Lyle Stevens and the arithmetic of being believed

Lyle Stevens had an unusual way of settling an argument between his younger brothers. They were debating who was more popular on Facebook. Their older brother, already accustomed to working with social data, decided that family opinion could use some numerical assistance. He examined their posts and the responses those posts attracted. Thanksgiving had acquired an analytics department.

The episode is funny because the ambition was so small. Nobody needed an enterprise software company to establish bragging rights between siblings. Yet Stevens returned to Boston with a question that had escaped the family gathering: if he could identify influence among people he knew, could a business find it among its customers?

That question became part of the origin of Mavrck. It also supplies a useful way into Stevens himself. Presented with a social mystery, he wanted to inspect the evidence. The attraction was never confined to the person making the most noise. He wanted to know whose words made something happen.

A business student with a second set of speakers

Stevens came to Northeastern University from upstate New York. In an early conference presentation, he described his upbringing as taking place in corn country. His family included serial entrepreneurs; the idea of running a business arrived before the idea for a business did. He graduated in 2009 with a business administration degree.

His college years included another kind of audience work. Stevens helped pay his way through school as a DJ. There is a pleasing connection between that job and the company he eventually built: both involve paying attention to how people respond. A dance floor, of course, delivers its performance review with rather less paperwork.

He also held leadership roles in his fraternity. Looking back in 2026, he described university life as a place to “practice being a founder.” Running an organization gave him somewhere to apply classroom lessons before a company depended on them. His later advice to students followed the same practical line: find a campus organization or a startup where you can learn by doing.

Before the influencers, the subject-matter experts

His route into marketing passed through Raytheon. Stevens worked on cybersecurity and internal technology, including a social network that allowed employees to collaborate within the company. The problem was organizational: how could people inside a large enterprise discover colleagues who knew something useful?

An internal social network leaves clues. People discuss particular subjects; colleagues respond; knowledge circulates. Stevens worked on a way to identify subject-matter experts from those interactions. The job involved studying relationships between people and their contributions, the same broad territory he would later explore in consumer marketing.

“I loved the product,” he recalled. It is a small statement, but it helps explain the transition. The bridge from aerospace to Facebook was a working interest in what networks could reveal. One environment contained employees with specialist knowledge. Another contained customers whose recommendations carried weight with their friends. The Thanksgiving experiment brought the second possibility into focus. Once back in Boston, he discussed it with Sean Naegeli, who became a co-founder.

The first cheque was considerably smaller

The venture initially operated as SplashScore, a Facebook application that measured social influence and connected influential users with brands. It was already in development before the Mavrck name appeared. The distinction matters: a company's launch date rarely captures all the evenings that preceded it.

Stevens returned to Northeastern's entrepreneurial community after graduating. Its student-run accelerator, IDEA, supplied SplashScore with $25,000 in gap funding, including $10,000 to launch and build the application. In 2014, he pointed out that the resource had remained available to him as an alumnus. His university connection had become a working relationship rather than an item on a résumé.

The company went through Techstars Boston in 2014 and emerged under the Mavrck name. Chris Wolfel joined Stevens and Naegeli among the founding team. For all the industry language that would accumulate around the business, its early proposition was easy to recognize: a brand's existing customers might include people worth listening to, and software could help find them.

By November 2015, Mavrck had raised a $5 million Series A led by Kepha Partners. Jo Tango's account of that investment stretched back to meetings with Stevens in 2013, a reconnection through Northeastern mentor Bob Lentz, and the Techstars period. The relationship developed across several versions of the venture. That history puts some useful texture around the word funding. An announcement arrives on a particular day; the conversations, introductions and judgments behind it can take years. Stevens's network included people who had watched the work develop, rather than encountering a finished pitch for the first time.

The awkward business of counting people

An influence score creates an obvious temptation: make the score bigger. By 2018, fake followers and manufactured engagement were a live concern for the business Stevens had entered. Mavrck's work included fraud detection intended to distinguish real audiences from inflated numbers.

