Rujul Zaparde on Building Zip, Hiring for Potential and Moving Fast

Rujul Zaparde remembers constantly working on small businesses with friends in middle and high school. They made rudimentary video games and sold them to students at other schools, then tried selling bumper stickers that let people report poor driving by teenagers—even though the founders were too young to drive.

Asked whether founders should be embarrassed by their first version, Zaparde says they learn more by putting something out. The first idea is probably wrong in ways they cannot see yet. Good service, better questions and fast iteration can compensate for missing features and turn early customers into advocates.

Zip began with a problem Zaparde and his co-founder Lou experienced at Airbnb, where he was a product manager and Lou was an engineering leader. A colleague left, and they were assigned ownership of a large software contract coming up for renewal.

Procurement asked them to create a purchase requisition in the ERP system. They had never been told their cost center and did not know what a general ledger code was, yet had to choose from thousands of options. Submitting the request turned out to be only the beginning.

They also had to upload the contract for legal review and file separate IT, security and privacy requests. A month or two later, it was hard to see where the renewal stood or what was blocking it. Even their procurement contact had to ask other teams and get back to them later. They thought there had to be a better way.

Harry Stebbings asks why they entered a market crowded with procurement and spend-management vendors. Zaparde says competition also signals a large market with real pain: they did not have to bet that the problem existed. The question was whether they had a distinctive insight into it.

His insight was that purchasing had become more distributed. When a small IT department chooses all the software, people can be trained on a specialized system. When thousands of employees initiate purchases, the experience has to work for people who do not know procurement terminology.

This does not mean spend decisions are entirely decentralized. Zaparde distinguishes decentralized initiation from centralized review: a company can still require a CFO to approve purchases above a chosen threshold. As budgets tightened, he saw many customers lower those thresholds, sometimes dramatically.

One customer champion told him that a CFO at a company with roughly 3,500 employees temporarily lowered the approval threshold to $5,000 to see more spending. The volume became too much, and the threshold rose again. Zaparde sees value in making those rules easy to change as circumstances evolve.

In a crowded market, positioning is an education problem. Zaparde says founders have to accept that many prospects will not understand the new approach yet, and distinguish early adopters from people who may buy later. Trying to persuade everyone at once is not necessary.

On whether to start with a narrow feature or a platform, Zaparde favors a single acute pain point. A young company has little capacity, which forces it to identify the thing someone urgently needs and cannot get elsewhere. Even as the product expands, that initial wedge can remain the reason its platform stands apart.

Zip's positioning took around six or seven months to clarify. The founders wrote separate short pitches for investors and customers, initially revisiting them every day, then every Sunday. They incorporated what they had learned from calls and tested the revised language in subsequent conversations.

They also built a spreadsheet of possible ideal customer profiles: IT, legal, procurement and finance all had a role in approving purchases. Their first handful of customers were in IT, but procurement and finance eventually became the core buyers. Early demand was useful evidence, not a definitive answer about the best customer.

Zaparde says a startup must sometimes turn down or even fire a customer. When Zip had about 12 to 15 employees, it signed a pilot with a company employing hundreds of thousands. During implementation, he realized making that customer successful would consume an excessive share of the roadmap.

He told the customer champion that Zip could not serve the company adequately at that stage. The champion appreciated the candor and asked to hear from them when they were ready; Zip contacted the company again roughly a year later. Stebbings notes how difficult such a decision can be when a large name offers credibility with investors and other buyers.

To test whether a problem is worth solving, Zaparde looks for repeatability: can the same narrative reach the same kind of person and consistently produce a similar result? Eventually other people must be able to repeat the process, even if the quality falls somewhat as it scales.

He is skeptical of free design partners as a substitute for evidence of demand. He and Lou decided their first ten customers would come from cold LinkedIn outreach, with no personal relationships, warm introductions or referrals. Each would pay real money, even if early prices were modest and increased later.

His second company brought a different mindset. In the first, he says, it was easy to worry about investor perceptions and whether every metric looked good. With Zip, the founders wanted to build a lasting business and preferred to discover a bad idea quickly rather than protect its appearance.

Another danger is customizing the product slightly for every early customer. With only a few customers, each request feels important, but five different versions make it hard to prioritize features or explain the product. Zaparde wants a sharp, repeatable proposition that the company can take to the market.

Stebbings argues that product-market fit can disappear when a company changes its channel, audience or price. Zaparde agrees that it is not binary: he sees a matrix of products, personas, channels and industries. The first successful cell does not prove that a new product or a new market will fit as well.

