Part IThe Story
The Suits Nobody Wore
In the third week of January 2017, about forty employees of a San Francisco software company called AppDynamics owned new suits. Some of them were expensive. They had been bought for a particular Thursday morning in New York, when the company would ring the bell at Nasdaq and become a public company. By then it was nine years old, employed roughly 1,200 people, was doing a little more than $200 million in annual revenue, and had raised north of $300 million across six rounds from Lightspeed, Greylock, and Kleiner Perkins. Its prospectus had been on file with the Securities and Exchange Commission since December 28. The roadshow was finished. The bankers had a number.
On Tuesday, January 24, Cisco announced that it would buy the company instead, for approximately $3.7 billion in cash and assumed equity. By the founder's account, that was about twice what the public markets had been prepared to pay. The reports of the timing disagree in the way reports of very fast things usually do. Wikipedia says a day before the offering. Other outlets say "days before." The founder, years later on a First Round podcast, said three days pre-IPO. Nobody disagrees about what came next. There was a party at the office in SoMa. People from Cisco came. By every account it was a very happy celebration. Then the founder walked home. He was the chairman and the largest shareholder, he was thirty-nine years old, and his 14 percent stake had turned that week into something like $525 million.
— Jyoti Bansal, to CNBC Make It, 2024We announced the sale, and then we had a huge party at our office in San Francisco. Folks from Cisco were there. It was a very happy celebration. But walking home after, I was sad and depressed. I spent nine years of my life fully dedicated to what we were building there. Suddenly, it's the end of a chapter.
He has called it the saddest day of his life. He has called the sale one of his biggest regrets. He says that in the same interviews where he explains, with an engineer's patience, why it was the correct decision. Both things appear to be true for him, and he has not tried to make them agree.
The suits are worth starting with because Bansal brought them up himself. Founders tend to remember details like that; reporters rarely think to ask. He mentioned them on the phone with the Economic Times the day after the deal was announced: "On Thursday, we would have been ringing the bell on Nasdaq, about 40 of our employees were planning to be there, some of them bought expensive suits for the IPO event." It is an odd thing to offer in the afterglow of a half-billion-dollar payday, unless what you lost that week was a morning and not money.
Nearly everything since can be read as a long, methodical, slightly obsessive effort to get that morning back. That includes the startup studio, the venture firm, the two companies run at once, the merger of those two companies into each other, a $5.5 billion valuation, and a Forbes billionaire listing in December 2025. The bell has not rung yet. In late 2025 Bansal said so plainly: "We want to IPO. We didn't get to do it last time."
By the Numbers
The Bansal Ledger
7 yearsOn an H-1B visa before a green card let him found a company
20Venture firms pitched before AppDynamics got its first term sheet
$3.7BCisco's price for AppDynamics, announced January 24, 2017
~400AppDynamics employees reported to have become millionaires in the sale
$5.5BHarness valuation after its $240M Series E, December 2025
16Product lines run as 'startups within a startup' inside Harness
~4,000People Bansal estimates he has hired since becoming a permanent resident
Seven Years Without Permission
The childhood barely appears in the record, and Bansal does not dress it up. He grew up in Ajmer, a small city in Rajasthan. On weekends he helped his father run a small business selling machinery to farmers. When a Bangalore podcaster asked him in 2018 to describe where he came from, the transcript gets as far as "So I grew up in a small town in Rajasthan and when you grow up in a—" and the excerpt stops there. That seems about right. He usually treats the origin as a clause and not a chapter. The one image that survives is plain and specific: a boy watching his father run a "mom-and-pop business," as Bansal put it decades later. Mostly that means watching a man sell things to people who had to be convinced to buy.
In 1995 he went to the Indian Institute of Technology in Delhi to study computer science. His dormitory wing was nicknamed the Tiger Wing, and his friends called him JKB. In 1996 he bought a scooter, and it became, in a batchmate's phrase, "public property for errands by many; he never objected." The batchmate was Neelam Dwivedi, who later co-founded a fintech startup called Funds Tiger. She remembered him as "kind-hearted, jolly and passionate," and also as someone who could think "10 steps ahead." The two descriptions don't obviously fit together, and that mismatch shows up again and again later. The generous dorm-mate who lends out his scooter and the strategist ten moves ahead turn out to be the same person, and the scooter detail will come back nineteen years later in a much bigger form.
Two visitors to his imagination did more than any professor. One was Bill Gates, whose campus visit he has cited as an influence. The other was Sabeer Bhatia, the Indian-born engineer who co-founded Hotmail, sold it to Microsoft, and walked away with tens of millions of dollars. Bhatia was the story a generation of Indian engineers told themselves late at night. "That's what brought me to Silicon Valley," Bansal has said, and he means both men, and the idea that you could start a company from nothing, as he later put it, "that provided outstanding services, technologies, and jobs."
He landed in California in 2000, at twenty-one, with a few hundred dollars and an H-1B visa. His first employer had about thirty people. He found a buzz in the Valley, he remembered, "everywhere, even in the coffee shops." He wasn't sure he would stay. The plan was to get experience and then start his own company, "maybe in the US or maybe in India."
The visa decided the timeline for him. An H-1B lets a foreign engineer work for an American company. It does not let him found one. He has explained this many times with a mildness that seems to cost him some effort.
— Jyoti Bansal, to Forbes, 2025The challenge unfortunately is if you're on an H1-B visa, you're not allowed to start a company and create more jobs, which I find very ironic.
So he waited seven years. He worked as an engineer and architect at three small enterprise software companies that sponsored his visa. They were Wily Technology, later acquired by CA; Datasweep, later acquired by Rockwell Automation; and netLens, acquired by NextPage, whose assets went partly to FAST Search and Transfer and then to Microsoft. All three ended in acquisition. None of them rang a bell. A young man watching each employer get absorbed into a bigger one might have drawn a lesson about the usual fate of enterprise software startups. If he did, he didn't say so at the time. "I knew I could create the technology to solve a specific problem," he wrote much later. "I just needed the right immigration status."
The green card arrived in 2007. Strictly speaking, it hadn't arrived yet. Bansal quit his job "even before the final paperwork came through." He incorporated AppDynamics in April 2008, a few months before Lehman Brothers collapsed. He became an American citizen in 2016, the year before the Cisco sale. He says citizenship mattered a great deal to him personally and changed almost nothing about his work, except that "business travel became easier once I had an American passport."
