SparkToro’s Finally Here

Rand FishkinBy Rand Fishkin

Almost two years ago, Casey Henry and I joined forces on SparkToro in the hopes of solving a thorny, expensive, long-standing problem in the market research world: How to discover the publications & people that influence any group of people online.

Though we’re launching at a very scary time in the world, we’re excited to finally share that product with you. In recognition of reality: budgets are tight and the future uncertain, we’re giving anyone who signs up 10 free searches every month, no obligation. We’ll also be donating $1 for every unique account that runs a SparkToro search (free or paid) over the next 72 hours (up to $10,000) to GiveDirectly’s amazing Covid-19 fund.

You can take SparkToro’s Audience Intelligence Tool for a free spin right now (it’ll ask you to register an account & confirm your email address). My one suggestion—watch the 2-minute intro video before you dive in—it’ll help you get the most from it.

Curious for more detail about our process getting to launch, the features SparkToro offers in this first version, and how to best take advantage of the data within? There’s plenty more detail on that below.

But before we dive in, I want to address how strange it feels to launch a product we’ve been working on for years in the midst of both a deadly pandemic and a painful economic crisis. It’s weird. Uncomfortable, even. How do you ask people to try a new marketing tool when their friends and family might be sick, out of work, afraid for their lives and livelihoods?

We wrestled with these questions, too. It’s part of the reason we pushed our launch from mid-March to today. Our logic went like this:

  • Yes, it’s a hard time, but we hope our product might be able to help businesses survive and thrive in the months and years ahead.
  • Like most businesses, we don’t have infinite cash or runway; we need to launch soon to give ourselves a chance of building something that can outlast this crisis. Could we wait another month or two? Yes, probably. Would that be better for anyone? Probably not.
  • In the initial wake of COVID-19, we sensed that a marketing tool launch was not only inappropriate, but would land flat. Since then, however, we’ve seen a lot of renewed attention in the web marketing and tech startup fields. In particular, rising traffic on IndieHackers and ProductHunt told us that folks were, once again, paying attention to these topics.

Still, something was nagging at us about promoting our product, exclusively, in a time when things are so challenging for so many. So, we chose to make this launch about something other than just ourselves, and help anyone who’s talking about SparkToro feel like they, too, are having a positive impact on more than just one little startup’s odds. Hence, our decision to donate, and not just a lump sum, but an amount based on every person who tries the tool. If you’re not familiar with GiveDirectly’s Project 100, I urge you to check it out. They make direct cash payments to those with the most need, for whatever those needs are. $10,000 is a lot for a small company like us, but I hope we wind up donating every dollar.

Behind the Product’s Curtain

What is SparkToro? I think Jasper Kuria nailed it in his recent review:

“In essence, SparkToro collates information about people’s digital habits, but removes any personal identification… Rather than leaving you to search with individual sites and tools for the audience information you want, SparkToro offers it all in one place. You can almost think of it as a search engine for audience behavior.”

The heart of SparkToro’s value is uncovering which sources of influence reach your audience. Marketers, product designers, entrepreneurs, non-profits, artists, and creators of all stripes desperately need this information, but today, the options are:

  • Run expensive, hard-to-design surveys and deal with the challenges of accessing the right group, biased respondents, and self-reported data.
  • Hire a professional market research/survey design firm to run a process like the above… Which often costs $25,000+ or more.
  • Search Google, social networks, or 3rd-party tools for potential sources, in which case the challenge is guessing which ones are actually popular with the audience you want to reach (as opposed to just popular overall or high-ranking by given search criteria).
  • Investigate your audience at scale by visiting their public social and web profiles, crawling and extracting data about what they follow, visit, share, link-to, comment-on, talk-about, put in their bio, etc. It’s time-consuming and difficult, but there’s no better data available.

This last process is a variant on how SparkToro works (without the personal/potential-privacy-violation issues). Our system builds a database of anonymized profiles by crawling hundreds of millions of social accounts and websites, building connections between those that appear to be the same entity, and then searching across this dataset for profiles that match your query’s criteria. There’s more detail in the video below:

With a massive amount of data (70+ million unique profiles representing billions of social accounts, shares, and websites) compiled and regularly updating (~500K profiles updated daily; ~120 days for a full refresh cycle), SparkToro has the power to search for any audience and get remarkable data, at scale, in seconds, with verifiable accuracy:

  • Social accounts they follow
  • Websites they visit
  • Podcasts they listen-to
  • YouTube channels to which they subscribe
  • Breakdown of the words and phrases they talk about most often in their public shares
  • Commonly-appearing words and phrases in their profile bios
  • Frequently-used hashtags
  • City-level geographic distribution
  • And more…

If you already know your niche well, this data will likely confirm a lot of your existing assumptions (and almost certainly show you some opportunities you didn’t know). But if you need proof of which sources are popular (say, to convince a boss, team, or client to spend in the right places rather than the wrong ones), or if you need to better understand another audience with whom you’re less familiar, SparkToro becomes invaluable.

Even if you already have a great understand of your niche, you’ll almost always find surprises in the data. Scroll down far enough in the sources of influence and you’re certain to discover channels, accounts, sites, and podcasts you haven’t heard of that are surprisingly well-followed in your space. Or, try using the Hidden Gems filter, one of our early users’ favorite parts of the product:

What (and Who) is SparkToro For?

Our early customers include a range of professionals, all of whom care about better understanding their audiences’ sources of influence. Those include:

  • Entrepreneurs and product builders seeking to tap into the zeitgeist of a field as they build for their customers.
  • Market researchers tasked with understanding a field for their projects or clients.
  • Ad buyers seeking sources for sponsorships, partnerships, or paid ads without going through the creative and targeting constraints of the major ad networks.
  • Content marketers & PRs looking for sources of coverage for their pitches.
  • Social media marketers looking to earn amplification for their messages in ways that will reach a targeted audience

The video below shows off a number of these use-cases, and how to get that value in the tool.

