By 3 Years ago, SparkToro launched in the midst of a world-changing pandemic. In 2021 and 2022 I wrote updates about this business’ journey, both because I love to transparently share the adventure, and because we’re passionate about spreading the model SparkToro’s used to fund, grow, and bring value to customers. Let’s continue that tradition today 😊
Here’s the good news:
- We’ve been profitable every month since Nov. 2020
- We’ve just issued checks to our 35 investors; they’re receiving 100% of their initial investment. Those lovely folks who believed in us can now expect to participate pro rata in future dividends or M&A (should we ever go that route).
- Growth has come in fits and starts, with periods of acceleration followed by plateaus; it’s hard to know what the future holds, but we wouldn’t be surprised if this pattern generally continues.
- While we’ve had periods of stress and longer hours, we (Amanda, Casey, and Rand) don’t engage in hustle culture. We’re an all-remote team that takes care of family, sees friends, travels, and leads multi-faceted, fulfilling, Chill Work lives. We aim for ~30-40 hours/week on SparkToro, done whenever we work best.
- We’ve helped more than 100,000 marketers do audience research, including thousands of paid subscribers. And we don’t obsess about churn; focusing instead of true lifetime value.
And the bad news:
- Following the API changes at Twitter, we were forced to shut down our popular free tools and sunset the Audience Tracking feature. Our Q1/Q2 feature roadmap also took a backseat to an overhaul of SparkToro’s profile connector system.
- Our reliance on Twitter as a connector network is ending—in 3-4 months, we expect to launch a new, V2 of SparkToro, with additional features, more robust data, and a network-agnostic infrastructure.
- Growth in Q1 was strong, but Q2 is down (perhaps as a result of the above)
More about all of this in the post.
Chart of SparkToro’s MRR from our favorite tool, Profitwell
Investor Payback
Let’s start by talking about the milestone SparkToro’s just hit: returning 100% of our investors’ original funding. For us, this is a huge proof-of-concept moment for our unique (and open-sourced) investment structure. Our goal is still to return multiples of capital over the years ahead, but we’ve hit, at the least, a minimum bar. Here’s how the history played out:
- March, 2018: Rand leaves Moz and starts SparkToro the next day
- June, 2018: We close our investment round of $1.3M from 35 individual angels (no institutional capital) who put in between $10,000 and $100,000 each. Casey Henry joins as cofounder.
- June, 2018-March, 2020: We spend ~$700,000 building V1 of the product, taking us to ~$600K in capital. The majority is salaries & healthcare, but R&D, data purchases, cloud computing, legal & accounting, customer research consulting, and a bit of T&E is also included.
- April, 2020: SparkToro launches
- November, 2020: We hit breakeven profitability, with ~$400K left in the bank
- July, 2021: SparkToro hires Amanda Natividad as VP Marketing
- December, 2022: Profitability’s brought us back above our post-investment cash level, but rather than immediate payback, we put $1.2M in a US Treasury Bond with the intent of earning a little interest while we build up a cash cushion.
- June, 2023: We’re at a healthy cash cushion with continued profitability and the T-Bill payback, and send an email to investors informing them of the upcoming payback.
More than a few of our investors wrote back with notes like Matt’s and Dan’s:
Sorry. Short break. I got something in my eyes.
Here’s the strangest part… my previous company, Moz, employed 200+ folks, raised $30M+ in investment, and produced total revenues of nearly a quarter billion dollars ($250M) in the 17 years I was there. But, despite those metrics, that company failed to live up to the promises I made to our investors, team, customers, and community. SparkToro is 1% Moz’s size by most metrics, but is already closer to living up to its more modest pledge. I still want to write about the end of Moz someday, but in the meantime, you can watch this 9-minute video (filmed in 2021) about the differences between the two ventures:
My biggest takeaway: make reasonable promises.
I know some entrepreneurs love to chase the 1-in-1,000, become the $1B Unicorn or bankrupt-the-company-trying model that’s compatible with the VC asset class. Not us. And we’re not happy quietly doing a Zebra-style startup by ourselves, either. We want to inspire other founders and other investors to consider startups like SparkToro—ones focused on profitability, long-term survival, and dividends.
I hope this payback is far from the last capital we deliver to the people who believed in us. And I hope it also inspires many folks who read this to tell their entrepreneur and investor friends that there’s another way. For example… this amazing group of indie founders I’ve been dragging to Bologna the last two years:
Like us, they didn’t raise venture. Like us, they prioritize things in their life beyond just their businesses. Like us, they care about profits > growth, customers & employees > investors, and work enabling wonderful lives > living only to get big or die trying. You’re not alone, indie founders (despite the media narrative that we don’t exist 😉).