“Influencers are a whole new kind of currency,” Stevens said that year. Currency invites counterfeiting. Once attention can earn money, somebody will try to manufacture a convincing imitation. The challenge for an influence platform is therefore partly one of selection: what deserves to count, and what merely looks impressive on a screen?

That challenge gives his career a consistent thread. The Raytheon project involved discovering useful expertise inside a large network. Mavrck involved finding useful advocates and creators inside much larger networks. In each setting, the visible hierarchy offered only part of the answer. A person with a small, attentive audience could matter for reasons a raw follower total would miss. The software had to make room for those differences.

Fifty-eight filters, one very human inbox

The public version of Stevens contains plenty of spreadsheets, but also some reassuringly ordinary tastes. He has described himself as a dog lover and an audiophile. His early professional biography mentions music production as well as DJing. Sound has evidently remained an interest beyond the college job.

His approach to work can be more elaborate. Asked about productivity in 2018, he described checking email in three scheduled sessions and using 58 Gmail filters to organize the traffic. Team members could mark a subject line “Action Required” when something needed his attention within a day or two. An urgent matter warranted a call.

Fifty-eight filters is a wonderfully specific answer. It suggests someone willing to build a system around a repeated nuisance. An investor, Jo Tango, who had followed Stevens since the SplashScore period, described his integrity, drive and leadership when announcing an investment in 2015. Together, these details show two aspects of the job: organize the information, and earn the confidence of the people bringing it to you.

“Data.”

Lyle Stevens, asked for one word describing how he works, 2018

When the customer stood on both sides of the deal

In December 2021, Mavrck announced a $120 million investment from Summit Partners. The stated plan included product expansion and recommendation technology to help marketers decide which creators to work with and how to structure those relationships. It was a larger version of the original analytical problem, with commercial decisions attached.

Four months later came another $135 million investment from Summit and the acquisition of Later. The funding figure described the investment; it was not an announced price for the acquisition. Keeping those numbers straight is part of telling the story honestly.

The combination brought together different working lives. Mavrck served enterprise marketers managing creator relationships. Later supplied social publishing tools used by creators and small businesses, along with its link-in-bio product. Stevens was now overseeing a business that could see both a brand's campaign and a creator's everyday publishing needs. The commercial question widened: how could better measurement for a marketer also help a creator earn a living?

A new name, then a different chair

In January 2024, the two businesses united under the Later brand. Their products acquired names that explained their jobs: Later Social, Later Influence and Later Link in Bio. A brand manager planning a campaign and a creator scheduling a post were now working within the same broader company.

The following month, Stevens moved from chief executive to chief strategy officer. Scott Sutton, previously the chief operating officer, became CEO. Stevens remained on the board and executive leadership team, with responsibilities covering innovation, partnerships, corporate development and future planning.

A founder's title can become so closely associated with a company that a change looks like a full stop. The actual division of responsibilities here points to continuing work. Stevens retained a role in choosing where the business should go, while Sutton took the chief executive position. Later's current leadership listing still names Stevens as its co-founder and chief strategy officer. His connection to the business continued through a different job description.

The next question is what the machine cannot prove

Stevens's writing in 2026 returns to trust, with a new complication. His January predictions for the creator economy considered AI-generated work, content authentication, audience ownership and more disciplined measurement. He anticipated creators becoming more structured businesses and brands asking harder questions about what their spending actually changes. Those were predictions, rather than a report that every change had happened.

By September, he was discussing how AI had lowered barriers to creative expression and made online content creation accessible to more people. It is a development with two sides for a business built around influence. More people can produce material. Audiences and advertisers still need reasons to believe the person or message in front of them.

That is a recognizable continuation of the Thanksgiving problem, even as the scale and tools change. Stevens began by asking which brother could get a response. He now works in a market where generating a response and earning trust can require separate explanations. The useful answer still depends on looking closely at what people actually do.