For example, success with technology companies does not establish fit in financial services. Stebbings asks whether sales should be organized by industry to sharpen the pitch. Zaparde says Zip had not done so, partly because a fully verticalized sales organization can be very expensive at its stage.

He agrees that a resourceful zero-to-one operator could bootstrap a promising vertical: learn its systems, develop the message and create materials that help the sales team succeed. The company might then fold that knowledge into the broader organization, keeping a dedicated group only where the product and market require it.

Building a sales organization was a learning curve for the founders, who came from consumer businesses. Zaparde initially questioned the need for sales engineers, assuming an account executive could handle demonstrations. Technical questions exposed the limits of that assumption and showed why specialized help mattered.

Sales metrics are useful, but Zaparde says understanding a problem often requires listening to the actual customer call. Confidence, for instance, matters enormously when a salesperson speaks to a prospect, but is difficult to measure or repair with a dashboard. Better training and more practice give people the knowledge to answer confidently.

Zip hired a salesperson before a sales leader. The new hire shadowed Zaparde, then took over the initial presentation while Zaparde continued to demonstrate the product; eventually the salesperson handled demos too. It was a gradual transfer built on trust, not a handoff to someone with a complete playbook.

The trigger for delegating was that Zaparde was no longer learning much from first calls: he was conducting them mainly to win deals. At a company of four or five people, little was documented, so shadowing and repeated practice were essential. Some founder friends hire two reps to create a comparison, though Zip did not.

Zaparde strongly favors speed as a startup advantage. Given a choice between a slower team with a thoughtful plan and a faster team willing to change course as it learns, he would generally bet on the latter. Unless the original plan is exactly right, rapid iteration offers more chances to find the destination.

Zip sometimes imagines competing against an identical version of itself in an alternate reality; the answer to winning is often to iterate faster. Zaparde acknowledges that the company has slowed as it has grown. Founders can still set a high initial pace, lead by example and publicly reward fast action and learning, including when an attempt fails.

Stebbings asks how a repeatable process can coexist with creativity. Zaparde suggests encouraging people to speak up, rewarding unconventional ideas and valuing first-principles thinking. He often prefers a high-potential person who has not done the exact job before, because that person must work through the problem instead of simply applying an inherited playbook.

He does not rule out experienced hires or existing playbooks; the balance depends on the situation. His concern is that relying exclusively on them can reduce originality. Developing leaders from within the company can also help preserve its way of thinking as it grows.

For a difficult problem, Zip gathers a small, cross-functional group for an hour. Participants spend the first 15 minutes writing ideas independently, then share them and narrow the list. The goal is to leave with a handful of possibilities, including some the group might not have reached without the exercise.

Focus also depends on the function. Zaparde thinks marketing execution can explore a wide range of activities, from growth to content, while the message presented to the market still needs to be coherent. He does not believe every function must be equally focused in how it experiments.

Speed can be misapplied. For his first company, FlightCar, Zaparde and a friend agreed to meet for an hour and pursue the best idea they could devise. Inspired by Airbnb, they chose to offer free airport parking by renting travelers' cars to others, and then spent five years working on it.

He would not choose a company idea on such a short deadline again, although he does not regret what he learned. FlightCar became a demanding operational business, with 17 airport locations and hundreds of hourly workers at its peak, washing cars and moving customers. That experience made the relative simplicity of a B2B software business attractive.

He compares FlightCar's margins to the last drop squeezed from a lemon: mistakes earlier in the process could wipe out the return. Later, working at Airbnb helped him see what excellent designers and engineers looked like. As a first-time founder without that reference point, he had been learning the talent bar by feel.

Stebbings favors repeat founders because they have seen hiring and other mistakes firsthand. Zaparde adds that first-time founders may be more willing to take fundamental market risk, partly because they do not yet know what is hard. That willingness can produce exceptional outcomes, even as experienced founders benefit from their past lessons.

With Zip, he and Lou deliberately chose high execution risk with lower fundamental market and product risk. He offers Snap as an example of a company that took substantial market risk and succeeded, and wonders whether a more seasoned founder might have been too cautious to try it.

Asked to answer the criticism that experienced entrepreneurs waste their talents on enterprise software, Zaparde argues that improving procurement can create value across a huge volume of business-to-business transactions. Those efficiencies can eventually reach consumers, although the connection is less immediately visible than in a consumer product.

At Airbnb, Zaparde watched Brian Chesky pay close attention to details in design reviews, including spacing in the mobile app. A few pixels alone may not matter much, he says, but that attention communicates the company's standard for product quality. Like speed, the bar set by founders becomes part of the culture.