Lessons About Growth
Every year, I learn new things, moderate my positions on old things, and add gray to the black-and-white opinions that dominate the worlds of marketing and growth. This past year’s lessons are a wide, disparate array, but worth sharing:
- The Long, Dark Teatime of the Self-Service Sales Cycle: In businesses like ours, most top-of-funnel marketing happens months or years before conversions do. When someone buys SparkToro, we have no way to attribute it to the three videos they watched on LinkedIn or the word-of-mouth recommendation from an ex-colleague at their previous agency, or the podcast they heard Amanda on last month. This would drive a lot of CMOs and CFOs bananas, but if you can lean into the process of trusting your “vanity metrics” (views, likes, comments, shares, emails, I-heard-about-you-ons), you can build a marketing flywheel that’s almost entirely devoid of competition.
- There’s Something Weird in the Macro-Economic Neighborhood: I ran a very similar business to SparkToro in 2008/09 when the macro economy was, supposedly, melting down to a far greater degree than in the last few years. But, the previous 12 months of “When is the recession coming? Is it here? What’s going on?” has been, oddly, worse. I think it’s the anxiety of not knowing. In 2008/9/10, everyone could point to the meltdown and recovery trends as they happened, and everyone else could nod and say, “yup, same here.” For the last year, no one seems to be on the same page about where the economy is broadly, and investors, businesses, business owners, C-suite, agencies, consultants, and service providers are all caught between lackluster-to-mediocre growth and fear of the bottom dropping out. It’s not financially worse than a true recession, but it feels, mentally and emotionally, worse than a known recession. Everyone’s in their numbers asking: “is it something we’re doing?”
- Some Features Sell and Some Features Gel: We’ve seen a surprising lack of correlation between adding new things to SparkToro that our audience and potential customers seem to get excited about (and produce upticks in conversion rate around launch) vs. those that quietly get used the most in the product. Occasionally, a feature does both. But, more often, it’s one or the other. Oddly enough, I don’t regret investing in both kinds—if feature X brings you to SparkToro, but feature Y is the one you use every month, that’s fine by me.
- Industry Averages are a Godsend: Casey, Amanda, and I were recently on a call with Asia Orangio of DemandMaven (she’s helping us with customer and stats research for the upcoming V2 of SparkToro), and Casey shared that ~1,200 sign up for new, free SparkToro accounts each week, but only ~2.3% become paying customers. “There’s gotta be a way to help those other 1,172 see more value, right?!” he grumbled, and I agreed.
Then Asia told us that a 2-3% conversion rate on freemium SaaS is very much the industry standard. And while we could maybe bump that a few tenths-of-a-percent, expecting to increase it by 50 or 100% is unrealistic. Not only did that information help us feel better about our conversion funnel, pricing, packaging, and homepage changes, it helped us put the focus in the areas it belongs: top of funnel, retention, upgrades, and reactivations. If you can get industry averages and bird’s-eye-views from consultants who’ve worked with dozens or hundreds of folks in your field, take ’em. - Zero-Click Everything is Here to Stay: Big Tech’s growth issues are real, and they’re going to use AI, dirty tricks, and every bit of technical leverage and monopoly power they can to make sure to keep people on their platforms, bribe us into making them free content, and discouraging any outlinks to the open web. Marketers, if you’re not already planning for how to benefit from zero-click searches in Google, and zero-click content on LinkedIn, Facebook, Instagram, TikTok, YouTube, Twitter (well, Twitter you probably don’t need to stress since the owner seems to have a death wish for the site), and Reddit (maybe same here?), you’re in trouble. These platforms have proven that the way to play is native content, delivered for maximum exposure/consumption/engagement and entirely devoid of attribution.
- Word of Mouth > Everything Else: Nine times out of ten, when your marketing team is seeing positive top-of-funnel numbers, but struggling to get growth, it’s because the product’s not resonating with customers. Customers talk to each other. It’s not just public reviews, social media/dark social, and trackable sales funnels that drive behavior. My emerging theory is akin to the iceberg analogy: what’s visible on the web and in your analytics is just the tip; word-of-mouth is the gargantuan segment underneath.
Great products aren’t enough, either. To be “great” is, in my opinion, not nearly as valuable as being irrelevant to 99% of people, but exactly perfect for the 1% who deeply care about the problem you solve. Extra bonus points: target your product at a group that’s well-connected to others in their field, and gets value from sharing new things. Nothing’s better than word of mouth marketing. Nothing. - Hype Train’s Gonna Hype: Generative AI and LLMs are cool. They can be quite useful for certain tasks (Britney Muller summarized many of these applications in her recent Office Hours). And it’s a huge relief to see tech folks focused on something other than blockchain and NFTs. But, much like the insanely-hyped rise of mobile apps (~2009-2012), a few hundred companies and products will build phenomenal value here while a few million others shouldn’t be investing nearly as much as they are.
