Onsite Pro, incorporated as Greenwork Inc. in March 2021, was a San Francisco-based startup that passed through Y Combinator's Summer 2021 batch. Over roughly three years, the company executed two full pivots: first from a clean energy w…
Onsite Pro, incorporated as Greenwork Inc. in March 2021, was a San Francisco-based startup that passed through Y Combinator's Summer 2021 batch.[1][2] Over roughly three years, the company executed two full pivots: first from a clean energy workforce training platform ("a Coursera for solar") to a hiring marketplace connecting trade schools with clean energy employers, then from that marketplace to HVAC digital proposal and pricebook software sold under the Onsite Pro brand.[3] The company raised approximately $2.4 million in pre-seed funding from GV, YC, Kleiner Perkins Scout Fund, and Global Founders Capital — all of it before November 2021.[4]
The company failed to achieve durable product-market fit across either pivot before exhausting its runway. Two pivots in under three years on $2.4 million left insufficient time and capital to validate each thesis; the final HVAC software bet was still in early stages when co-founder and CEO Sam Steyer stepped down citing personal fatigue in November 2023.
YC now lists Onsite Pro as "Inactive" with a team of four.[5] No acquisition or formal shutdown announcement has been published. Steyer returned to climate tech at Halcyon; Gautam Jayaraman, who became CEO at the end, has not publicly commented on the outcome.
Sam Steyer and Gautam Jayaraman incorporated Greenwork Inc. on March 31, 2021.[6] The two co-founders brought complementary but unusual credentials for a trades-focused startup.
Steyer held an A.B. in Applied Mathematics from Harvard and an M.S. from Stanford.[7] Before Greenwork, he had co-founded Station A, a clean energy analytics platform, where he served as Head of Analytics. He also worked as a surrogate speaker on Tom Steyer's 2020 presidential campaign — an experience that embedded him in climate policy circles and gave him a macro view of the clean energy transition's workforce bottleneck. That bottleneck became the founding insight: the US was building out solar, EV, and HVAC infrastructure faster than it was training the workers to install and maintain it.
Jayaraman's background was almost entirely inside the YC ecosystem. He held M.Eng. and S.B. degrees in Computer Science from MIT and had co-founded two prior YC companies: Jamglue (YC S06, an online music remixing platform) and Rickshaw (YC W14, a commuter shuttle service). Between those ventures, he worked as an engineer at Amazon and Dropbox, and as an Engineering Manager at DoorDash Labs.[8] Jayaraman later noted with some pride: "since 2006, I've worked at five YC-backed companies, and zero non-YC-backed companies."[9] That depth of YC-network embeddedness likely helped the company secure a strong pre-seed syndicate quickly, but it also meant neither founder had deep roots in the trades industry they were entering.
The founding thesis was straightforward: the clean energy transition required a massive skilled-trades workforce, but the pipeline connecting training programs to employers was broken. Greenwork would fix it by building what Steyer called "a Coursera for solar" — a platform where trade schools and workforce development programs could upload curricula and connect graduates directly to clean energy employers.[10] The timing felt right: the Biden administration's infrastructure bill was moving through Congress in 2021, and the policy environment was explicitly focused on workforce development for clean energy.
The company entered YC S21 with this training concept. Within roughly a year, Steyer acknowledged the model hadn't held up: "It's been a humbling experience. Not all of our assumptions were correct and we've had an incredibly steep learning curve over this first year."[11] The specific failure mode of the training model — whether it was supply-side (trade schools unwilling to pay or participate), demand-side (employers preferring existing channels), or structural (the market wasn't ready to pay for a digital training intermediary) — was never publicly disclosed.
What followed was a rapid repositioning, and then another. The company's story became less about its original climate mission and more about the founders' willingness to chase product-market fit wherever the data pointed — even if that meant abandoning the mission that had attracted their investors.
Onsite Pro's final product was a B2B web application targeting the HVAC distribution channel — a layer of the industry that sits between equipment manufacturers (like Carrier, Lennox, or Goodman) and the independent contractors who install and service HVAC systems in homes and commercial buildings.
