FriendlyData: Six Employees, $1.2M Raised — Sold to ServiceNow for 8 Figures the Month After Closing Its Seed Round

All dollar amounts are given with a rough yen conversion at ¥150/$1.

A company whose story ended in under three years

FriendlyData was an enterprise software company founded in 2016 by three people: Michael Rumiantsau, Alex Zaytsev, and Alexey Zenovich. It did one thing: convert plain-English questions into SQL so people could query a database directly. The goal was to let business staff who couldn’t write SQL touch their own company’s data without going through an engineer, the Text-to-SQL space.

In October 2018, the company was acquired by the publicly traded ServiceNow (Santa Clara, California). According to They Got Acquired, the sale price was in the “low 8 figures”, roughly the low tens of millions of dollars, or a few billion yen. The company was under three years old, had six employees, and had raised a total of $1.2 million (about ¥180 million). At the time of the sale, it had “only just started landing enterprise customers.”

How did a valuation that revenue and customer count can’t explain come about? Let’s lay out the timeline first.

2016 → 2018 in sequence

What stands out is the density of the last three rows. The company sold the month after closing its seed round. Normally, closing a seed round is a declaration of “we’re going to grow this over the next several years”, a sale immediately after signals an entirely different decision.

What they were building

Rumiantsau is a software engineer from Belarus who had served as CTO of a local startup called Flatlogic. He moved to San Francisco in 2017 and initially ran a contract development shop alongside FriendlyData.

There’s one decisive fact about the product: it beat every known Text-to-SQL benchmark. Rumiantsau’s own explanation of the acquisition rationale has been preserved:

ServiceNow was interested in building the world’s most advanced NLP (natural language processing) capability. FriendlyData’s technology outperformed every known Text-to-SQL benchmark. (paraphrased)

What was for sale, rather than revenue, was technical superiority you could measure and compare.

The ten months everything changed

The turning point is concentrated in 2018. For the two years before that, FriendlyData had been “a product its founders ran alongside contract work.” That changed with acceptance into 500 Startups Batch 20. Rumiantsau has called that acceptance “a turning point in my life as a founder,” explaining why:

As an immigrant, having social proof opened a lot of doors for me. (paraphrased)

After the acceptance came a seed close in September 2018, followed immediately by ServiceNow’s outreach, and the deal closed in October. Laid side by side, the shift in trajectory is obvious.

The technology itself had been under continuous development since 2016. It didn’t suddenly get better in 2018. What changed was that the technology became visible.

Why did a company with almost no revenue command tens of millions of dollars?

Three structural factors overlapped.

The verification cost was low. When a buyer evaluates a private startup, the most expensive question is “is this technology actually good?” Text-to-SQL has a public benchmark, a shared ruler, and FriendlyData was at the top of it. Even without a revenue track record, a buyer could confirm the edge in a few hours. Conversely, in a field without such a ruler, the same technical strength wouldn’t fetch this kind of price.

The buyer, meanwhile, was making a strategic acquisition, not a financial one. ServiceNow was investing heavily in AI and machine learning at the time, with a stated goal of “the world’s best NLP capability.” In this scenario, the pricing benchmark shifts from “some multiple of this company’s profit” to “the cost and time of building the same thing in-house.” For a public company, a one-year delay in opportunity can easily exceed tens of millions of dollars. That’s why the price held even with the seller’s revenue near zero.

Add the fact that the whole team moved over. All three co-founders joined ServiceNow and worked on integrating the technology into the parent company’s platform. The buyer got more than the code. The people who could keep it running came with it. This deal carries a strong flavor of what’s called an acqui-hire.

A second point: Rumiantsau’s view on sales. He has said “your first sale should happen before the MVP is finished,” and he actually started reaching out to companies from the prototype stage. His reasoning: “feedback from people actually using it is the best kind.” As a result, by the time of the acquisition, FriendlyData had already reached the state of “just starting to land enterprise customers.” The gap between zero and “just starting” looks large from a buyer’s perspective.

Three versus seventy — an asymmetry

For the most visceral number in this story, skip the sale price and look at the headcount involved in due diligence.

Rumiantsau has recalled that “selling to a public company was hard because of the sheer volume of scrutiny and due diligence you have to pass.” When a small team sells to a public company, negotiation and document requests consume roughly as much effort as running the business itself. The sale process becomes, for a six-person company, a major project in its own right.

The lessons he later shared on social media assumed this same asymmetry: negotiate with multiple prospective buyers, secure an executive sponsor if you’re selling to a large organization, and have a fallback plan ready in case the deal falls through. If the business stalls while a three-person team handles DD and the deal doesn’t close, the company can suffer a fatal blow.

What this story doesn’t tell us

To be candid, there are a lot of blanks here. The sale price is given only as a range (“low 8 figures”). The exact amount is undisclosed. Neither ARR nor customer count was released. So you can’t learn “how much revenue gets you what price” from this case.

Further, deciding to sell the month after closing a seed round is also a decision to walk away from the premise of that fundraise. A sale right after closing is a good deal for investors, a short return horizon, but there’s no way to verify how big the founders could have grown the company over the following years. A sale is always a choice to “step off at this point,” and this story doesn’t disclose the cost of moving that fast.

What’s reproducible and what isn’t

Two things are reproducible. One is starting conversations with prospects before the MVP is done. The other is a design philosophy: choose a field where an external yardstick exists. Benchmarks, public scores, third-party certifications, any indicator that lets others confirm your edge quickly is a field where you can be valued even with a thin track record.

What isn’t reproducible is bigger. The timing of 2018 itself, for one. Text-to-SQL back then was a specialized machine-learning problem, and topping the benchmark had real value. Today, foundation models widely offer similar capability, and it’s hard to imagine the same technology earning the same recognition now. For another, acceptance into 500 Startups Batch 20 itself, not something you get just by applying, and in his case it worked in the unusual way of substituting for credibility checks as an immigrant founder. And finally, the buyer came to them. ServiceNow moved first. This wasn’t a deal the seller went hunting for, and a buyer showing up unprompted is not a condition you can engineer on purpose.

Rumiantsau later left ServiceNow, founded the AI analytics startup Narrative BI, and now invests in other entrepreneurs through an investment platform called @founders.ai. The exit wasn’t the end. It became capital and credibility for the next round.

Related reading

• ScrapingBee — how a small SaaS built the numbers that led to a sale
• MENTA — a Japanese example of an acqui-hire by an operating company