Later he would estimate that he had hired about 4,000 people since becoming a permanent resident. Seven years of waiting for permission, followed by a hiring count in the thousands: the gap between those two numbers is the closest thing in his story to a political argument. He states it once and then lets the numbers carry it.
Fifteen Months of Being Wrong
There is a common story about the immigrant engineer who sees a problem nobody else sees, builds the obvious solution, and gets rich. Bansal's version is less flattering, and he is the one who tells it that way.
He had spent seven years inside other people's companies, and he said later that he learned more from what went wrong there than from what went right. The first startup he worked for was founded by four people, "all PhDs from MIT and such," and it had one fatal defect: "there was no one willing to buy their product. It's frustrating." He has described his first two employers as failures and said, "you learn a lot more from failed startups than successful ones." So when he started his own, he set a rule. He had to have at least four customers to talk to, real ones, before he believed anything.
The underlying insight was sound. By 2007 he could see that applications were becoming distributed, with more moving parts, more virtualization, and more code spread across more machines. "People were building apps everywhere," he said. "And every time a company produces code, it needs to be monitored. The existing solutions would not be good enough for modern applications." What he built on top of that insight was mostly wrong. He has listed the mistakes publicly, in order, the way an engineer writes up a postmortem.
The first was timing. He assumed that within two or three years everyone would be on the cloud and would need cloud monitoring. "The reality was that it was a much longer timeline and there's no funding for being right eventually." The second was architecture. He imagined a SaaS product that customers would adjust their systems to fit, and customers told him they wanted everything on-premise and wanted to change nothing. The third was ambition. He was building "a complex product that would manage the entire monitoring process — from finding to fixing issues," and customers only wanted the "find it" part. In each case the fix was to do less: work for companies not yet on the cloud, require no changes to their application architecture, cut the roadmap down to plain visibility. It took fifteen months and several pivots to ship a decent first version. "If we hadn't pivoted," he wrote, "we would have died."
He did this during the worst financial crisis in eighty years. The markets he had assumed, finance and banking among them, had stopped taking any spending risks at all. He had no customers and no prospect of more funding. He did the unglamorous thing. He searched LinkedIn for people responsible for IT operations, sent hundreds of cold InMails, and then studied the replies the way he would study a log file. Were small companies answering, or large enterprises? Which job titles? He asked for feedback, not money. He wanted people to feel "like they're partnering with you." Then he sorted them, in a phrase that has since become Unusual Ventures orthodoxy, into "the desperate" and "the curious." The curious were pleasant and useless. The desperate were customers.
Raising money worked the same way, with the same odds. "Back when I was a first-time founder in my 20s, I pitched 20 VCs before one finally gave me a term sheet." (He was actually thirty when he founded AppDynamics. Memory compresses things kindly.) The investor who said yes was John Vrionis, then at Lightspeed Venture Partners. Vrionis is from Georgia. He went to Harvard, studied economics, and played varsity soccer there. He then picked up a master's in computer science at the University of Chicago, apparently out of sheer appetite, arrived in the Valley in 2002, worked in product management and sales, got a Stanford MBA, and became a general partner at Lightspeed. "I gravitate to people who have an immense drive to overcome and the humility to keep learning," he has said. By the 2020s he would describe Bansal as his friend of twenty years and his co-founder, and note that he had been an early investor in all three of Bansal's unicorns. In 2008 he was an investor writing a check to a thirty-year-old with no customers, no co-founder, and a product built for a cloud that didn't exist yet.
The missing co-founder bothered Bansal more than he let on at the time. "There are depression moments, there are moments like when things are very hard, you want people with whom you can talk to, even though it's a lonely job," he told a room of Accel invitees in Bangalore in 2018. "In my case I didn't have a co-founder which is one thing I really think was a bad idea." His first hire became what he called "a de facto co-founder," and that was enough. The sentence in the transcript trails off unfinished — "you want someone who you can talk to but everything…" — which tells you more about how lonely it was than any finished sentence would.
The early customers were companies whose problems were urgent and very public. Priceline used AppDynamics to monitor a sprawling set of applications. Netflix was moving its whole business online. "Imagine you're a consumer on Netflix who gets frustrated because video is buffering," Bansal explained to Forbes. "We help their engineers to make sure no glitches happen or if glitches happen, they can fix them very fast." Later HBO became a customer, and in August 2015 CNBC ran a story under the headline "How this developer's code saved 'Game of Thrones.'" That is about as much glory as application performance monitoring is ever likely to get.
The Engineer Who Learned to Sell
Early on, Bansal held the engineer's faith in its purest form: a good enough product sells itself. He calls this "a dangerous mindset" now. At the time it was simply what he believed.
It held up for a while. By landing a few big accounts like Netflix, AppDynamics reached several million dollars in annual revenue, and Bansal "was feeling confident about our product." Then a fellow CEO said something that, by Bansal's own account, changed the course of the company: if AppDynamics ever hoped to cross $100 million, it needed to "get scientific about sales."
He took this literally. That is his method, and it is also a little funny. He approached sales the way he had once approached distributed systems, as a discipline that could be studied. "If a software engineer can learn programming, they can learn sales," he wrote. "That said, I won't sugarcoat it — learning sales is a years-long process." He studied companies like his own that were three to five years ahead. How did they run their sales process? Whom did they hire? He hired John McMahon, whom he calls "a top authority on enterprise software sales," for weekly whiteboard sessions. Around him gathered a group of advisors he still lists with something like reverence: Dev Ittycheria, John McMahon, Dali Rajic, Jeremy Duggan, Joe Sexton. The company developed something called a Business Value Assessment, a structured way of measuring in dollars what a customer stood to gain. Bansal has called it "a game changer." Eventually he hired a VP of sales, which he counts among the best decisions he made as a founder.
This education produced a new vocabulary that he still uses. "Product-market fit," in his view, is not enough. What matters is "product-market-sales fit," the point where you not only have something people want but can also explain its value and sell it repeatably. A successful company, he says, is a "three-layer cake." A great product sits at the bottom, and "without a solid base, the business is doomed to failure," and on top of it sit go-to-market excellence and operational excellence. The job of the founder is "to keep taking a company to the next level — and that can't happen unless you bring in the right expertise."
