By What does your audience talk about online? How do they describe themselves? What hashtags do they use? Prior to social media aggregation software, answering these questions was a maddeningly manual, lengthy process. Unfortunately, most of those products are exorbitantly expensive, enterprise tools. After SparkToro’s launch last year, we were somewhat shocked to find:
A) How valuable these text insights were to our users (most of whom are not in enterprise/bigco world)
B) How creatively folks applied this data (see below)
C) How hard SparkToro previously made it for folks to reach and use this information (doh!)
Today, we’re fixing that with the new Text Insights tab.
When you perform a search in SparkToro, you’ll see it on the left-hand menu (beneath “Overview”). From here, you can:
- Research far more text & content data than previously available (up to 250!! vs. the previous 30 rows)
- See the percent of the searched-for audience that’s employed these words, phrases, & hashtags in the last ~120 days (the average time period over which SparkToro’s index fully refreshes)
- Select text and hashtags to add to your Lists (for export, use in campaigns, ad targeting, content, etc)
- Get quick access to search those keywords/hashtags on social platforms and Google Trends
- Export a selection of text insights (based on which boxes you check) or the entirety of the data to CSV
Long story short, it’s vastly more robust, functional, insight-rich, and integrated with the rest of the SparkToro application. To those who’ve nudged us to invest more in the text data we index across Twitter, Instagram, LinkedIn, Facebook, and more, thank you! You were absolutely right—this wasn’t a use-case we initially conceived of in SparkToro’s development, but it’s proven to be a core part of why marketers of all stripes use our application.
What’s Inside Text Insights?
There are four unique sections in Text Insights: Words in Bios, Hashtags Used, Phrases Used, and Words Used.
These four subsections contain a wealth of easy-to-explore data, but the more you know about how we discover and calculate it, the better your results when using it are likely to be.
Words in Bios: This data, inclusive of both individual words and multi-word phrases, comes from the text of the social profiles that make up the search results. In the example above, SparkToro’s database found 1,566 profiles containing the word “mixologist.” Below, you can see the words we frequently saw in the bio/profile text of those accounts, the percent of audience, and get quick links to search those words in Google Trends, Twitter, Facebook, LinkedIn, and Google search (those icons will open the link in a new window).
Of those 1,566 profiles, 133 also contained the word “bartender” (8.5%). Another 110 contained the word “amateur,” (7.0%). And, so on. This profile text may not necessarily be from the same network/profile page. For example, my Twitter account (@randfish) has the words “Moz,” “marketing,” “tech,” and “startups.”
Rand’s Twitter (above) and Instagram (below)
Meanwhile, my Instagram profile (@randderuiter), has the words “pasta,” “life,” and “form.” SparkToro’s database contains all of these words, and a search for people whose profiles contain “pasta” or “marketing” could include my data (though individuals are anonymized and aggregated in the SparkToro search process, so you’d never actually see me in the results, just the aggregation of my data with hundreds or thousands of others).
Hashtags Used: Possibly my favorite data in Text Insights because of its numerous, powerful applications, hashtags also come from a variety of networks (primarily Twitter, Instagram, Facebook, and Linkedin). Hashtags are often one of the best ways to search SparkToro, too, because profiles using a hashtag frequently have distinct, marketing-applicable traits than broader words and phrases.
The hashtags tab will also show quicklinks to searches on Twitter, Instagram, Facebook, and LinkedIn for that tag, along with the percent of audience we saw using them. Note that just because the numbers might be in single digits doesn’t mean a much wider group isn’t seeing or following those hashtags! SparkToro can only measure actual use, so 5% of VPs of Sales using the #B2B hashtag likely implies a much wider segment of that audience sees content containing those tags.
Phrases Used / Words Used: These two tabs work exactly the same, but Phrases Used shows multi-word text found in a profile’s posts and content, while Words Used shows only single words. These are sourced from posts, replies, and comments on Twitter, Instagram, Facebook, LinkedIn, Pinterest, Medium, Quora, Reddit, YouTube, and Github. It’s rare that all ten of these networks are present for any given profile, but most profiles in our index have at least 2-3 of the above.
In the example above, we see 4.7% of profiles that have talked about “raku pottery” also using the phrase “throw down” (likely many of those reference the popular, utterly soothing HBO show Geraldine and I have been watching, The Great Pottery Throw Down). That doesn’t mean that *only* 4.7% of people who’ve talked about raku pottery are watching or engaging with the phrase “throw down,” but rather that, of the 881 people in our database whose public tweets, Instagram posts, Facebook posts, etc. used those words multiple times in the last 120 days, 4.7% of their posts that SparkToro crawled & indexed also contained the words “throw” and “down” in succession.
Basically, I’m saying you shouldn’t view small numbers, even as small as 0.5-1%, as meaning that there’s little discussion or engagement on these topics. In fact, 1% of all online pottery fans discussing something publicly likely means there’s a much bigger number of engaged, aware, potentially-curious people in that world.
How Can I Apply This Data?
