A strategy implementation example from a three-sided marketplace: Jonathan Stull of Handshake

If you want a strategy implementation example with real detail attached, this is one of the better ones on record. Jonathan Stull walks through how Handshake solved a three-sided marketplace, why the company publicly promised never to charge for the thing it could most easily have charged for, and what that decision bought them years later.

Stull is COO at Handshake, which exists so that students have equal access to job opportunities regardless of where they go to school, who they know, or what they studied. He joined in 2015. Since then higher education partnerships grew from 60 to 1,400, employer partners reached 600,000 including every company in the Fortune 500, and students on the platform went from 100,000 to 10 million. The company raised an additional $200 million on top of $230 million at a $3.5 billion valuation, from investors including True Ventures, Coatue, Kleiner Perkins, Lightspeed, GGV and the Chan Zuckerberg Initiative.

Waiting tables, the 9/11 Commission, and Beijing

Stull's path is worth telling because it explains what he does now.

He had a job lined up for after college. It fell through after he moved to Washington DC, and he waited tables full time for about nine months. Then he took a job on Capitol Hill answering constituent letters for around $20,000 a year, still waiting tables at weekends. He recommends waiting tables to everyone as an experience.

From there he did speechwriting, communications and op-eds for the 9/11 Commission during its investigation and reporting. Then London, doing emerging markets banking for Citigroup. Then a private equity firm investing across India, Africa, Southeast Asia and China, which took him to Beijing for two years working with Chinese portfolio companies expanding internationally.

He met his wife in London, moved to China with her, and returned to the US for business school. An internship there put him into product management, where the international portfolio was handed to him on the basis that he had lived abroad. That was Lookout, a mobile security company backed by Andreessen and Accel, where he went from intern to senior director of consumer product in three years. Then Flipboard, doing business development and corporate development with Samsung and others.

Everything except engineering

At Handshake, Stull has held nearly every function. He was the first product manager and built the product team. He was CFO until they hired one. He ran legal until they hired a chief legal officer. He has been CMO at various points.

He is clear about what the traditional route looks like by comparison. If there is a standard COO background, it usually runs through sales and go to market, pairing with a product or engineering-led CEO. His own is broad-based, and sales was the one thing he had least of before Handshake.

His description of running legal without being a lawyer is the most honest account of this I have heard. He finds external counsel, sets up the operating process, and then takes the hard questions they escalate, pushing on the legal issues until he can make a business call. He is not doing the lawyering. He is deciding how to balance the risk.

Underneath it is a specific claim to leverage: he knows Handshake better than almost anyone outside it, and most people inside it, with the exception of the two founders. Combine that with knowing enough about each discipline to ask sharp questions, and you can hire well and hold people accountable in areas you could not do yourself.

His summary of the role early on is the one every startup operator will recognize. A startup moves forward on momentum, because individuals push it. Much of his job was filling gaps, and knowing enough to find the right external person, assess them and set them up properly.

The commitments they made in public

This is the part of the episode I would send to a founder.

Handshake sells first to universities and now to a broader set of higher education partners including community colleges and boot camps, acting as the system of record for how students move from education to employment. Those institutions then invite their students and their employers onto the platform, which makes the sale a partnership and a distribution strategy at once.

The principles were set early and stated out loud. Students first. A network that is Handshake branded rather than white labeled, because a single branded network is what makes it possible for every employer to reach every student. Democratize access to information, to social capital, to credentials.

Then came the guardrails, which are the interesting bit. Handshake told universities it would never charge for access to a school and never charge for a job posting, because it wanted the most liquid marketplace possible. What it expected to charge for was premium tools covering branding, sourcing and analytics, which is roughly what it built.

Stull acknowledges the counterargument directly. Most people want optionality and do not want to be boxed in. His response is that when you are managing several constituencies at once, you sometimes need to create constraints on yourself that other people can believe in. Handshake could have monetized job postings early, and it would have killed the network and contradicted the mission.

The dividend arrived years later as trust. His example is people telling Handshake that they had been worried about something, checked with existing customers, and heard that those customers had the same concern three or four years earlier and watched Handshake honor its commitments.

The general principle he draws: live up to commitments, do the work every day, and it comes back in spades.

Non-negotiable on vision, flexible on tactics

The reviews feature illustrates how the two halves work together.

Handshake wanted students to be able to review internships, and now hosts more reviews of internships and first jobs than Glassdoor. In retrospect it looks obvious. At the time, career center leaders raised every way it could go wrong, including what happens when a board chair's company is reviewed badly, or when a review touches a Title IX matter.

Handshake's stance was to hold the vision and negotiate everything else. We will work these problems out with you, we will listen, but in service of a mission we already agree on.