The phrase "next level" is not a metaphor for him. He thinks of a company as a video game. "To me, the level zero was getting funded, then getting the product market fit, then figuring out the early sales machine, hiring executive level people for the first time, and so on." The biggest mistake entrepreneurs make, he says, is "spending time on skills they need for level 3 when they haven't even passed level 1." Asked once for the two things that turned him from an engineer into a CEO, he refused the premise: "there are two hundred small things that make this transition happen, not just two things."
For problems he didn't understand, he had what he called his "rule of 10." "If I didn't have experience on building a sales funnel, I'd ask 10 smart people for their opinion or solutions." He was suspicious of the general-purpose advisors who gather around first-time founders. "Being a first-time entrepreneur, there were a lot of people offering to connect me with someone or other, and there were many of them giving advice." He came to prefer what he calls "topic-based mentorship, not generic mentorship." He tried a CEO coach once. It "didn't work very well for me."
At this point, to be fair to how well it went, AppDynamics was among the fastest-growing software companies of its era. By early 2015 it had reached a $1 billion valuation in under seven years. When Business Insider asked how that felt, the headline used his one-word answer: "Scary." By the end of 2016 the company had more than 1,600 customers, double the year before, including Barclays, JetBlue, Expedia, Kraft, Nasdaq, HBO, and, as it happened, Cisco. In September 2015, with more than 900 employees, Bansal handed day-to-day operations to an outside chief executive, David Wadhwani, and became chairman and chief strategist. The next June, Ernst & Young named him Entrepreneur of the Year for Northern California.
He has never said much publicly about the decision to step aside. It's possible that a man who learned sales like a second language also saw that running a 1,000-person company going public was its own level with its own skills, and that he had made himself an apprentice again. It's also possible the explanation is less tidy than that. The record doesn't say. What it does say is that sixteen months later, as chairman, he was the one who had to decide whether to take Cisco's money.
The Arithmetic of Yes
Bansal does not describe the decision as a feeling. He describes it as three variables.
The first was strategic: "We could be part of a bigger platform, like Cisco, and their customer base and market." The second was cultural: "what kind of home your employees get." He gives Cisco credit, years later and with no apparent reluctance, for granting "a high degree of independence to the AppDynamics unit." The Economic Times reported at the time that the offer was accepted partly because the nine-year-old startup's identity would be preserved. The third was financial: "Is it the right thing for the stockholders — founders, investors and employees?"
The third variable is where he spends most of his time when he tells the story. He projected the IPO path and estimated it would take three to four years of strong public-market execution to reach a $3.7 billion valuation on its own. Cisco was offering that number now, in cash. "By selling, we reduced that risk for all the employees. That's a significant impact." And then: "When weighing the options, it wasn't about me. The biggest factor was our employees."
The number of employees who became millionaires has grown with retelling. Headlines in India and the United States settled on 400 out of 1,200, one employee in three. When Bansal himself gave CNBC the figure, it was smaller and stated more carefully: "we had about 300 or so employees who made more than a million dollars in the final offer we accepted. We had dozens of employees with $5 million-plus outcomes." Founders are rarely accused of underselling. Here the man at the center of the story is quietly correcting the legend downward, which says something about him, or at least about how he relates to numbers.
He remembers what the money did. "People in the company were ecstatic; they had never imagined so much money. Many were buying new houses, new cars. I know one person who took six months off, rented an RV, and travelled around the country." The story has an older counterpart. In 1998 Jay Chaudhry, another Indian-born founder who would later build Zscaler, sold his company SecureIT to VeriSign for $70 million. Two years later, when VeriSign's stock surged, seventy of SecureIT's eighty employees became millionaires. "One employee took six months off and traveled the country," Chaudhry recalled. The two founders, a generation apart, tell nearly the same story down to the six months and the open road. It's tempting to treat the parallel as myth, as if every immigrant founder's windfall ends with an employee on an American highway. It's more likely just true twice. Six months is about how long it takes a person to find out what they want when nobody is making them work.
It took Bansal about six months too.
The arithmetic was correct, and he has never seriously argued otherwise. What the arithmetic left out was the one stockholder whose return wasn't measured in dollars. "I really enjoyed building the company, creating products, solving problems, competing in the market, every element of company building that I'd gone through," he told CNBC. "It was intense and stressful, but I really enjoyed it. I also felt like we didn't fully finish what we could've." AppDynamics, he says, could have become "much bigger." The Economic Times reported in 2025 that the unit now produces more than $1 billion in annual revenue for Cisco. He doesn't need to point that out, and he doesn't.
Asked whether he would accept a similar offer for Harness today, he said he almost certainly would not.
After the Bucket List
"Too young to retire" is how Bansal described himself to the Economic Times the day after the sale. It was mostly a joke. He took several months to test it anyway.
"Everyone told me, 'You should retire. Go on the beach. What else do you need to do?'" he recalled. "That was my first instinct, as well. I wanted to trek in the Himalayas, hike Machu Picchu, do a safari in Africa, see the fjords in Norway. In six months, my bucket list was done. And I started to realize: That's not it for me."
The bucket list is a revealing document. It reads like a product roadmap: four line items, each completable, each checked off. A man who thinks of company-building as levels in a video game had given himself a vacation with levels, and finished it the way he finished everything, on schedule. Then there was nothing left to ship.
"I tried to retire," he told Forbes years later, with the flatness of someone reporting a failed experiment. "People say, 'Once I retire, I'm going to do what I enjoy.' I asked myself, 'Do I enjoy playing golf all the time or being on the beach all the time?' I don't really."
Sometime in 2017, back on his couch and, as Fortune put it, "not actually sure what he was supposed to be doing. As in, with his life," Bansal watched Jiro Dreams of Sushi. The documentary is about Jiro Ono, a sushi chef who was eighty-five during filming and had spent essentially his whole life perfecting a single craft in a tiny Tokyo restaurant. In the first three minutes Ono looks into the camera and states his philosophy: "Once you decide on your occupation… you must immerse yourself in your work. You have to fall in love with your work. Never complain about your job. You must dedicate your life to mastering your skill."
On paper it is a strange match: an eighty-five-year-old chef who does exactly one thing, and a forty-year-old software founder who was about to start doing four. The match makes sense once you see what Bansal took from it. "If you look at the movie, it really isn't just about the outcome, it's about the process of getting better and better." The sale had given him an outcome. What he was missing was the process. Ono does not sell the restaurant. He goes back in the morning.
