$2,500 × 10,000 Customers Leaves $75M of SmartBear Unexplained

Nathan Latka did the multiplication out loud and it came up three quarters short. SmartBear had about 10,000 customers. Its average transaction, CEO Justin Teague had just said, was about $2,500 a year. “if I take 10,000 customers times that $2,500 ACV average you just gave me, obviously that’s 25 million annually and missing 75 million bucks of revenue — where is that?” Teague did not dispute the gap. He said there were two answers, and between them they describe how this company was actually assembled.

When this conversation happened

The tape dates itself twice. Reaching for a market benchmark, Latka says SendGrid “just went public” — SendGrid listed on the New York Stock Exchange on 16 November 2017. Pressing Teague on whether the year would close above the round number, he notes there are “about 35, 40 more days left in the year,” which puts the recording in the back half of November, and his sign-off says “as we wrap up 2017.”

For once the database agrees with the content. GetLatka’s capture stamp for this interview reads 16 November 2017, the episode is numbered 960, and the dated revenue, headcount and churn rows all hang off the same day. The one company detail that does not line up is older than the tape: Teague puts the founding at 2003, an engineer building a code collaboration product so developers could keep track of who changed what, and Latka repeats 2003 in his summary. SmartBear’s GetLatka profile carries 2009 as the year founded. On a fact its guest lived through, the tape is the better source.

• 2003 · Founded an engineer builds a code collaboration tool to fix a problem other software engineers had.
• 2015 · Teague arrives brought in as COO by an investor who liked the try-and-buy concept and wanted it made scalable, then moved into the CEO seat.
• Spring 2017 · Francisco Partners the private equity firm buys the company, in what Teague calls building SmartBear 2.0.
• Nov 2017 · This interview about 10,000 customers and roughly 400 staff across six sites.

The round number was still a target

The milestone this URL promises is not in the conversation. Asked for the top line, Teague says SmartBear is “approaching about a hundred million dollars.” Latka asks whether he will break it with 35 or 40 days left; Teague’s answer is that the target for the year is “just under,” and that they will “definitely break it next year.” The episode itself went out under a headline of $90 million in ARR.

GetLatka’s row is more precise than either. The revenue recorded against the date of this interview is $99,999,996 — four dollars short of nine figures, a database doing exactly what Teague did and refusing to round up. The next dated revenue row for SmartBear is $139 million, recorded December 2020, three years later. That is where the crossing is on the record. It is not on the record here.

The growth rate underneath it is the part Teague gives freely. Twelve months earlier the company was at $78 to $79 million — Latka proposes the range, Teague confirms it. Organically SmartBear “had been growing in the mid-teens”; with acquisitions counted, “we’re now in the mid 20s.”

$2,500 buys the door, not the customer

The entire business runs off what Teague calls a trial-and-buy concept, open source tools included: SmartBear owns and develops two of the leading ones in its market, Swagger and SoapUI, and they sit at the top of the same funnel. Products cost anywhere from a hundred dollars to a couple of thousand. Some are pure cloud subscriptions billed monthly with spike usage on top — the cross-browser testing product lets a team skip the phone lab and pay for the configurations it uses — and others are traditional annual contracts. Twenty million downloads a year produce about 400,000 leads, people raising a hand to say they are trialling and would like to buy, and about 7,000 of them become customers.

So the $2,500 is real, and it is also the smallest number in the business. Latka works the arithmetic the other way — divide the annual figure by twelve, then by 10,000 customers — and lands on accounts paying “much closer to like eight or nine or ten grand a year versus twenty five hundred.” Teague agrees, and supplies the distinction that resolves it. The $2,500 is “the average first-time purchase,” the year-one contract value of a brand new logo, not what an installed account pays.

“we have developer tools, so five developers writing the next mobile application download and like it, it spreads, and we might wake up one day, you know, two years later, and find out that there’s 200 users in the same account.”

That is expansion described as a mechanism rather than a target, and it is what puts net retention in the 115 to 120 percent range Teague gives. Asked to credit more of the growth to new customers or to expansion, he splits it evenly: about 50/50. The new half he attributes to the API wave — Swagger and SoapUI put SmartBear in front of every team trying to build, deploy and test APIs, and he describes the company as riding that rather than creating it.

One figure in the funnel does not sit with the others. Adding about 7,000 customers a year to a base of about 10,000 implies either a customer count climbing far faster than the revenue line, or a large share of new logos gone inside twelve months — and the retention figures Teague gives a few minutes later describe neither. The likeliest reading is that the 7,000 counts first-time transactions rather than surviving accounts, which would be consistent with the $2,500 being an average first purchase. The tape never reconciles it, and neither number was walked back.

The acquisition engine

Half the growth was bought

The rest of the story never appears in a funnel chart. SmartBear spent years acquiring small, founder-led tools companies and running them on as innovation centers — six locations globally, with many of those founders still in the building. The example Teague reaches for reframes the whole revenue line: “one of them for example was as small as a million dollars — it’s now about forty five million dollars of our revenue.”