The core problem it addressed: HVAC contractors have historically relied on paper catalogs, phone calls to distributors, and manual spreadsheets to build customer proposals. A typical residential HVAC replacement proposal — which must account for equipment specifications, pricing, available rebates, financing options, and matched system components — could take hours to assemble. The result was slow sales cycles, inconsistent pricing, and lost deals when contractors couldn't produce a professional quote on-site.
How the product worked: Onsite Pro gave HVAC distributors a digital pricebook preloaded with their product catalog, pricing, availability, and AHRI (Air-Conditioning, Heating, and Refrigeration Institute) system matchups — the regulatory certifications required for matched equipment combinations.[24] Contractors who worked with those distributors could then access the pricebook through a web application (designed to run on a laptop or tablet) and use a step-by-step proposal builder that pulled in product data, current pricing, applicable utility incentives, and financing options from partners like FTL Finance.[25]
The stated goal was a professional, customer-ready proposal in under ten minutes — compared to the hours a contractor might otherwise spend.[26]
The go-to-market wedge: The distributor was the paying customer, not the contractor. Onsite Pro charged distributors a flat fee per pricebook — "one low cost for as many users as you need within a single pricebook" — and contractors used the tool for free.[27] This B2B2C structure was deliberate: a single distributor relationship could unlock dozens or hundreds of contractor users, making the unit economics theoretically attractive if distributors could be convinced to pay.
Domain-building signals: The team joined HARDI (Heating, Air-conditioning & Refrigeration Distributors International), the primary trade association for HVAC wholesalers, signaling an attempt to embed in the industry's institutional infrastructure.[28] They also recruited advisors with direct HVAC industry experience, including a sales professional whose prior HVAC software had been acquired by Goodman (one of the largest HVAC manufacturers) and a finance executive from FTL Finance.[29]
Technology stack: The product was built on Node.js with a Next.js frontend, deployed via Vercel, with Segment for analytics and NextAuth.js for authentication.[30] This was a modern, lightweight web stack — appropriate for a small team building a data-driven SaaS tool, but not unusual or proprietary.
What distinguished it from alternatives: Existing HVAC software (ServiceTitan, Jobber, FieldEdge) focused primarily on field service management — scheduling, dispatch, invoicing — rather than the pre-sale proposal process. Onsite Pro's specific focus on the proposal-and-pricing workflow, integrated with distributor pricebooks and AHRI data, was a narrower and more specific wedge. The closest analog was Contractor Commerce, which also targeted the digital sales layer for HVAC contractors, but approached it from the contractor side rather than the distributor side.
Onsite Pro's primary paying customers were HVAC distributors — regional wholesale businesses that purchase equipment from manufacturers and resell it to independent HVAC contractors. The US has thousands of such distributors, ranging from single-location independents to regional chains. The end users of the product were the contractors themselves: typically small businesses (2–20 employees) that install and service residential and light commercial HVAC systems.
The distributor-as-customer framing was strategically important. Distributors already had financial relationships with contractors and a direct incentive to help those contractors sell more equipment — more contractor sales meant more distributor revenue. Onsite Pro was positioning itself as a tool that distributors could offer to their contractor networks as a value-added service, rather than asking individual contractors to pay for yet another software subscription.
The US HVAC market is large. The residential HVAC replacement market alone generates over $15 billion in annual equipment sales, with an estimated 5–6 million replacement systems installed per year. HVAC distributors collectively handle the majority of that equipment flow. The addressable market for distributor-facing software is harder to quantify, but the number of HVAC distributors in the US is estimated at several thousand locations across hundreds of companies — a fragmented market with no dominant software incumbent specifically serving the proposal workflow.
The broader field service software market (which includes HVAC as a major vertical) was valued at approximately $3 billion in 2021 and growing, driven by the digitization of the trades. However, Onsite Pro was targeting a specific sub-segment of that market — the pre-sale proposal layer, accessed through distributors — which was meaningfully smaller than the total field service software opportunity.