Years later, on Alisa Cohn's podcast in May 2026, the identity crisis after the exit was still the first topic of conversation. "Entrepreneurship as a craft, not a one-time event" was how the episode summed up his conclusion. The word craft matters here. A craft has no exit. You get better at it, and you die.
Parallel Entrepreneurship
In October 2017, about nine months after the Cisco announcement, Bansal launched two things at once. He would keep doing that.
The first was BIG Labs, short for Bansal Innovation Group, which he called a startup studio. Its purpose sounded like a lab notebook: "really my lab, a research lab for me to experiment with ideas and projects that I'm excited about." The studio's stated principle was "parallel entrepreneurship," building several companies in tandem, each one starting with "ideation, a market thesis, and initial funding" and moving, if it survived, through prototyping, validation, and product-market fit into an independent company. The idea behind it was simple and a bit greedy. If building is the thing he enjoys, why do it only once at a time?
The second was Harness, the studio's first company. Its premise came from an old source of insight: AppDynamics customers. Many of them were unhappy with how they shipped software, the continuous integration and continuous delivery pipelines that carry code from a developer's laptop into production. Bansal noticed that for every ten to twelve engineers at these companies, there was at least one DevOps engineer whose job was to build delivery automation and, as he put it, "babysit deployments." The large banks he talked to employed around 10,000 engineers each, and hundreds of them did nothing but write automation scripts. "People don't often realize writing code is just the first 30% of the job," he said later. "Then you have the next 70% of the job to make sure you test that code properly." Or, as he put it in the line that gave the company its name: "The whole world is running on code, whether banking or transactions or airlines or everything — and that code needs a safety harness."
The company's own description promises something closer to a personal favor: Harness would make software delivery simpler, "giving devs their nights and weekends back." The engineer who spent seven years working for other people's companies built a product for the engineers who still do.
In May 2018 came the third thing, a venture firm, Unusual Ventures, co-founded with Vrionis. The founding announcement described a seed fund of $160 million. A $400 million second fund followed in November 2019, a $485 million third fund after that, and by the mid-2020s the firm described itself as managing more than $1 billion. Bansal's title there is "Entrepreneur Partner," which is both precise and a hedge. He designed the firm's "Get Ahead Platform," which provides seed-stage founders with dedicated operational staff. The idea came straight from his own early years: "the importance of doing the work WITH entrepreneurs versus just giving advice." He teaches in Unusual Academy alongside people like Andy Rachleff and Adam Grant.
The firm's fundraising produced one of the few moments in Bansal's public life that could be called a stand. In late 2018 the New York Times reported that peers had advised the two founders to go after the easy money, the sovereign wealth funds of Saudi Arabia and Abu Dhabi, which by then were funding much of the Valley. "People would say, 'It's really easy — they'll give you as much money as you want,'" Vrionis told the paper. The two "didn't feel comfortable making investments on behalf of repressive governments." They raised from nonprofit groups, historically black universities, and children's hospitals instead. Bansal reposted the article on LinkedIn without comment beyond the excerpt.
Investing changed him in ways he describes with unusual candor. "Early on, when a founder pitched us at Unusual Ventures, I'd think: 'Great idea. Big market. I know exactly how I'd build this.' I'd start visualizing the product roadmap, the go-to-market playbook, how I'd scale it." He had to train himself out of it. "You can't assess a startup based on what you would do." It's an honest admission, and an unusual one. The thing he enjoys most, imagining himself building a company, is the exact instinct that makes him worse at judging other people's.
Then, in July 2020, came Traceable, an API security company and the second product of BIG Labs. Bansal was CEO of it as well. He was now CEO of two companies at the same time.
The investors were not uniformly comfortable. Steve Harrick, who joined Institutional Venture Partners in 2001 and sat on the boards of both Harness and Traceable, gave CNBC the most honest description of the arrangement anyone has offered.
— Steve Harrick, Institutional Venture Partners, to CNBC, 2025I've never done this before, backed a CEO to run two companies simultaneously. But Jyoti is that good. He's not only a great executive, but he hires well and he delegates well, and so I just talked to Jyoti. I said, 'This is a major risk.' I got his assurance he wouldn't do a third one.
The last sentence is the one to hold onto. A twenty-year venture veteran felt he had to get a verbal promise, effectively a handshake, that the man would stop at two.
Matt Murphy of Menlo Ventures invested in both AppDynamics and Harness. Before that he spent more than fifteen years as a general partner at Kleiner Perkins, where he was an early observer at Google and launched the iFund with Apple; later he led Menlo's investment in Anthropic. He has seen as many founders as almost anyone. "You could count the entrepreneurs on two hands that have been able to do something like this, maybe one," he told Fortune. He offered Jack Dorsey as the best comparison. Fortune's headline called Bansal "the 'Elon Musk of enterprise software'" and then added "(sort of)," which is about right.
The way he actually works is less dramatic than the comparisons suggest. Asked how he ran two companies and a venture firm, Bansal gave an answer that sounds almost disappointing.
"It's very ironic that people think, [because] I'm CEO of two companies, I should be the best at multitasking. Actually, my cheat code is to not do too much multitasking." At AppDynamics, he said, "I used to be in a meeting and I'm on my phone replying to emails and responding to messages all the time." Now he works in blocks of thirty minutes, sometimes an hour. One block goes to engineers on "technical problems I'm interested in." The next goes to investors and the board. The next goes to a customer. "And I can do it between the two companies." He summed up the method in one sentence that could hang on his office wall: "So you do multitasking on an aggregate, but not at any one moment."
The parallel entrepreneur turns out to be a serial worker, cutting his day into very small sequential pieces.
No Prize for Being Eventually Right
Anyone who assumes a second unicorn is easier than the first should look at what happened to the third.
Bansal answered the question himself in a chapter of David Reske and Mary J. Cronin's The 100X Entrepreneur: Stories of Extreme Growth. The chapter asks: "Does building a unicorn once make it easier the next time? The short answer? No." Some parts did get easier. Investors now came to him, board meetings no longer intimidated him, and he knew what to expect. The parts that matter most did not. "Finding the product-market fit is still the same iterative process. We had to pivot at Harness as well."