SparkToro’s customers have been both immensely creative and thoughtfully pragmatic in their applications of text-based data. I talk regularly to hundreds of SparkToro users, and get to hear stories of all kinds. Many of those include use-cases like:
- Content creation – search demand data is the standard “go-to” for many content marketers, but if you want to stand out from the crowd, be among the early creators of cutting-edge topics, or find content affinities your search-data-only peers are missing, text from social discussions and use can be invaluable. A pottery website might be inspired to consider including content on sterling silver (I honestly don’t know how these two topics overlap, but the data above suggests they definitely do). VPs of Sales might be more reachable through cybersecurity topics than another debate on sales-qualified-leads. And now might be the season to get a jumpstart on gin recipes, recommendations, and deals for the mixologists in your audience (I know, I know, the Negroni’s never gone out of fashion).
- Social media marketing – If you’re looking to engage around topics and hashtags on Twitter, LinkedIn, Instagram, and Facebook to build your audience and brand, SparkToro’s text data is hard to beat. In addition to inspiring valuable listening and engagement opportunities, you can also analyze a competitor’s social account and see what topics are engaging their audiences… Chances are, you’ll find useful ways to interact that they’re missing.
- PR pitches – getting a journalist, content outlet, industry publication, or blog to cover a topic and include your story/mention/brand/link is a painful, low-success-rate process. But a pitch that shows, with data, how interested their readers (or potential readers) are in that subject, has the potential to go much further. Explaining that “12% of your readers talked about XYZ in the past quarter, but you haven’t published anything since 2018,” makes for a solid nudge (though I might suggest you break the news less bluntly).
- Social ad targeting – Facebook, Reddit, Twitter, Pinterest, and LinkedIn’s ad builder tools leave much to be desired in terms of topic targeting. In Facebook, I find it particularly infuriating that you have to *know* what to start typing in order to receive the category suggestions their platform will allow. Thankfully, SparkToro’s text insights will give you precisely those starting points for building the ad audiences you’re after. Search iterations and the social tab can inspire even more opportunities on this front (and of course, hashtags are pure gold for this).
- PPC & SEO keyword research – Similarly, text insights can quickly be checked off, added to a list, and exported for keyword research in Google or Bing’s search ads. Having built and used plenty of keyword research tools in the past, I can promise that the text data you’ll get from social profiles and conversations is massively different from what’s shown in standard keyword tools. That likely means more work, as not every word or phrase will be a match, but it also means a competitive advantage over those who ignore this datasource.
- Market research – if you’ve been tasked with understanding an audience or group of people, there’s no better way of getting a bird’s eye view of who they are and what they’re about then looking at the words & phrases they use to describe themselves, the topics they discuss, the hashtags they post, and the sources they follow. This weekend, I was analyzing the crowd that follows Popular Mechanics, a longstanding publication in the everyday engineering and sciences field. And, sure enough, 3.4% of their followers use the word “engineer” in their profile. But, 1.4% use “designer,” 1.5% “photographer,” and another 1.5% “author” or “founder.” The market research reveals that Popular Mechanics has a more diverse, more creative audience than I’d have expected.
- Persona building – I’ve shared some thoughts on my approach to personas (essentially: use them to solve for problems you have rather than building them and then finding problems you can apply them to), but no matter how or why you’re building them, statistically sound, sampled data at scale is essential to a quality model. Surveys might help you get at some of this, but can’t compete for accuracy, coverage, scale, or depth with profile data from thousands of anonymized, unbiased, real world samples.
- Copywriting and messaging – we all make assumptions about what our customer are familiar with, how they talk about a topic, what words and phrases they know, use, and recognize… but often, our assumptions are wrong. Data about what an audience *does* use and doesn’t, how they describe themselves, how they adopt hashtags, etc. can be invaluable for writing ad copy, product copy, messaging materials, even content headlines.
The above are just a sample of the applications. I’ve talked to folks whose use-cases stretch beyond marketing, advertising, content, or influence to all sorts of profile-data-at-scale needs. The bottom line is that when you need to know what a distinct group of people are talking about online, or how they describe themselves, there’s no substitute for the precision and clarity of direct aggregation. Surveys and interviews are amazing for all sorts of other functions, but will never reveal this type of information as accurately or comprehensively.
One last bit of advice: Scroll down!
ABOVE: I’ve copied+pasted rows 150-155 below rows 1 and 2 to illustrate
You’ll be amazed by how relevant and useful the text insights are, even those used by small percents of an audience. E.G. Who knew SparkToro’s fans had as many musicians as SEO consultants?! Might even be worth us building a case study or two on using the product for promotion of an album launch 🙂
If you haven’t already, give the new Text Insights a spin! It’s available to all free SparkToro users, and of course paying subscribers as well. Data in the tab is limited if you’re a free user, but you can still get a great deal of value from the samples provided. As always, if you’ve got feedback or questions, drop us a comment below or an email. Casey and I take feedback to heart (this very feature was inspired by your suggestions!) and love to hear from you.