Onsite Pro competed on two dimensions simultaneously: product depth (a specialized proposal tool with AHRI data and distributor pricing integration) and distribution reach (access to contractors via distributor relationships). This dual-axis positioning created structural vulnerabilities on both sides.
From the field service management incumbents: ServiceTitan, the dominant HVAC software platform with over $500 million raised and tens of thousands of contractor customers, had both the distribution reach and the financial resources to add proposal features natively. Any feature that Onsite Pro built could theoretically be absorbed by ServiceTitan as a module — and ServiceTitan's existing contractor relationships gave it a distribution advantage that Onsite Pro, selling through distributors, could not easily replicate. The risk of being a feature rather than a platform was structural, not just competitive.
From manufacturer-owned tools: Goodman, one of the largest HVAC equipment manufacturers, had already acquired at least one HVAC sales software company (as noted by Onsite Pro's own advisors).[31] Manufacturers have a direct incentive to provide proposal tools that favor their own equipment — and the resources to offer those tools for free or at subsidized cost to distributors and contractors. Competing against a free tool backed by a manufacturer's equipment margin is a difficult position for an independent SaaS company.
From the relationship layer: Perhaps most importantly, HVAC distribution is a relationship-driven business. Distributors choose software vendors the way they choose equipment lines — based on trust, long-term relationships, and peer recommendations within a tight industry network. A two-year-old startup with no prior HVAC history, entering the market via a pivot from a clean energy hiring marketplace, faced a credibility gap that HARDI membership and advisor recruitment could only partially address. The sales cycle for distributor software is long, and the switching costs for distributors (who must migrate pricing data, train contractor users, and manage the transition) are high — meaning the first mover with a distributor relationship has a durable advantage.
Onsite Pro charged HVAC distributors a flat fee per pricebook, with unlimited users per pricebook included in the price.[32] The contractor-facing product was free, consistent with a B2B2C model where the distributor subsidizes contractor access as a value-added service.
The company never disclosed revenue publicly. The absence of any ARR announcement, press coverage of customer wins, or investor update citing growth metrics is itself a signal — companies with strong early traction in B2B SaaS typically publicize it.
The only third-party revenue estimate available is a low-confidence LeadIQ figure of approximately $750K in annual revenue as of June 2025.[33] This figure is internally inconsistent with YC's "Inactive / team of 4" listing and should be treated as unreliable — LeadIQ's revenue estimates for small private companies are algorithmically generated and frequently stale.
Inferred burn rate (labeled as estimate): With approximately $2.4 million raised and a team that LeadIQ estimated at 12 employees at peak,[34] a rough burn estimate would be $80K–$120K per month at full headcount (assuming San Francisco-area engineering salaries and standard startup overhead). At that rate, $2.4 million would support 20–30 months of operations — consistent with the November 2021 funding date and a late 2023 / early 2024 runway exhaustion. This is an inference, not a confirmed figure.
The hiring marketplace phase used a subscription tier model charged to employer-side customers, with the supply side (trade schools) onboarded for free.[35] No revenue figures from that phase were disclosed.
The most concrete traction data available is from the hiring marketplace phase (early 2022): approximately 500 job seekers on the platform and "tens of companies recruiting for more than 100 jobs" across solar installation, HVAC, EV, and advanced manufacturing.[36] These numbers suggest early-stage marketplace activity but not meaningful scale — 500 job seekers and tens of employers is a proof-of-concept, not a defensible marketplace.
The White House recognition (representing the US at the Infrastructure Talent Pipeline Challenge celebration) and the Allume Energy partnership were meaningful credibility signals for the hiring marketplace phase, but neither translated into disclosed revenue or user growth.[37][38]
No traction data — customer counts, ARR, distributor partnerships, or contractor users — was ever disclosed for the Onsite Pro HVAC phase. The absence of any public metric from the final product is consistent with a company that did not achieve the scale needed to make those numbers a marketing asset.