At Harness the founding assumption was that the ideal customer would be a large enterprise like the banks with 10,000 engineers. Early conversations found more urgency in mid-sized companies starting a new initiative or moving to a new architecture such as Kubernetes. His discovery method had become more rigid, if anything. He insisted on at least fifty "cold hard conversations" with prospects outside his own network, because strangers are "more direct and blunt with their feedback." He ran every prospect through what he calls the "3 Whys": Why buy anything? Why buy now? Why buy us? He stopped using the phrase "minimum viable product" in favor of "minimum sellable product." In an existing market with an established incumbent, a new entrant has to match everything the incumbent does and then add something the incumbent can't, or nobody will pay.
He also overcorrected. Having learned at AppDynamics that hustle doesn't scale, he "tried to force that machine into Harness a bit too early," installing a systematic sales operation in a company still in the hustle stage. He had learned the lesson so well that he applied it before it was time.
Traceable gave the clearest demonstration. Bansal has written about it several times, with the pleasure of someone who now finds his own past mistakes useful. "When I started Traceable," he wrote in 2025, "we got almost all of our assumptions wrong — and this was my third company." On what to build, they assumed companies wanted API protection; it turned out they first needed API discovery, because they didn't know what APIs they had. On whom to sell to, they targeted mid-sized cloud-native companies, but the real pain and urgency was in large enterprises in finance and healthcare. On how to sell, they started with product-led growth, and classic enterprise security sales turned out to be the answer.
That is three wrong assumptions out of three, from the founder of a $3.7 billion exit and a $3.7 billion unicorn, in a company that started in a dedicated BIG Labs room at Harness's San Francisco headquarters, where co-founder Sanjay Nagaraj remembers working on it.
"Eventually, we were able to make Traceable a success and years later, many assumptions proved true," Bansal wrote, "but there's no prize for being 'eventually right.'" It's the same sentence, nearly word for word, that he used about AppDynamics and the cloud in 2008. Sixteen years apart, two companies, the same lesson learned again from scratch. He doesn't present this as failing to learn. He presents it as the nature of the work. You can learn to recognize that you are wrong. You can't learn how to avoid being wrong in the first place.
One Company, Sixteen Startups
By 2024 the two companies had begun to merge on their own, through their customers.
"The Harness team would go set up a meeting with an executive at a bank or some of our customers," Bansal told CNBC. "I would go into the meeting and the executive would say, 'It's a one-hour meeting. Can we save the last 15 minutes? Because I also want to talk about Traceable.'" Employees at all-hands meetings at both companies kept asking about consolidation. The original reason for separating them had been that their products went to different buyers within a company. That reason was disappearing, as engineering leaders began buying security tools too. A company spokesperson said 70 percent of Traceable's largest customers were already Harness customers.
There is one small, almost comic detail in the CNBC account. Bansal had been "effectively the first IT person at both companies," setting up the same Google productivity apps and the same Carta equity-management software for each. One man configured the same software twice for two companies that were going to merge.
The merger was announced in February 2025 and completed in March. The combined company would have about 1,100 employees, an expected $250 million in 2025 annualized revenue, a 50 percent growth rate, and a valuation around $5 billion. Bansal pointed out the comparison himself, unprompted: "It's about the same size that AppDynamics was when we were about to go public." This was eight years later and a different company, but it had brought him back to the same place. He was at the front door again.
He had learned not to need a separate company for each idea. "If I was starting Traceable now," he told a reporter later that year, "I would have just launched it inside Harness because now I have this startup-within-a-startup infrastructure." The phrase "startups within a startup" now describes how the whole company is built. Harness runs as a set of sixteen product lines: continuous integration, continuous delivery, feature flags, cloud cost management, chaos engineering, an internal developer portal, and others. Each is built by its own team, and each can be bought on its own without committing to the rest of the platform. New ones launch at a rate of two or three a year. Each starts as a team of four or five people in an ideation phase and gets more funding and headcount only if it shows promise, the way a venture firm makes follow-on investments. Bansal describes the portfolio in exactly those terms, with "stealth, seed stage, early stage and late stage pieces" inside a single company. AI, he notes, has lowered the cost further by cutting the number of engineers a new bet needs.
So BIG Labs did not really close. It moved inside Harness. The venture firm moved inside too, in the sense that he now reviews his own product lines with an investor's habits. The parallel entrepreneur found a way to run parallel companies without asking any board for permission to start a third.
The product had followed the technology. On July 21, 2026, Harness announced something called Agent DLC, a lifecycle for building, testing, and securing AI agents. Bansal's explanation of the opportunity sounded like his earliest insight at AppDynamics, scaled up: "If AI is helping write, the volume of code is going up something like 10 times and now people are struggling to test it all." The engineer who once noticed that every new piece of code would need monitoring now notices that machines are writing code faster than people can check it. It is the same observation applied to a much larger flood of code.
The Bell, Again
In December 2025, Harness closed a $240 million Series E. Two hundred million dollars was new money led by Goldman Sachs Alternatives. Forty million was a tender offer for existing shares, with IVP, Menlo Ventures, and Unusual Ventures among the buyers. Bansal's own fund was among the buyers of his own company's stock. Nobody seemed to find this odd, and maybe by then nothing in the arrangement was odd. The valuation was $5.5 billion. Forbes estimated that his roughly 30 percent stake, plus what was left of the Cisco money, put his net worth at about $2.3 billion. Indian outlets ran the story the same week under some version of the same headline: IIT grad who went to US on H-1B visa becomes new billionaire.
The timing gave the news a political edge he didn't choose. By late 2025 the H-1B program was under heavy pressure, with new compliance checks, policy changes, and open hostility. Bansal's story was suddenly an argument people wanted him to make. He made it in his usual way, carefully and without heat. Immigration, H-1B talent included, is central to Silicon Valley. "I always prefer to hire Americans. And yet, that's not always possible." The irony he had pointed out years earlier, that the visa that brought him lets you work but not create jobs, needed no new emphasis. Four thousand hires made the point.
The story has a second geography, and it closes a loop. By the end of 2024 Harness and Traceable employed about 400 people in India, across Bengaluru, Delhi-NCR, and Hyderabad. Bansal planned to raise that to 1,000 in three to four years, and said 60 to 70 percent of product development already happened there. He hires from the IITs, the IIITs, and the NITs, the institutions that produced him. He has a theory of India in three waves: services exporters like Infosys and Wipro, then domestic consumer companies like Ola and Flipkart, and now Indian companies selling software products to the world. He mentored the Pune-based Mindtickle through a $12.5 million raise from Silicon Valley investors in 2015, before anyone called him a billionaire. In December 2024 he was guest of honor at IIT Delhi's Alumni Day. A classmate in the comments mentioned seeing him at their silver jubilee reunion, twenty-five years after the Tiger Wing and the scooter.
