NEW: Custom Search & Custom Audiences Offer Incredible New Ways to Use SparkToro

Updated April 10, 2024: We relaunched SparkToro with new data sources. In that update, we also renamed “Custom Search” to “Advanced Search.” If you have a paid SparkToro account, you can access Advanced Search in your dashboard.

Three years ago, if you wanted to uncover which podcasts Data Scientists listened to, or find out which social accounts were most followed by people who talk about Organic Chemistry, you’d need months of surveys, interviews, and manual data collection.

Then, in April of 2020, SparkToro launched, and high-quality answers to those questions could be found in seconds.

Since launch, SparkToro’s had a big weak point: multiple vectors for audience searches weren’t possible in the app. If you wanted to know, for example:

  • Only those Data Scientists who talk about R (if that’s your target persona, digging into that specific group vs. all R-enthusiasts or all data scientists is a huge win)
  • Only those folks posting about Organic Chemistry who also follow the publication: New Scientist (if you’re considering a partnership or ad deal with New Scientist, you might first want to confirm that you can reach the right audience)
  • Only those startup founders who follow @Hubspot on one or more of their social profiles (if Hubspot’s content team is trying to reach more founders, this query will show exactly where and how to do that)

If you want to see it in action rather than read about it, these 107 seconds should do the trick:

The best part is how easy this is to do:

On the Custom Search tab, just enter the criteria you’re seeking, and click “Test & Save Query.”

SparkToro’s database will quickly run a test version of the query to see how many profiles match the criteria you’re after. You’ll then receive a message like this:

Every custom search you set up is saved in your account, so you can easily run them again later. The results automatically appear, usually in 10 seconds or less (though more complex, larger queries can take more time), with the same visualization you’re used-to from an Audience Research query:

I find it fascinating that a large percent of founders who follow HubSpot are talking about e-commerce, and equally interesting that so many use “author” in their bio. When you scroll down and see the sources of influence, it’s equally illuminating—a near-perfect overlap of entrepreneurship publications and marketing-centric ones.

Three final notes:

  1. Because SparkToro does sampling on large audiences (e.g., if 20K profiles match a search, we might analyze 4-5K of them rapidly to get faster/less-processing-intensive results), it’s often the case that a Custom Search query will show more or fewer results than the percent of audience for a behavior multiplied by the audience size. I’ve noticed it more with hashtags and websites than with words in bio, frequently talked about phrases, or followed-social accounts, but YMMV.
  2. Custom searches cannot currently be tracked in Audience Tracking, but if you’re passionate about doing so, please drop us a line direct: [email protected]. It’s not available quite yet, but we hope to be able to support it.
  3. Custom Search also supports “but, does not” as a modifier to your query, so you can specify things like “My audience uses these words in their profile: doctor, AND frequently talks about: Covid, BUT, DOES NOT frequently talk about: hoax”

Just be aware that the more specific your query gets, especially with very narrow groups, the harder it becomes to find enough profiles that match. We recommend starting broader vs. narrower.

Custom Audiences Lets You Analyze Your Email List

Since launch, we’ve received a lot of requests from folks who hoped to send us their email list and get back an audience research report on the folks who’ve subscribed. It’s a very smart way to think about marketing: if the folks who signed up for your newsletter/product emails/etc. are good fits for what you offer, the next logical step is to find more of them!

For the last year, we’ve only been able to do this once you applied enrichment data to your email list, which meant going first to Clearbit, FullContact, Pipl, or a similar provider and paying those firms to append Twitter or other social media URLs to the emails. Only then could we run a report on that data.

To a few folks who jumped through these hoops, the results were powerful. Agencies and in-house analysts took SparkToro’s audience research report back to leadership teams to prove the case for investments like:

  • PR campaigns with the publications their email subscribers read
  • Advertising on the YouTube channels, websites, and podcasts they subscribed-to
  • Content campaigns with the words, phrases, and hashtags popular among these subscribers
  • Social media marketing through and with the accounts they paid attention to

Incredibly cool. But a real pain in the tuchus*.

Starting now, we’re able to manage the custom audiences process directly. Here’s how it works:

  1. Send an email to [email protected] with details about your list (e.g. “we’ve got 2,000 Twitter URLs” or “our email list has 15,000 folks in the B2B sales space”)
  2. Casey will ask you to share your list and can then tell you how many matches we’ve got (if you need an NDA specific to the list shared, we can sign that). If we can’t find enough matches in our existing system, you can use identity resolution services like the aforementioned FullContact/Clearbit
  3. If you want to proceed, sign up for a SparkToro Agency tier account (or upgrade to one if you’re not already on that plan), and we’ll convert your list into a custom SparkToro search you can find in your account

Ta da! Custom audiences, right in the SparkToro UI you’re used to, but with data on exactly the people you care to analyze. If you’ve always wondered which social accounts, YouTube channels, hashtags, words + phrases, podcasts, and press sources were popular with your email list, this is massively faster and cheaper than anything else we’re aware of to solve that problem.


In other news, SparkToro’s data structure is getting a bit of an overhaul this year, and I thought it would only be right to mention it here. Long story short: we want to ingest and display more kinds of data from more sources, in more ways, and not be tied to any singular identity system (right now, Twitter sits at the center of our ID system, because it’s more likely to be connected to websites and other social accounts than others). But, since November, we’ve worried that Twitter is no longer a reliable player, and its usage is both more fragile and on the decline in ways that have us worried.

See? We’re hard at work on a new SparkToro infrastructure.

In the next 6-9 months, you’ll likely see SparkToro rolling out a new, more robust, more data-rich system. Amanda, Casey, and I did some great brainstorming at a rare in-person meeting last week, and we’re both excited and relatively confident in this new concept.

* that’s Yiddish for a person’s backside