The way to combine detail and speed is selective micromanagement. Founders cannot inspect everything, so Zaparde suggests choosing a few things that embody the company's core values. He spends considerable time on sales calls and what specific people say and do, rather than trying to manage the entire go-to-market organization through metrics alone.

His job, as he sees it, is to focus on what is not working. An area that currently operates well may eventually break, but it need not receive the same immediate attention as a visible problem. One live challenge was scaling business development while maintaining productivity and a career path toward account executive roles.

Despite its brand, Zip still acquired most new customers through outbound sales at the time of the interview, though inbound had grown. Enterprise selling often requires finding the right person inside a large organization, and titles can vary by company and country. Stebbings worries AI-generated volume could overwhelm outbound channels.

Zaparde is less convinced that outbound will disappear, although he cannot predict the distant future. He sees an opportunity to use tools such as ChatGPT to automate manual work and increase the team's leverage, while treating the outcome as something to reassess after more experience.

Zip was founded during the pandemic, making the return to shared offices complicated. Its engineering, product and design staff were required to come in on Mondays and Thursdays. The team tried to concentrate meetings on those days, leaving more uninterrupted time elsewhere in the week while addressing employees' desire for social contact.

Early in the pandemic, Zaparde and Lou kept a video call open all day while working from different locations. As the company grew, employees joined the same room, eventually with around 20 people. Zaparde left himself unmuted during sales calls, so engineers and others could hear customers and learn directly from the conversations.

His biggest hiring mistake has been putting too much weight on someone having held a similar role at a similar scale. A high-profile executive may succeed, but every company differs and the old playbook may not transfer. Zaparde would often rather back someone who has proved capable and has room to grow into a bigger role.

That bet requires an honest conversation: both company and employee are taking a risk, and both should say if the arrangement is not working. It may attract less external recognition than a famous hire, he says, but can be the better decision more often than people assume.

Founders are often late to recognize a bad hire because the person's team and peers may have seen the problem for months. Zaparde seeks signals from trusted, high-performing people across the organization. Multiple independent reports are a serious warning, though any individual assessment can be wrong or nuanced.

He disagrees that a performance improvement plan is always pointless. If a plan never works, he suggests, the company may be starting it too late. Detecting a problem earlier and making expectations clear can give someone a real opportunity to correct course.

Raising a lot of money did not mean Zip had to hire immediately. Zaparde believes founders often expand before establishing even initial product-market fit. A small team can candidly change direction when an idea is not working; with 20 people who have joined on the strength of the original vision, that becomes harder.

Zip remained disciplined about headcount regardless of cash in the bank and was cash-flow positive at one point while still small. The company accelerated hiring after gaining more confidence in 2021. On valuation, Zaparde says the founders believed the market need and growth justified using the capital to pursue opportunities sooner.

Zaparde says he wants to lead Zip for decades. When the company began in 2020, he and Lou wanted it still to be known in 2040. Stebbings says that sense of the company as a founder's life's work is a strong signal to him as an investor.

In the quick-fire round, Zaparde describes changing how he organizes his day. An early shared spreadsheet where everyone listed daily tasks broke down as the team grew. He now writes his own tasks on a card, keeps it in his pocket, crosses things off and starts a fresh card the next day.

His concern about generative AI is less an immediate collapse in enterprise jobs than malicious use, such as imitating someone's voice for fraud. He agrees that enterprise adoption can be slow, but worries that people with harmful intent may adopt the tools quickly. He does not claim to have a solution.

Founder and investor incentives can diverge because of their different time horizons and exposure. A venture investor has a portfolio and does not need every company to succeed in the same way, while a founder has committed a major part of their life to one company. Zaparde describes this as a structural difference rather than bad intent.

He would choose Frank Slootman for his board. The advice he finds misleading is that once a company has product-market fit, the rest is simply a race to scale. He also criticizes the habit, common in the low-interest-rate era, of solving every problem by hiring another person before considering alternatives.

What he wishes he had known before starting Zip is how hard it would be to build a world-class product in procurement. Some enterprise customers have over a thousand possible questions, conditional logic and more than 20 kinds of risk review. Even compared with the search systems he worked on at Airbnb, he found the domain remarkably complex.

Knowing that complexity might have led Zip to invest earlier in architecture that made its platform easier to extend. But Zaparde immediately qualifies the hindsight: building that architecture earlier might have taken too long and been the wrong tradeoff when speed mattered. Decisions have to be made with the information available at the time.

Looking toward 2044, Zaparde expects still to be working on procurement at Zip. He estimates that well over 95% of approval requests are approved and sees an opportunity to automate much of that routine review. He stresses that the technology is not there yet, but imagines a future in which software handles many approvals without a person reading and approving each one.