPC, and other keyword metrics
  4. Web crawl data run by SparkToro itself to collect information on YouTube channels, podcasts, Subreddits, and more

We build systems to intelligently merge this data and produce the insights you see about any given audience: the websites they visit, keywords they search for, podcasts they listen to, YouTube channels to which they subscribe, and many more. This core data is then also passed to LLMs to provide additional insights into things like topics of interest and opportunities to take action (content ideas, broad interests, etc.).

When you enter a search, SparkToro identifies matching profiles, website visitation and keyword search data, and uses these to extract information about what else the audience does on the web and who they are. For any given bit of information, we offer:

  • A 0-100 affinity score – the relative degree of usage/subscription/engagement/visitation between an audience and an activity. For example, the “High-Affinity Websites” tab score might show that a particular website is 60/100 and another is 30/100, indicating that first is twice as likely to be visited than the second by the searched-for audience.
  • Some measure of popularity – e.g. the number of subscribers a YouTube channel or subReddit has, the number of reviews from a podcast, the traffic estimate for a website, or the estimated volume for a Google search term.

SparkToro’s data is not without bias! Our clickstream panel doesn’t include every device on the planet, and we knowingly overweight to profiles present on LinkedIn (our primary data source), which means you’ll generally see data about more educated, employed, and “professional” individuals. We also intentionally exclude data on anyone under 18 and work to remove all adult content websites, social accounts, and keywords.