Chorus Turned On Pricing in 2017, Hit $8.4M, Then Passed $10M

Roy Raanani was explaining why Chorus.ai’s customers sold it for him, and the illustration he picked was a sentence a sales rep says out loud, to a prospect, in the first minute of a call.

“Oh, Nathan, yeah, just so you know, there’s another participant here. It’s Chorus. It’s a product that’s going to take notes for me so that I can be more focused on our conversation.”

Roy Raanani, co-founder and CEO, Chorus.ai

Dating the tape

This conversation happened in the autumn of 2017, and the internal evidence is unusually tight. Raanani says Chorus launched in 2015 and that “the company is actually only two and a half years old.” He describes 2016 as a year of proof of concepts, 2017 as the first year the product had a price on it, and he answers a question about the $10M mark with “not by the end of this year” before naming 2018 as the year they would push. He had hired a head of marketing “a few weeks ago.” GetLatka’s own rows for Chorus put team size at 40 and revenue at $8.4M on 12 October 2017 — which is precisely the company on the tape, “a little over 40 people.” The video you can watch today was posted in July 2021, when Nathan Latka put it back up as a lookback.

The black box in the middle of every sales org

Raanani’s read on the category was that the first companies to attack calling had aimed at the wrong seat. They made it easy to record and to dial more, which serves the outbound prospector doing a hundred dials a day. The opportunity he saw — and he claims Chorus was among the first into it, starting about three years before the tape — was one level up.

That absence is the whole pitch. “In every other part of business you look at the fundamental drivers of performance, you measure them and you improve them, and when it comes to closing revenue and renewing revenue that most fundamental activity is a conversation, and there was no data on it.” Managers were forecasting against a black box.

Latka pushed for something the software could actually say out loud, and offered his own hypothetical: could you tell a CRO that the rep who always beats quota says the word “price” three minutes in while everyone else waits until minute 18? Raanani did not take that example, but he took the shape of it, and answered with discounting.

• It comes up constantly — across more than a million calls analysed by the time of the tape, discounting surfaces on roughly 25 to 30 per cent of conversations.
• Almost nobody trades for it — “everybody offers the discount but very few people get something in return for it.”
• Timing separates the reps — some offer it very early in the cycle; the experienced ones wait until they know there is a real deal and a real timeline, and spend it on something concrete.

The promise attached to that was specific and, for a 2017 machine-learning product, unusually falsifiable: within a week of switching Chorus on, the platform should surface moments worth more than what the entire company pays for it for the year.

What it costs, and why a ten-person team buys it

Asked for an average revenue per customer, Raanani gave the answer every enterprise seller gives: “we can’t, we don’t want to talk about pricing until we’ve looked at value.” What he would say is that Chorus priced itself as infrastructure — “competitive with any other core piece of the sales stack, right? So competitive with something like a Salesforce.com.”

Latka then built a hypothetical sized to his audience: a ten-person company at about $1M in annual run rate, two salespeople going to four, maybe raising a seed. A hundred dollars a month, a thousand, or ten thousand? Raanani took the bottom bracket. “It would be more around the 100 per month range.”

The interesting part is his account of why a company that small buys at all, because it has nothing to do with analytics. Every such team has one account executive who has been there two or three years, who knows the product cold, who knows how to position against competitors and how to create urgency. Buying Chorus is buying an insurance policy on that person — if they leave for a family emergency or the shiny startup around the block, the gold does not leave with them. Latka’s summary was that in the worst case you are auto-creating a training manual for scaling a sales team just by recording calls, and Raanani agreed that was the simplest version of it.

Moving up-market changed the reasons, not the mechanism. A deep bi-directional Salesforce integration, and the ability to push everything Chorus generates into a Redshift instance so a customer can build dashboards in Domo, Tableau, Periscope or Looker, are features that only matter to a large team — and the larger the team, the less visibility a leader has, so the value scales with the problem. The sharper claim is what it displaced: some customers told him they spent more time in Chorus than in Salesforce, because Salesforce is a lagging system that depends on a rep going in and updating fields according to their own interpretation of a meeting. In Chorus, a manager opens the deal in one click and sees what was and was not discussed, and whether it is really closing this month.

Latka’s response to that was to ask whether Chorus was in acquisition talks. Raanani did not blink: “if we had to sell we’d be really disappointed.” Latka told him he had clearly practised it.

The round that priced itself

Chorus was never going to be bootstrapped, and the reason Raanani gives is a sentence he says he used with his own team: there is “a billion dollar enterprise value company five to seven years from now that’s ours to lose,” because no company on earth was going to leave its customer conversations unrecorded and unread forever. The early decision that followed from it was to build the machine learning in-house rather than outsource it.

The seed was meant to be a SAFE. They ran a friends-and-family round on one, planned to keep the institutional round on the same paper for simplicity, and then Emergence Capital moved fast enough to price it immediately — converting the friends and family in the process. Raanani calls the seed “a pretty substantial six million dollar seed round,” then Redpoint led the A. Total raised at the time of the tape: “over 20 million dollars.” GetLatka’s funding rows for the same period record $6.3M dated June 2016 and a further $16M line dated October 2017 — about $22M together, which squares with his figure.

“We’re not there yet”

The exchange the headline rests on is short. Latka asked, as vaguely as Raanani wanted to answer it, whether Chorus had passed $10M. “We’re not there yet.” By the end of the year? “Not by the end of this year.” 2018? “We’re going to be pretty aggressive in 2018.”

Then came the detail that makes the number remarkable rather than modest. Chorus did essentially no revenue in 2016 — Latka put it at under $100,000 all in, and Raanani agreed; the year was spent on proofs of concept and on getting the product right. Pricing went on in 2017. So the $8.4M that GetLatka records on 12 October 2017 is the first full year of charging for the product, and it corroborates his “not there yet” from the other direction: below ten, well above anything a two-and-a-half-year-old company has a right to expect.