$45Mrevenue attributed to a company SmartBear bought when it was doing about $1 million

“we’re sort of, I would say, entrepreneur friendly, right? You can build a million dollar company, get a payout to us, and then become part of the company where we have innovation centers that are still driven by our founder led owners.”

Justin Teague, CEO, SmartBear

What SmartBear did not do is fund that with venture money. Asked how much the company itself had raised before the private equity sale, Teague does not have the figure to hand and estimates the “million-and-a-half range,” adding that SmartBear was in the couple-million-dollar range when it was first put together with another company. Each of the acquired companies, he says, had its own self-funding or bootstrapped model, and a previous investor put money behind the roll-up. GetLatka’s funding file disagrees and the tape cannot settle it: the record carries $25 million in total across two rows dated March and October 2020, nearly three years after this conversation and well inside private equity ownership. Neither corresponds to anything Teague describes.

Retention measured from the floor

Teague tracks the number that flatters least. Net retention he calls the easy metric, the one with add-ons already baked in. “my favorite metric is gross dollar churn”: start the year with ten customers paying ten dollars each, end the year, and count how much of that original ten dollars is still being paid, ignoring everything they bought since.

Over the prior twelve months that number sat “right around high eighties” measured as retention, which Latka rounds to a little over 80% in his summary and which GetLatka records from the opposite direction as 16% gross churn. Either way, something close to a sixth of the starting revenue base walks each year, and expansion has to cover that before any of it counts as growth. Latka at one point restates the combined picture as about 150 percent and gets a “you got it” back; the revenue churn and net retention figures Teague himself gave are the high eighties and 115 to 120, and those are the ones to use.

Lifetime value gets the same treatment, which is to say he refuses to reduce it to one number. Contract sizes are too far apart, so SmartBear cuts it by logo, by product and by total spend, and Teague names the complication honestly: a mobile project ends, its licences migrate to a new project, and whether that is a loss depends entirely on which view you take. What he watches is the trend rather than the industry benchmark.

$1.25 to buy a dollar

SmartBear does not track the cost of a customer at all. “we track CAC on a dollar basis rather than trying to acquire just a customer,” Teague says — what does it take to acquire a dollar of new ARR — “and our metric is we try to stay below a dollar 25.”

CAC ratio = acquisition spend ÷ new ARR acquiredTeague’s ceiling is $1.25 of spend per $1 of new ARR, which Latka reads back as a payback period a little over a year. Teague: “that’s right.”

The spend behind it is about $750,000 a quarter, excluding headcount, and it is not mostly bought clicks. Teague describes a big inbound demand engine built on content, because a high-velocity try-and-buy motion “demands us getting people to come to us.” Against CAC benchmarks, $1.25 per dollar of ARR is the figure to compare. GetLatka’s record also carries a $12,500 customer acquisition cost for SmartBear, which is not a number Teague states anywhere on the tape and does not reconcile with a $2,500 first transaction.

The IPO question he has not had to answer

Latka builds the ladder: SendGrid just went public, markets feel good, and by his reading of other CEOs a company starts thinking seriously about listing somewhere in the $110 to $130 million ARR range. Teague declines the frame. “we try not to focus on an outcome as much as building just a great company,” he says, and there is a dangerous trap in believing you can architect your outcome. Pushed harder — if the same $200 million could come from a private investor or from public markets, which cost of capital is cheaper? — he does not pretend to have modelled it: “I haven’t had to make that decision yet.”

The company he was running at the time was about to hit 400 employees globally across six innovation sites, the largest just outside Boston — the headquarters sits in Somerville, Massachusetts — plus Galway and an offsite center in Tula, Russia. GetLatka’s dated headcount rows carry it forward: 433 in December 2018, 489 in December 2019, 553 in June 2020, 535 in December 2020, and 826 by July 2023. The sales column on those same rows runs 109, 113, 126 and 129 through December 2020 — a company at that revenue scale with fewer than 130 people in sales, which is what a trial-and-buy funnel is for.

Ben Franklin and the everything store

The closing round is quick. Favorite business book: a biography of Ben Franklin, because innovators turn up in every era and this one invented the library. What he is actually reading is The Everything Store, the Amazon book, from which he takes a single lesson — a maniacal focus on the customer, and how much closer to it SmartBear is trying to get. Favorite tool that is not his own: Slack, for a company spread across six sites, and Flipboard on his phone. Close to eight hours of sleep. Married, three boys, 45 years old.

Then the question Latka ends every episode on, and the answer has nothing to do with SmartBear.

“I wish I had taken more risk, right, because the time to take it is when you’re young and the implications are not quite as, let’s say, dramatic. But I accumulated a lot of experience going overseas, living in London, of different types of software companies, and I think accumulate as much experience as you can early because you can always leverage it later.”

Justin Teague, CEO, SmartBear

Sources Nathan Latka’s November 2017 interview with SmartBear CEO Justin Teague, episode 960; GetLatka’s dated revenue, headcount, churn, customer-acquisition and funding records for SmartBear; SendGrid’s New York Stock Exchange listing on 16 November 2017.