I think about it this way: if you have real customer problems you were already working to solve, and AI is the obvious, best solution (plus you actually understand what it does and how it works), use it! But if you’re searching for problems AI might address, just so you can get executive/investor buy-in and ride the hype train… re-think those priors.
I’ll undoubtedly have new learnings next year.
Lessons About Chill Work
Chill work is our professional philosophy, but that doesn’t mean it’s a religion. It’s not set in stone. It evolves as we learn more about how best to practice and integrate it in our lives. Broadly, we have beliefs:
- The most important thing we can do is make great decisions about the design of our business, our product, our marketing, and the problems we choose to solve (and don’t).
- The best way to make great decisions is to be well-rested, healthy, happy, secure, and patient.
- More work is rarely useful compared to the right work. A hundred more blog posts or a dozen new features pale in comparison to the right five posts and that one truly valuable feature.
Applying those beliefs is different for each of us. That’s part of the philosophy: what works for me doesn’t necessarily work for Casey or Amanda. For example, I love working late nights. I hate early mornings. And I find myself pretty unproductive between 3-6pm. So, I work a couple extra hours at night, do a few hours most weekends, and use my afternoons for long walks and elaborate meal prep. Amanda, on the other hand, keeps weekends almost entirely work-free. Casey’s more of a mix. Since we’re entirely remote, and only have a meeting every few weeks, this works great.
Chill work ebbs and flows, too. February, March, and early April were incredibly stressful for me this year, but Casey’s quiet confidence that we’d find a way through the Twitter issues (below) helped me stay grounded… mostly. The last few weeks, I’ve been doing more of the re-assuring as Casey analyzes our hopeful new data sources and ran into numerous road bumps (with something of a breakthrough this past week).
I think that sometimes, we portray SparkToro as too chill.
Compared to 95% of tech startups, we are. But that doesn’t mean we don’t have high tension days, occasional weeks of existential dread, and familiarity with the emotional rollercoaster that comes with startup life. Our first few years of existence have been remarkably smooth, but even we have restless nights spent worrying about the future. My goal has been to use those times to empathize with fellow entrepreneurs, and remember to prioritize sleep and health—the ingredients most likely to help me make good decisions.
The extended SparkToro team: Rand, Casey (and his girls), Lindsay, Geraldine, and Amanda
As I write this, I’m feeling optimistic and energized. The problems we’re struggling against are fixable. The market we serve is resilient and growing. The way we provide value is going to get better. And the big, thorny problem that’s been weighing on us for a year… is only a few months away from being gone forever.
Lessons About Relying on an External Platform
Before we launched SparkToro as a paid tool, we built three free tools, all centered on Twitter: Fake Followers, SparkScore, and Trending. Each of them had thousands of users and helped both our traffic and brand reputation. Fake followers, and its underlying methodology was so well-liked that when Elon Musk tried to get out of buying Twitter last summer, his people contacted us about leveraging a more sophisticated version of the methodology behind it.
I’ll save that story for the sequel to Lost & Founder, but the TL;DR is that we didn’t work together, Musk bought Twitter, and he then made it impossible for us to operate any of our free tools. Simultaneously, his changes at the company had two other impacts:
- The APIs no longer enabled the type of access required for Audience Tracking, a feature we’d been building for the prior year. Even if we wanted to pay hundreds of thousands each month, it simply wasn’t possible.
- The activity on Twitter shifted dramatically—in many professional and interest spaces, it declined, while conservative politics, hate speech, harassing behavior, and meta conversation about Twitter itself increased.
SparkToro has long used Twitter as our primary connector network, i.e. finding web and social connections starting from Twitter and then branching out to 13 other networks to build an index of the 86M entities we search, aggregate, and anonymize. But, with the new owner’s changes, that became both more challenging (no APIs for some things, ludicrously expensive APIs for others) and less valuable (worse data coverage, more hateful accounts that don’t add value for our customers, and more meta noise from people who used to share, follow, and engage with other topics).
For years, we’d been friendly with folks on Twitter’s API team, trust and safety team, and even the partnerships team. But, within 60 days of the new owner’s takeover, everyone we knew there was fired or left. There was no one to contact as, little by little, Twitter’s features and value diminished.
Thankfully, we’re far less affected than many other Twitter data products. SparkToro continues to run on the other 13 other networks we index, and we’ve still got access to enough data to make our index functional and valuable. Amazingly, apart from closing our free tools and Audience Tracking, we haven’t had a customer-facing impact to the product.
But, Twitter cannot be our future. Its days as a platform for valuable audience research are over. Even if the APIs still functioned, were still free, and still enabled the kinds of requests we used to make, Twitter’s activity graph has become more like Parler’s or Truth Social’s. Pre-Elon Twitter was already a bit culture-wars and politics-centric. Since his takeover, those subjects are almost all the platform offers. That’s not what our customers care about mining for marketing insights. We have to move on.
