HackerEarth Refuses to Give Its CSMs an Expansion Quota
HackerEarth's CEO pays customer success managers on retention and never on expansion, because a quota can be hit while a fifth of the accounts walk out the door.
On this page
There is a fight in customer success circles about whether CSMs should carry an expansion quota. Sachin Gupta has picked a side, and the reason he gives is a specific failure mode rather than a philosophy.
If you give your CSM a revenue quota, then they are only worried about revenue — in the sense that even if 20 percent of my accounts are churning, but two big accounts can compensate for the revenue loss, they’ll go for it. And as a company I don’t want that. Every customer is important to me.
Sachin Gupta, founder and CEO, HackerEarth
The quota rewards the aggregate and hides the distribution. A CSM who hits their expansion number while losing a fifth of their accounts has done their job by the metric and damaged the business. Gupta pays for retention instead, and treats expansion as a consequence.
What he pays for instead
“I personally believe that CSM should be responsible for driving product adoption, building a customer champion, and making your customers referenceable,” Gupta said. “Expansion will happen naturally if all those three things are happening.”
The compensation follows that literally. An enterprise CSM at HackerEarth is measured against a retention rate, and the target is set as an increment on what the team already achieves.
- Start from the actual baseline. “We are already operating at say about 95 percent retention, and this is not including the expansions.”
- Set the goal just above it. “The goal that the person gets is in the range of 97, 98 — so they should be able to make an incremental improvement in terms of our historical retention rates.”
- Pay referrals separately. Referrals sit outside the core structure — “additional cherry on the cake” — so they can never substitute for keeping accounts.
- Segment the book by how much hand-holding it needs. An enterprise CSM owns 50 to 60 accounts; an SMB manager can own up to 100.
The blended result at the time: gross revenue churn around 15 to 16 percent, and net revenue retention of 106 percent last quarter, 110 projected, with 120 as the objective. The two halves of that blend are wildly different — enterprise revenue churn “in early single digits”, SMB at 20 to 22 percent. Gupta did not dress it up: “Not something that I’m too proud of, but it’s just the nature of our business. SMBs, some of them stop hiring altogether year on year.”
A friend who did not get the job
HackerEarth exists because of one specific rejection. Gupta, a software engineer by training, watched the recruiting process fail somebody it should not have failed.
A good friend of mine who was top of our batch, top of the class, didn’t get through some of the top companies. And we were shocked — we all think this is the guy who’s going to get into Google, Facebook. And he just didn’t get through, because the recruiting process, the interviewing process, is in some sense broken.
Sachin Gupta, founder and CEO, HackerEarth
The product that came out of it is a coding assessment tool: candidates write code in the browser, companies use the results to decide who advances. Deliberately, it is not a recruiting agency and takes no percentage of anyone’s first-year salary. “We’re not sourcing candidates, we are more of an assessment tool and we fall after the sourcing stage.”
Pricing is per recruiter seat, with bundled assessment credits — and the choice of seats over assessments is a deliberate trade of upside for predictability. “Number of assessments could vary year on year. You sometimes have a slow time in terms of hiring, sometimes you have a sudden surge. So to introduce more predictability in the business, it’s better to price on number of recruiter seats.”
That last figure is the one Gupta volunteered as reassurance, and Latka agreed: “You’re not at the mercy of just top three accounts.”
From 5,000 assessments to a million
In its first year, around the end of 2012, HackerEarth ran about five thousand assessments in total. By 2020 it was running close to a million a year.
4.5Mdevelopers in the community, alongside roughly 500 paying customers
The community is a second product, not a marketing line: developers come to self-assess, and companies sponsor hackathons and coding challenges inside it. It also changes how investors price the business — “we typically get valued on SaaS plus community.”
The customer base widened in a way Gupta found surprising at the time and obvious in hindsight. “A mega trend that has taken place is, obviously, software is eating the world, and most of the businesses today are primarily software businesses.” So an aerospace company like Boeing, or GE in healthcare, builds internal dev teams — “and then obviously they had to build that recruiting muscle, and that’s where HackerEarth comes into the picture.”
Around 2,500 recruiters were active at any one time, with more than 10,000 having passed through the system over its life.
Burn as a step function
HackerEarth raised $11.5 million across three rounds — a $500,000 seed, a $4.5 million Series A, and a $6.5 million Series B in late 2018 that it never announced. Those three rounds match the funding history on the GetLatka profile, and the total is exact.
Gupta’s framing of burn is the clearest statement of the bootstrapper-with-VC-money posture in the archive.
Your burn should never be out of control. Obviously to grow the business you need to spend. So the way I look at our journey has been more of a step function: you invest, and you reach that level, so you kind of break even, and then you say, again, I’m going to invest. And that’s what we’ve always done.
Sachin Gupta, founder and CEO, HackerEarth
When COVID hit expansion, they cut marketing immediately and aimed to break even by the end of the year, burning under $500,000 a month. It is a company that treats profitability as a rung on a ladder rather than a destination.
Two numbers that do not agree
Two figures on the profile deserve a caveat. Only one of them got a marker, and the reason is worth stating.
The first is revenue. Gupta never states one. Latka does the usual arithmetic — “500 customers at around a thousand dollar ARPU, so somewhere around 500,000 bucks a month, or about a six million dollar run rate” — and the interview’s own title says the company “breaks $5m run rate.” The profile records $4.8 million for July 2020. Three different numbers, none of them from the founder. The recorded figure now carries an estimate marker.
The second is headcount, and it cannot be reconciled at all. Gupta gives a team of 140 and then breaks it down by function — about 30 in sales, six or seven in customer success, 40 to 45 in product including engineering, design and product management, and another 10 to 15 in marketing. The profile carries 394 employees for June 2020 and 443 for December 2020, on either side of a July conversation in which the CEO said 140.
One of those lineages is wrong, and the founder’s own itemised count is the better-evidenced one. Both are being left on the record rather than quietly reconciled, because the disagreement is the finding.
Asked what he wished he had known at twenty, Gupta gave an answer about geography rather than product. “We started our business in India and then we expanded globally. I think if we had known how big the market could be in the US, we would have probably made that move earlier.”
Sources — Sachin Gupta interviewed by Nathan Latka, recorded 29 July 2020. Revenue, headcount, customer and funding figures from the HackerEarth profile on GetLatka, with dates as recorded.