Limitations of SparkToro’s V1

The first version of this tool is, we think, fairly remarkable, but far from perfect. Some of our initial limitations include:

  • V1 is English-language only and the dataset is almost entirely from accounts and websites in English-speaking countries.
  • Our coverage of the United States, United Kingdom, Ireland, Canada, Australia, and South Africa is reasonably decent. Outside those countries, our data tends to be less reliable because the audience sizes aren’t large enough.
  • We don’t collect (and, due to privacy issues, don’t plan to collect) demographic data like gender, race, sexuality, income, ability status, etc. We also don’t buy this data from third parties or try to make guesses or inferences based on behavior. For now, SparkToro’s exclusively focused on literal text that appears on publicly accessible social accounts and websites; nothing else.
  • SparkToro is useful if your audience describes themselves or talks about their topic online. It’s not very applicable, however, if they don’t. As I often told our beta testers over the last year, “SparkToro’s great if you’re seeking data about real estate agents, not so great if you want data about homeowners.”
  • In very small niches, especially when those niches are combined with smaller geographies, SparkToro’s data is often not enough to provide extensive information. If you narrow an audience of a few thousand to a single city, you may only have a few dozen profiles worth of data. Those can still be revealing, but they may be less valuable than a larger group. Expanding to a bigger geography or using broader descriptors can help.
  • Our most common request (that we haven’t fulfilled yet)is for PDF or CSV exports of the data on the Audience Insights tab. That’s in progress, we promise 😉

We’re always looking for feedback and suggestions on how to get better, so please don’t hesitate to drop a line to [email protected] if you’ve got ideas for us.

Our Launch Process

It’s certainly the case that we took an unusual, longer-than-normal path to reach public launch. In part, that’s intentional. We wanted to get the tool far past the “Minimally Viable Product” and into “Exceptionally Viable Product” territory. That meant spending a significant time on iterative development — building things, launching them to a small, private group of testers, and seeing how they reacted. Then, inviting more small groups and watching relative excitement and product-use of those folks over time.

Whiteboard of the problems SparkToro helps solves.

In the first year, we encountered plenty of challenges. There were days and weeks when both of us worried about whether the core concept behind our solution (crawling & aggregating billions of social account and website data points into distinct profiles, then analyzing their shared attributes & behaviors) would even work. But in the Spring of 2019, we started to get excited. The results looked… promising.

From July to November, we ran several beta testing cohorts, iterating on the product each time. We added features folks wanted, removed some that didn’t work well, upgraded our data, and tinkered with the interface. We worked with the superb team at Elevate to analyze the results of our beta and recommend steps that needed to be taken before a public launch. In mid-November, it became clear that while the product was useful, the UI was outdated.

Image

Above: the beta interface of SparkToro’s tool

So, we contracted Monty Hayton after finding his work on Dribbble, and after a furious two months of revisions, launched a new interface for the product. The reception was immediate: this was a dramatic improvement, and we were ready to move to early access.

Above: SparkToro’s updated interface, designed by Monty Hayton

In late February, we sent our first early access emails, with a terrific reception. A week later, our second cohort didn’t perform nearly as well; the threat of COVID-19 was clearly showing. We hit pause on our upcoming launch, stepped back for a month, kept working on tweaks and tune-ups, and then, as we settled into early April, determined we’d aim for this month in spite of the challenging climate.

I’m really glad we raised as much money as we did. It gave us the patience to build carefully, test thoroughly, and tune-up until we had something to be truly proud of. The old startup adage goes: “if you’re not embarrassed by your first version, you launched too late.” We rejected that nonsense, and I’m glad we did. I think SparkToro’s worth the wait, and had we launched that early Alpha version last Spring or the beta version in the Fall, we’d be providing a far less useful, usable product.

Don’t Take Our Word For It

As part of SparkToro’s launch, we offered press review access to a number of folks in the media, marketing world, and blogosphere. We asked that they be honest about their thoughts on the product and not pull punches about the downsides.

Our requests for press reviewers of SparkToro’s early access

I’m truly humbled to say that of the folks who’ve published reviews so far, well… I’ll let you judge for yourself.

Reviews of SparkToro

And they’re not alone. The talk on Twitter has been encouraging, too.

There’s dozens more, hundreds even, and I’m grateful for every one.

I’m a second-time founder. And after Moz, I’ll be honest, I feared—hell, I still fear—that I might not be able to build something people would want and use again. I think a lot of entrepreneurs have that fear. It probably keeps many of us from even trying again. So, yeah, when I see people saying kind things about this product we’ve developed, it’s special. It’s affirming.

I hope you’ll give SparkToro a spin. Even if you don’t need it now, I hope you do someday. Building things that help others, and reaching them in the places where they pay attention… without having to give the big ad monopolies your money? That’s a company I’m excited to build.

P.S. SparkToro might look like a two-person effort, but that’s far from reality. Casey and I would like to extend huge thanks to the folks who helped make this product and company happen, including our 35 investors, Christine Ryu (who helped make SparkToro’s UX exceptional), Dawn Shepard (the genius behind our branding, artwork, and visual design), Tyler Foster (who filmed & edited all our videos), Ashley Long (who made sure our legal & privacy issues were rock solid), Jennifer Newkirk (our immensely kind, responsive accountant), Claire Suellentrop and Georgiana Laudi (our incredible SaaS consultants), and Geraldine DeRuiter and Lindsay Henry (our patient, loving, ever-supportive partners). Thank you all!

Keep up with SparkToro®