The most direct explanation for Onsite Pro's failure is arithmetic. The company raised approximately $2.4 million — all of it by November 2021 — and executed two full pivots over the following two years.[39] Each pivot required rebuilding the product, reorienting the go-to-market, and re-establishing credibility with a new customer segment. At an estimated burn of $80K–$120K per month, the company had roughly 20–30 months of runway from its last funding close. That runway had to cover the wind-down of the training model, the build-out and operation of the hiring marketplace, the decision to abandon the hiring marketplace, the pivot to HVAC software, the product rebuild, the HARDI membership and advisor recruitment, and the early sales motion — all before any of those bets could compound.
The hiring marketplace phase lasted roughly 12 months (early 2022 through late 2022 or early 2023) before being abandoned. The HVAC software phase was still described as "attacking a new problem" — implying early-stage — when Steyer stepped down in November 2023.[40] That means the final pivot had at most 12–18 months of runway behind it before the company effectively ran out of both capital and leadership energy.
The team attempted to address the capital constraint implicitly by keeping headcount lean and by framing each pivot publicly as a success rather than a failure — a narrative strategy that may have been intended to preserve investor confidence and recruiting ability. But no additional funding round was raised after November 2021, suggesting the strategy did not attract follow-on capital.
The clean energy hiring marketplace was a reasonable pivot from the training model — more transactional, more immediately monetizable, and aligned with real employer demand. But the competitive landscape was structurally unfavorable. Indeed, LinkedIn, and ZipRecruiter already dominated general hiring; niche job boards for trades (like iHireConstruction or Tradesmen International) had existing networks. Greenwork's differentiation was its focus on clean energy specifically and its supply-side relationships with trade schools — a genuine but narrow moat.
The marketplace also faced a classic chicken-and-egg problem. Employers pay for access to job seekers; job seekers come for employer density. With 500 job seekers and "tens of companies," the platform was too thin on both sides to generate the match quality that would justify employer subscription fees over free alternatives. The White House recognition and the Allume Energy partnership were credibility signals, not demand signals — they did not translate into disclosed revenue or accelerating user growth.
The team's attempted remedy was to focus on the clean energy vertical specifically, where general-purpose platforms had less depth. But the clean energy trades market in 2022 was still nascent — the Inflation Reduction Act, which dramatically accelerated clean energy installation demand, was not signed until August 2022, and its workforce effects would take years to materialize. The market Greenwork was building for was real, but it was 3–5 years from the scale needed to support a standalone hiring marketplace.
The pivot to HVAC proposal software was analytically defensible — the problem was real, the market was large, and the distributor-as-customer model was clever. But the HVAC distribution market has structural characteristics that make it difficult for a new entrant to penetrate quickly.
HVAC distributors are relationship-driven businesses. They have long-standing ties to manufacturers, contractors, and software vendors. Switching costs are high: migrating a pricebook requires re-entering pricing data, retraining contractor users, and managing a transition period during which the old and new systems must coexist. A distributor evaluating Onsite Pro was not just evaluating the product — they were evaluating whether a two-year-old startup with no prior HVAC history would still exist in three years to support them.
The competitive threat from manufacturer-owned tools was particularly acute. Goodman had already acquired at least one HVAC sales software company.[41] Manufacturers can offer proposal tools for free or at subsidized cost because the tool drives equipment sales — a fundamentally different economic model than a standalone SaaS company that must charge for the software itself. Onsite Pro's flat-fee-per-pricebook model was competing against tools that manufacturers had every incentive to give away.
The team attempted to address the credibility gap by joining HARDI, recruiting HVAC-specific advisors, and building a product that integrated AHRI matchups and financing — features that required genuine domain knowledge. These were the right moves, but they take time to translate into sales. The company did not have time.
Steyer's November 2023 departure was unusually candid about its cause. "I am also, honestly, a little tired after 3 years as CEO, on the heels of a presidential campaign and another startup founding experience," he wrote.[42] This is a rare public acknowledgment of founder fatigue as a factor in a company's trajectory.