The Future of SparkToro and Our Upcoming V2
Since Q3 of last year, we’ve been planning for a move away from Twitter as our central connector network to a network-agnostic infrastructure. But, those plans have been theoretical, exploratory, and un-tested until the last 90 days. Thanks to some data partnerships, new data sources, and a lot of elbow grease by my co-founder, Casey, we believe we’ve got a promising architecture that offers many of the same features and benefits folks have received from SparkToro over the past 3 years, and a lot of exciting new ones we’ve never before been able to offer.
How will SparkToro V2 Work?
The old version works like this:
- Index every real, useful-looking Twitter profile
- Crawl, search, and match profiles across 13 other networks (and the open web)
- Build an index of searchable profiles
- When someone searches, anonymize and aggregate matching profiles to return relevant data
The new version works in a network-agnostic fashion:
- Start with websites and keywords (from our existing index, plus clickstream and SERP data sources)
- Crawl profiles associated with those websites on any available network (expanding from 13 to ~20)
- Add profiles with no direct domain associations (these can still provide value when topics/keywords/behaviors/demographics are called for)
- Build an index of network-agnostic entities (i.e. a profile doesn’t need to exist on any given network or have a specific association to be included in our system)
- When someone searches, anonymize and aggregate matching profiles to return relevant data
It’s not a huge philosophical change. SparkToro continues to be a “show me the behaviors and demographics of people with characteristic X” tool. But, it gives us a far less brittle, single-point-of-failure architecture to aggregate public information and make it easily searchable for actionable insights.
In the short-term, we’re also going to test a simpler, more streamlined version of SparkToro’s search interface. The current drop-down doesn’t get nearly enough use—almost everyone tries “My audience frequently talks about: __________________” but…. they try it with SEO-style keywords! That’s not how it works!
Well, that wasn’t how it worked.
Instead of fighting the tide and trying to re-educate 2,000+ folks who use SparkToro each week, we’re making the way you think SparkToro works into the way it does work. Analyze audiences based on search keywords and websites. We’ll see how it goes, and adjust from there.
It also lets us provide unique, useful data available nowhere else.
Above: an early wireframe for SparkToro V2
The extensibility of the architecture is what excites me most. No more “hey Casey, can we build feature X?” followed by “Well, Rand, not unless you can think of a way to tie that to Twitter accounts.“
What New Kinds of Data Could Be Available?
So. Much. Potentially. Cool. Stuff.
- Search keywords – our most requested data feature has long been “show me the keywords this audience searches for!” and now, we can finally provide that.
- Most-Used social networks – a lot of free users (and a few paid customers) have hoped that we can show which social networks are more or less used by the audience(s) they care about. That’s something we can finally do: providing high-quality numbers around your audience’s use of LinkedIn or TikTok or Github vs. the average web user.
- Popular apps & productivity tools – Does your audience use Canva, Adobe, OpenAI, Spotify, or dozens of other popular web apps or tools more/less than the average web user? For persona builders, especially in certain B2B and B2C niches, we suspect this can be incredibly valuable information that SparkToro will be uniquely able to provide.
- Brand affinities – Not every brand that’s high-affinity among a particular audience is necessarily a competitor, but we suspect a lot of folks can get value out of knowing which brands (in their space or beyond it) are attracting their customers and potential customers.
- News stories, popular articles, trending topics & keywords – By looking at the stories getting attention in places like Google News, Google Discover, Twitter, Mastodon, and other social networks, we may be able to ID both individual URLs and story topics/news-relevant keywords for any given audience. Not quite what Buzzsumo does. Not quite what Newzdash offers, but something uniquely SparkToro-esque (akin to what we briefly planned with Audience Tracking).
- Email newsletters – We got pretty close to launching this in SparkToro before February’s Twitter changes put our attention elsewhere, but Casey has some good ideas and pre-built infrastructure ready to manage this one. I think it could be huge, especially for those spaces where email subscriptions drive a ton of awareness and action.
- Events – Another near-launch, this one from 2021. Derailed by the complexities of identifying “events” on the web (and the cancellation of so many during the pandemic), we suspect it’s possible again in the future.
I can’t promise that all of these features will be available in the near-term, but all of them should be possible in the long-term. If you’re passionate about one or more of these, please drop me a line ([email protected]); I’d love to have your input on prioritizing what we include and understanding how the data will help you!
Thanks, as always, for following our journey and supporting an alternative to the Unicorns-or-GTFO startup world. Year three is a big milestone for us, but I hope its just the beginning of a multi-decade business that provides amazing audience research while also delivering great lives to the team, positive returns to the investors, and a model that other companies and investors choose to adopt.
Amanda, Rand, & Casey, goofing around during a SparkToro V2 planning meeting
Stay tuned for more on V2, Chill Work, Zebra startups, and all things audience research.