He has said more than once that building a company takes "ten years at least." Harness turns nine in October 2026. AppDynamics was nine when it sold. The comparison is obvious, and he has made it himself. Whether Harness reaches ten as an independent company, and whether it reaches a Thursday morning in New York, is the one question in his career he can't settle with a framework. "We want to IPO," he said in December 2025. "We didn't get to do it last time."
There is one more detail, from earlier that year, and it is the one this story keeps coming back to.
In May 2025, Bansal went on stage at TiEcon in Santa Clara for a fireside chat about go-to-market strategy in the age of AI. Afterward he wrote that they had "touched on growing up in India — watching my father run his own mom-and-pop business in Rajasthan." Anurag Wadehra, who moderated, added one line beneath the post: it had been "a special treat to have your parents in the audience."
So in a convention hall in Santa Clara, the man who spent his weekends in Ajmer selling machinery to farmers sat in the audience while his son explained to a roomful of founders how to sell.
Part IIThe Playbook
What follows distills Bansal's methods into a working set of principles. They come from three companies, a venture firm, and a large public record of his own mistakes. Few of them are glamorous. Most are corrections, beliefs he held, found wrong, and replaced with something more disciplined. What sets the playbook apart is not any single idea. It is the consistency: one engineer's way of debugging, applied to customers, sales teams, capital, his own calendar, and finally the shape of the company.
The Bansal Playbook
- 1.Treat the waiting room as an apprenticeship.
- 2.Hunt the desperate, not the curious.
- 3.There is no prize for being eventually right.
- 4.Learn sales the way you learned to code.
- 5.Play only the level you are on.
- 6.Ship a minimum sellable product.
- 7.Mine your last company's customers for your next idea.
- 8.Price the exit in other people's risk — and your own regret.
- 9.Multitask in aggregate, never in the moment.
- 10.Run a venture portfolio inside the company.
- 11.Hire hunger, not trophies.
- 12.Choose qualified soft advisors and topic-specific mentors.
J
The Decision Timeline
Key moments where a single choice shaped what followed.
1995
Enters IIT Delhi to study computer science; influenced by Bill Gates's campus visit and Sabeer Bhatia's Hotmail sale.
2000
Arrives in California at 21 on an H-1B visa with a few hundred dollars.
2007
Receives green card; quits his job before the final paperwork arrives.
2008
Founds AppDynamics in April; spends 15 months and multiple pivots finding product-market fit.
2015
Hands the CEO role to David Wadhwani in September; becomes chairman and chief strategist.
2017
Cisco announces the $3.7B acquisition on January 24, days before the IPO; BIG Labs and Harness launch in October.
2018
Co-founds Unusual Ventures with John Vrionis; declines sovereign wealth capital.
2020
Launches Traceable in July; becomes CEO of two companies simultaneously.
2025
Merges Harness and Traceable (February–March); raises $240M at a $5.5B valuation in December.
Principle 1
Treat the waiting room as an apprenticeship
Bansal spent seven years legally barred from doing what he came to America to do. The easy story treats those years as lost, a policy injustice standing between a founder and his company. Bansal tells it differently. "I realized then how valuable my time working in other startups when I was on the H-1B visa had been," he wrote. "I learned a lot about what went right and — even more importantly — about what went wrong at those businesses."
The most important lesson came from a failure: a startup run by four MIT PhDs that could not find a single buyer. Watching that company taught him the rule he took into AppDynamics, never to build without at least four customers to talk to. All three of his employers were eventually acquired, and none went public. That gave him an unusually close look at the usual ending for enterprise software companies, years before he had to choose an ending himself.
The general point is that a period of forced waiting, whether from a visa, a non-compete, a vesting schedule, or a family obligation, can be turned into structured observation of other people's mistakes. Bansal did not wait passively. He used other people's companies as test cases and treated their failures as experiments run at their expense.
Tactic: If you can't start yet, keep a written log of every decision at your current company that you would make differently, and why. That log becomes your first operating manual.
Principle 2
Hunt the desperate, not the curious
His best-known contribution to founder craft is a sorting rule. In 2008, with no customers and a dead funding market, he sent hundreds of cold LinkedIn InMails to IT operations leaders. He asked for feedback rather than a sale and then analyzed who responded by company size, job title, and industry. Most respondents were curious, and he treats the curious as polite noise. The ones who mattered were desperate: they needed a solution right now and could explain to their boss why.
The questions that separate the two groups are blunt. If we build this product, would you be willing to buy it? What's pushing you to buy something now? How would you articulate the business case of buying our product to your boss? At Harness he formalized this into the "3 Whys" (why buy anything, why buy now, why buy us) and a minimum of fifty "cold hard conversations" with people outside his network, because friendly contacts give friendly, useless feedback.
The insight is not that founders should talk to customers. Everyone says that. It is that the distribution of urgency among prospects is the most valuable data a founder has, and finding it requires going wide first and then narrowing hard. Harness's urgency turned up among mid-sized companies moving to Kubernetes, not among the large banks it had assumed. Traceable's turned up in large financial and healthcare enterprises, not among mid-sized cloud-native firms. Each time, the only way to locate it was volume.
Tactic: Send enough cold outreach to get at least fifty real conversations, then rank every respondent by how urgent their problem is, and build only for the top tier.
Principle 3
There is no prize for being eventually right
Bansal used almost the same sentence about two companies sixteen years apart. In 2008 he was correct that everyone would move to the cloud, but wrong about when, and "there's no funding for being right eventually." In 2020 at Traceable, he was eventually right about many of his assumptions, but "there's no prize for being 'eventually right.'" His company-building philosophy comes down to keeping the urge to be right in its place.
This is a sharper point than "pivot fast." His ideas were usually directionally correct. API protection did matter. The cloud did arrive. The trap for a technical founder with good taste is mistaking a correct long-run thesis for a correct product today. AppDynamics survived by cutting its ambitions (on-premise, no architecture changes, "find it" before "fix it") until the product met customers where they were, not where they would be.