He beat his own forecast. The next dated revenue row is $15M in June 2018, so the $10M crossing happened inside the year he promised to be aggressive in, and it happened early enough that he had cleared it by half.

The rows after that are $20M in December 2019 and $55M in June 2020 — a steeper step than anything before it, and the one place in the ladder where the database jumps faster than the narrative on the tape would predict.

The product is the channel

Almost none of that came from spending. Raanani says most new customers arrived through referrals and inbound, that the head of marketing had been in the building a few weeks, and that growth to date had been “purely organic.” Pressed on paid, he would not give a figure and let Latka bracket him instead: under $50,000 the previous month. The strategy was to exhaust one channel before opening another, because the inbound was converting at a rate he described as incredible — and when Latka guessed “more than 20 trial to paid,” he confirmed it.

His explanation for where the inbound comes from is the thesis of this post. Sales communities are inherently viral, and so are meetings — and Chorus does not sit passively on the network, it joins as a participant. It compounds because of who Chorus sold to: several of the leading sales and marketing software companies were customers, and when their success teams ran quarterly reviews with their own customers, Chorus was in the room. He names Qualtrics, with hundreds of reps, alongside high-growth startups — Outreach, EverString, Visible, Uberflip, Vidyard — and Latka noted he had had many of those founders on the show inside the previous 60 days.

1. One rep switches it on alone. Chorus shipped a self-serve product for an individual account executive — one click, no rollout, no procurement — which Raanani says nobody else in the category had done.
2. It introduces itself in the meeting. The software joins the call as a named participant, so the rep explains it to the customer unprompted.
3. The customer is often another sales team. Chorus sold into companies that sell into sales, so the audience for that explanation is frequently a buyer.
4. Leadership is the last conversation, not the first. Once reps are using it and liking it, the VP’s standing objections — how hard is this to set up, will my reps actually use it — are already answered.

The four numbers he would not give

For a company built on measuring what everyone else leaves unmeasured, Chorus was running with a lot of instruments dark, and Raanani was candid about which ones.

• Fully weighted acquisition cost — “I don’t have great data on the CAC right now just because we’re experimenting with a lot of different approaches.” With most growth organic and the marketing hire three weeks old, there was no stable denominator to divide by, which is the usual reason acquisition-cost benchmarks stop being comparable at this stage.
• Paid spend — “I don’t feel comfortable sharing those numbers yet.” Only the bracket survived.
• Viral coefficient — they had started tracking it. No number, and no claim that it was above one.
• Negative net revenue churn — “we are going to find out a lot in 2018 because we’re going to have a lot of those contracts coming up for renewal.” They had only seriously gone to market at the start of 2017, so the first real renewal cohort did not exist yet.

He would go as far as the shape of the motion without the arithmetic: expansion inside year one on most deals, because the customers are growing companies, and a land-and-expand pattern that starts with account executives and moves to SDRs, BDRs and customer success. Whether that added up to negative revenue churn he refused to say, and his reason for refusing was the right one — the sample did not exist. Even the customer count was given as a range rather than a figure: “we don’t get into the specifics, but you could peg us around 100 enterprise customers.”

After the tape

What happened next, and when

Everything in this section postdates the October 2017 conversation and was unknown to both men at the time. The funding and headcount lines are GetLatka’s dated rows; the exit is public record, announced by ZoomInfo in July 2021 and filed with the SEC that month.

• June 2018 · Series B $33M recorded. Revenue $15M that month; headcount 72 by that November, 13 of them engineers.
• Dec 2019 · $20M Team of 95, with 24 in sales and 7 in marketing — the marketing function Raanani had just started hiring.
• June 2020 · Series C $45M recorded, revenue $55M, 117 people.
• Jul 2021 · Acquired ZoomInfo announced the purchase of Chorus.ai for approximately $575 million in cash. Headcount 191.

Two figures disagree and both are on the record: the title Latka gave the repost says Chorus sold for $550m, while ZoomInfo’s own announcement and GetLatka’s exit row both carry $575M. The larger number is the one with a filing behind it.

$575MZoomInfo’s announced cash purchase price, July 2021 — against the billion-dollar company Raanani said in 2017 was theirs to lose in five to seven years

Back on the tape, the last question

Raanani got to Chorus by a route he describes as deliberately slow. Bain out of college, an engineering science degree from Toronto, a Stanford MBA, then first hire at Eric Schmidt’s Innovation Endeavors — which he left because “it felt disingenuous for me to be advising these incredible entrepreneurs if I’d never started a business myself.” He went back to Bain to clear roughly $150,000 of business-school debt as fast as he could, then spent over a year looking at different areas, not for a gap in a market but for something he was authentically excited about and would give the next ten years to. Machine learning, voice and conversations was where that landed. He was 35 on the tape, married with two kids, aiming for seven hours of sleep, naming “The Hard Thing About Hard Things” as his business book, Bezos as the CEO he studies and Slack as his favourite tool that is not his own.

Then Latka asked the question he always closes on: take us back 15 years, what do you wish your 20-year-old self knew?

“Observe. Try to really understand what it was that the people that you looked up to, that you were working with, were doing. Try to really understand what it was about them that made you admire them and make you work your ass off.”

Roy Raanani, co-founder and CEO, Chorus.ai

Sources Nathan Latka’s October 2017 interview with Chorus.ai co-founder and CEO Roy Raanani, reposted to YouTube in July 2021 as a Latka lookback; GetLatka’s dated revenue, headcount and funding records for Chorus; ZoomInfo’s July 2021 investor announcement and its Form 8-K of the same month, for the acquisition price.