The timing was damaging. The HVAC pivot was still in early stages — Steyer himself described it as "attacking a new problem" — meaning the company was at its most vulnerable moment: a new product, a new customer segment, no proven revenue, and dwindling runway. A CEO transition at that stage forces the incoming CEO (Jayaraman) to simultaneously manage the existing team, re-establish investor relationships, and close the first meaningful customer deals — all while the clock runs down.
Jayaraman was technically capable (MIT, DoorDash, Dropbox, two prior YC companies), but his background was engineering-heavy. The HVAC software business at that stage needed a sales-intensive motion — cold calls to regional distributors, conference attendance, relationship-building with HARDI members. Whether Jayaraman had the sales orientation to execute that motion is unknown, but the outcome — YC listing the company as "Inactive" within roughly 18 months of his taking over — suggests the transition did not produce a breakthrough.
Zooming out from company-specific decisions, Onsite Pro's final product occupied a dangerous position in the software stack. Proposal generation is a workflow feature, not a platform. The dominant HVAC software platforms (ServiceTitan, Jobber) had the contractor relationships, the data, and the capital to add proposal features natively — and ServiceTitan in particular had been systematically expanding its feature set to cover more of the contractor workflow. A standalone proposal tool, however well-designed, faces the constant risk of being absorbed as a feature by a platform that already owns the contractor's daily workflow.
The distributor-as-customer model was an attempt to avoid this trap by approaching contractors through a different channel. But it created a different vulnerability: distributors are not software buyers by nature, and convincing them to pay for a tool that primarily benefits their contractor customers — rather than their own operations — is a harder sell than it appears.
Two pivots on $2.4M is a structural constraint, not just a strategic choice. Greenwork's capital efficiency was impressive in one sense — the team kept the company alive for over two years on a pre-seed round. But each pivot reset the clock on product-market fit validation. By the time Onsite Pro had a product worth selling to HVAC distributors, the company had roughly 12–18 months of runway left and no confirmed path to a Series A. Startups with thin capital reserves need to pick a thesis and stay with it long enough to generate the metrics that attract follow-on funding; Greenwork's pivot cadence made that impossible.
Climate mission and HVAC software require different founder profiles. Steyer's credibility came from climate policy networks, the Tom Steyer campaign, and Station A's clean energy analytics work. That credibility was an asset in the hiring marketplace phase — it opened White House doors and attracted mission-aligned employers. It was not an asset in HVAC distribution, where credibility comes from years of industry relationships and a track record of supporting distributors through software transitions. The team attempted to compensate with advisors and HARDI membership, but those are proxies for the relationships that incumbents had built over decades.
Distributor-as-customer models require longer sales cycles than pre-seed runway typically allows. Onsite Pro's B2B2C wedge — charge distributors, give contractors free access — was structurally sound but slow. HVAC distributors evaluate software vendors over months, not weeks, and the first distributor win is the hardest because there is no reference customer. Greenwork entered this sales motion with less than two years of runway remaining and a CEO transition mid-cycle. The model needed a Series A to fund the sales team and the time required to close the first cohort of distributors; without that capital, the motion stalled before it could generate the proof points needed to raise.
Manufacturer-subsidized competition is a ceiling, not just a competitor. The fact that Goodman had already acquired an HVAC sales software company — noted by Onsite Pro's own advisors — was a warning sign that the company appears not to have fully weighted. When a manufacturer can offer a comparable tool for free because it drives equipment sales, an independent SaaS company cannot compete on price and must compete on neutrality (supporting multiple brands) or depth (features the manufacturer tool lacks). Onsite Pro's multi-brand positioning was its natural answer, but it needed time and distributor relationships to prove that neutrality was worth paying for.
Founder fatigue is a business risk, not just a personal one. Steyer's public acknowledgment that he was "honestly, a little tired" after three years of pivots, a presidential campaign, and a prior startup is unusual in its candor — and instructive. The CEO transition happened at the worst possible moment: a new product, a new market, no revenue, and a shrinking runway. Investors and boards should treat founder energy as a finite resource that depletes faster under serial pivoting, and founders should plan for that depletion before it becomes a crisis.