✕
Assumptions vs. Reality Across Three Companies
Bansal's own accounting of what he got wrong at the start.
| Company | Initial assumption | What customers revealed | Category |
|---|---|---|---|
| AppDynamics | Everyone on the cloud in 2–3 years | Much longer timeline; serve on-premise now | Timing |
| AppDynamics | SaaS; customers adapt to us | On-premise; no architecture changes | Delivery |
| AppDynamics | Full find-to-fix platform | Customers only wanted "find it" | Scope |
| Harness | Large enterprises first | Urgency in mid-sized firms adopting Kubernetes | Customer |
| Traceable | API protection | API discovery first | Scope |
| Traceable | Mid-size cloud-native buyers | Large finance and healthcare enterprises | Customer |
| Traceable | Product-led growth | Classic enterprise security sales | Delivery |
The table is humbling because it covers a founder who had already succeeded. Experience did not make him less wrong at the start. It made him faster to admit it.
Tactic: Write down your three founding assumptions (what, who, how) on day one, and schedule a deliberate review against customer evidence every ninety days, assuming at least one is wrong.
Principle 4
Learn sales the way you learned to code
Bansal began as a pure believer that the best product wins. A few million in revenue from accounts like Netflix seemed to prove it, until a fellow CEO told him that crossing $100 million would require getting "scientific about sales." He took "scientific" literally. He studied companies three to five years ahead of his. He brought in John McMahon for weekly whiteboard sessions. He built a Business Value Assessment to put a dollar figure on customer outcomes. He hired a VP of sales, a choice he calls one of the best he ever made.
His framing reaches well beyond his own story: "If a software engineer can learn programming, they can learn sales." The point is not that founders should become salespeople. It is that sales is a system that can be learned, with inputs, outputs, and failure modes, not a personality trait, and that treating it as a talent rather than a discipline leaves engineering-led companies permanently weak at distribution. His later vocabulary of "product-market-sales fit" and the "three-layer cake" of product, go-to-market, and operations is the result of that education.
He also learned the limits of the lesson. At Harness he installed a systematic sales machine too early, before the hustle stage was over. Founder-led selling, he now says, has to continue until the first twenty to twenty-five customers are real success stories.
Tactic: Find the company in your category that is three to five years ahead, study exactly how it hires and runs its sales team, and copy the structure, not the charisma.
Principle 5
Play only the level you are on
Bansal thinks about company-building as a video game: level zero is funding, then product-market fit, then an early sales machine, then the first executive hires. "The biggest mistake entrepreneurs make," he wrote, "is spending time on skills they need for level 3 when they haven't even passed level 1."
This can sound like ordinary prioritization advice. It is really a defense against a founder's constant anxiety about everything he doesn't yet know. New founders, he observes, feel they "need to learn them all NOW." The level framework gives permission to stay ignorant about everything except the next milestone. It also explains why he rejects simple explanations of success, the "two hundred small things" rather than two big ones. Each level has its own small set of things to get right.
His first entrepreneurial lesson, listed on LinkedIn, is the same idea with no metaphor: "Don't overcomplicate things. Be laser-focused on getting to the next milestone."
Tactic: Write down which level your company is on right now, list only the three skills needed to clear it, and decline every learning opportunity that isn't on that list.
Principle 6
Ship a minimum sellable product
Bansal dislikes the phrase "minimum viable product" for any company entering a market with an incumbent. He prefers "minimum sellable product": something that includes all the features of the existing solution plus capabilities the incumbent can't match, because a buyer comparing products won't pay to switch to something that does less.
This builds on what he calls the "binary differentiator" rule for launching into crowded markets. A new product needs a difference so clear that a buyer can tell yes or no whether it's there, not a slight improvement. Inside Harness he asks product managers a version of this every time: "So what? What's in it for the customer?" A feature that is merely "better than what exists today" doesn't pass.
The principle explains how Harness can launch new modules into categories that already have strong vendors. Each module has to be complete at entry and clearly different, or it doesn't launch.
Tactic: Before building, list every capability the incumbent offers and then the one difference a buyer could verify in a demo; if you can't name that difference, you aren't ready to build.
Principle 7
Mine your last company's customers for your next idea
Harness was not discovered by searching for an idea. It came from listening to AppDynamics customers complain about something next to the problem AppDynamics solved. They were happy with monitoring and unhappy with delivery. Banks with 10,000 engineers had hundreds writing automation scripts and babysitting deployments. That was a problem visible only to someone already inside those accounts.
Then the customer overlap became the strategy. When executives began asking to save fifteen minutes of a Harness meeting for Traceable, and when 70 percent of Traceable's largest customers turned out to be Harness customers, the merger mostly wrote itself. Bansal's companies are related products sold to an overlapping base of buyers he already knows, which is why his "startups within a startup" model works. Each new module enters accounts where trust already exists.
This is the main advantage serial founders have, and most underuse it. They treat each company as a fresh start instead of a continuation of relationships they already have.
Tactic: Ask your ten best customers which problem next to the one you solve is costing them the most, and treat the most common answer as your next product hypothesis.
Principle 8
Price the exit in other people's risk — and your own regret
The Cisco decision is a lesson in two parts. The first part is the reasoning, which held up. Bansal weighed strategic fit, cultural home, and financial outcome. He estimated that matching $3.7 billion as a public company would take three to four years of strong execution, and concluded that selling "reduced that risk for all the employees." Hundreds of people got life-changing outcomes, and Cisco built AppDynamics into a billion-dollar line of revenue.
The second part is what the reasoning left out: the founder's own unfinished work. "I also felt like we didn't fully finish what we could've." The sale was right for nearly every stockholder and wrong for one, and that one went on to spend the next nine years trying to reach the bell again. Asked whether he would take a similar offer for Harness, he said he almost certainly would not.
$
The Cisco Decision, Reconstructed
Bansal's three stated factors, plus the one he discovered afterward.
| Factor | His reasoning | Verdict in hindsight |
|---|---|---|
| Strategic fit | Access to a bigger platform and Cisco's customer base | Held up |
| Cultural home | Would employees be treated well and AppDynamics kept intact? | Held up |
| Financial risk | $3.7B now vs. 3–4 years of public execution to match it | Held up |
| Founder's unfinished work | Not weighed at the time | The regret |
The general lesson is not "never sell." It is that the decision has a fourth column, and founders should fill it in before the offer arrives, not on the walk home.
Tactic: Long before any acquisition conversation, write down what "finishing the job" means to you in concrete terms, so that when an offer comes you are weighing it against that and not only against the IPO math.
Principle 9
Multitask in aggregate, never in the moment
The CEO of two companies says his secret is not multitasking. At AppDynamics he answered email in meetings. Now he blocks his day into thirty- or sixty-minute segments, each devoted to one company and one kind of work: engineers, then the board, then a customer. "So you do multitasking on an aggregate, but not at any one moment."
This is the operational foundation for everything else in Part II. The startup studio, the venture firm, and two simultaneous CEO roles would have been impossible for a scattered person. Steve Harrick's willingness to back the arrangement depended on Bansal's ability to "hire well and delegate well." The time-blocking is what lets the delegation work, because every block gets his full attention, so short blocks are enough.
There is a lesson here about scale. Doing more total work doesn't require doing more things at once. It requires breaking work into smaller sequential pieces and switching cleanly between them.
Tactic: For one week, schedule your day in thirty-minute single-purpose blocks with your phone out of reach, and notice which meetings shrink once they get your full attention.
Principle 10
Run a venture portfolio inside the company
Harness runs sixteen product lines as "startups within a startup." New ones begin with four or five people in an ideation stage, two or three launch a year, and resources are released in stages only when early signals justify it. Bansal describes his own company as having "stealth, seed stage, early stage and late stage pieces." It is a venture portfolio with one shared sales force and one balance sheet.
This structure is where everything he did after the exit ends up. BIG Labs taught him how to build parallel companies, and Unusual Ventures taught him how to judge early-stage bets. The Harness-Traceable merger convinced him the separate-company overhead had been unnecessary: "If I was starting Traceable now, I would have just launched it inside Harness." The structure also keeps a founder who loves the earliest stage engaged inside a company of more than a thousand people. He gets to start something new two or three times a year without leaving.
The open-ecosystem design matters too. Each module works with competitors' products, and customers can buy any one alone. That keeps every internal startup honest. It has to win each sale on its merits, not ride along on a bundle.
Tactic: Set aside a fixed share of engineering capacity for tiny, time-boxed teams exploring new products, and release headcount to each one only in stages tied to customer evidence.
Principle 11
Hire hunger, not trophies
One of the most counterintuitive things Bansal has said came a year after the Cisco sale, in Bangalore: "I will try and avoid hiring people who have been part of big exits." The reasoning, as FactorDaily's Pankaj Mishra summarized it, is that exits bring "a certain comfort, risk averse nature." Bansal wanted, in his own words, to "surround myself with people who are scrappier and hungrier than me."
It's a strange rule from a man whose own exit made hundreds of millionaires, including himself. He seems to have applied it to his own case first. His identity crisis after the sale showed him how comfort can dull ambition, and he concluded that the cure was more work, not rest. As an investor he looks for the same quality. Beyond the problem size, the team's edge, and validation, he asks whether founders are "committed for the long haul," with "ambition, not just a desire for a quick exit."
This isn't a blanket ban on experience. He relied heavily on experienced sales leaders and advisors. It is a way of telling apart people who have done the work from people who have been paid for it and are now mainly protecting what they have.
Tactic: When interviewing senior candidates with big exits behind them, ask directly what they still need to prove, and weigh the specificity of the answer more heavily than the résumé.
Principle 12
Choose qualified soft advisors and topic-specific mentors
Bansal's approach to advice is unusually methodical. For any unfamiliar problem, he applies a "rule of 10": ask ten smart people. He prefers "topic-based mentorship, not generic mentorship." He tried a CEO coach, and it "didn't work very well for me." For venture board members he uses a two-by-two grid: qualified or unqualified, soft (gives advice and trusts you to decide) or hard (insists you follow it). The best is the qualified soft advisor. The one to avoid, especially when you have given up real control, is the unqualified hard advisor.
His test is simple: "would you hire this person as an independent advisor to you as the CEO even if they were not an investor in your company?"
?
Bansal's Advisor Matrix
How to classify the people who tell you what to do.
| Type | Behavior | Bansal's ranking |
|---|---|---|
| Qualified soft advisor | Has done the job; offers advice, trusts your call | Best |
| Unqualified soft advisor | Hasn't done the job; offers advice, trusts your call | Common, acceptable |
| Qualified hard advisor | Has done the job; insists you follow advice | Manageable |
| Unqualified hard advisor | Hasn't done the job; insists you follow advice | Avoid |
He holds himself to the same standard he sets for investors. He has written about having to suppress the urge to picture himself running the companies Unusual backs: "you can't assess a startup based on what you would do." He is describing, from the other side of the table, how not to become an unqualified hard advisor.
Tactic: Before your next financing, put every prospective board member through the hiring test: if they weren't investing, would you pay them for advice? If not, limit their control accordingly.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Jyoti Bansal, to Fortune, 2024I remember being at the party after, celebrating, we'd sold the company for a pretty good price, and that was a high point. Then I came home, and I was depressed, thinking, 'This is the end.'
— Jyoti Bansal, to CNBC Make It, 2024When I started [AppDynamics], I didn't think about a financial outcome. I just thought in terms of: This problem needs to be solved.
— Jyoti Bansal, to Forbes, 2025People say, 'Once I retire, I'm going to do what I enjoy.' I asked myself, 'Do I enjoy playing golf all the time or being on the beach all the time?' I don't really. I realized why not just go back to what I enjoy, building a company.
— Jyoti Bansal, on LinkedIn, 2025Eventually, we were able to make Traceable a success and years later, many assumptions proved true — but there's no prize for being "eventually right."
— Jyoti Bansal, to CNBC Make It, 2024It's very ironic that people think, [because] I'm CEO of two companies, I should be the best at multitasking. Actually, my cheat code is to not do too much multitasking.
Maxims
- Use the wait. Years spent unable to build are years for studying how other people's companies fail.
- Sort the desperate from the curious. Only prospects who need a solution now give you feedback worth acting on.
- Directionally right is still wrong. A correct thesis about the future does not fund a company today.
- Sales is a system you can learn. If an engineer can learn programming, an engineer can learn to sell.
- Clear the current level. Skills for level three are a distraction until you have passed level one.
- Enter crowded markets complete. Match the incumbent, then add the one difference a buyer can verify.
- Your last customers know your next idea. The problem next to the one you solve is the cheapest one to find.
- Fill in the fourth column. Before any exit, decide what finishing the job means to you.
- Focus in pieces. You can do many things in a day as long as you do them one at a time.
- Ask whether you'd hire the advisor. If you wouldn't pay for someone's advice, don't give them control.