The mechanism was in-person and expensive. All-day sessions with 30 career centers in Philadelphia, then Miami, then San Francisco.

Solving a three-sided marketplace

Two-sided marketplaces are hard. Handshake took on three, and Stull explains why that was actually the shortcut.

The existing players had split into two camps. Some sold white-labeled software to universities without building relationships with students or employers. Others built relationships with students and employers and refused to touch higher education, because the sales cycles were long and the vision looked unbuildable.

It took founders naive enough to try both. Garrett, Ben and Scott kept observing that students and employers still interacted through the university, and that the university systems were bad. They got useful mentorship, including from Andy Chan, then head of the career center at Stanford and now at Wake Forest.

Stull calls it a textbook innovator's dilemma. LinkedIn was never going to attempt this. The founders did because they had nothing to lose, and then executed with more sophistication than their naivety suggested, including driving overnight and sleeping in the back of a McDonald's to make one customer pitch.

Once universities were on board, the chicken and egg problem dissolved, because the real marketplace is between students and employers and the university seeded both. When Eastern Michigan came on, every employer recruiting there had to be on Handshake, and every student was invited.

Then the network effect turned. An employer with 20% of its target schools on Handshake had one login for those and 80 separate systems for the rest. The following year it was half. By the third or fourth year, employers were telling schools they would not recruit there unless the school adopted Handshake, because posting a job 25 times across 25 systems was untenable. Schools reported enormous increases in job postings.

The growth stages Stull cites: five schools to 60, then 150, then 450, which is when it took off, then 700.

His closing note on the model is a caution against seeing universities purely as distribution. They are embedded partners who have been working on this problem for decades and function as genuine thought partners.

Why the timing worked

Stull's account of the competitive landscape is a good lesson in reading a market.

Career centers were digitized in the late 1990s and early 2000s. The eventual standard was Symplicity, originally a government contractor, whose white-labeled system became close to a default.

The trap in that design is worth understanding. Schools liked white labeling because it preserved their brand and allowed customization, which is reasonable when you care about your community's identity. But it meant every employer needed a separate login for every school, producing a fragmented system in which students lost the opportunities. Schools gained customization and their students paid for it.

Meanwhile the incumbent was running for profit at a smaller scale, charging more and investing less. Stull's phrase is a fat and happy incumbent who was not innovating.

Into that gap Handshake brought a clear students-first vision, plus a tactic that doubled as marketing: symposiums around the country where top career services directors spent all day designing what career services of the future should look like. That produced the roadmap and turned participants into evangelists.

The founders worked essentially unpaid for two years before their first customers, then had five customers to work with closely. Stull's point is that the growth curve everyone sees afterward sits on top of years of research nobody saw.

Generalists to specialists, forever

Asked when the company moved from generalists to specialists, Stull rejects the premise of a moment.

At 500 people it is still happening, and roles that today are two jobs in marketing will be three in a few years. There is no clean transition, and he offers grace to anyone in the middle of it, because it is hard for the individuals and hard for the company, which still wants to be scrappy.

The way through is diagnostic rather than scheduled. Look at what is breaking, and where. He borrows the framing that startups are a series of fires, and the question is whether each is an ember or a forest fire. You cannot specialize every role at once, because you cannot fund it and because most things are working.

He is honest about getting it wrong in both directions, noting they should have hired a head of legal a year earlier than they did.

His actual mechanism is conversation, repeated. Every few one-on-ones, discuss what this world looks like in six or twelve months, so expectations are set long before someone's scope changes. Talk about the evolution at all hands. Bring in outsiders who are a stage or two ahead so the message does not only come from leadership. Circulate Molly Graham's essay on giving away your Legos.

The goal is that nobody is surprised when their role narrows, and that leadership is planning six to twelve months ahead rather than a month.

What the money is for

Handshake closed its round in the autumn and announced it in January, and Stull is explicit that they did not need the money. They took it to accelerate against a long-term plan, and because favorable terms at low dilution are not always available.

The vision behind it is bigger than campus recruiting: being the first place anyone starts, restarts or jumpstarts a career.

His analysis of the competition is generous and precise. Indeed is an excellent job board. LinkedIn is genuinely useful and strongest for mid-career professionals, but it largely reflects your past, the people you already know and the skills you already have. Handshake's claim is about building relationships and skills you do not yet have, whether you are 18 with none of them or 40 and changing industry.

That is why partnerships with companies like Coursera matter, and why the product roadmap points toward career communities. The questions people actually have, how to run a first one-on-one with a manager, how to negotiate compensation, how to move into a different industry, are conversations rather than searches.

At the time of recording the company was around 500 people heading toward 800, approaching $100 million in ARR, and had been doubling revenue year over year.

How to choose your first international market

Stull's advice here is unusually practical, and he credits a blog post by John O'Farrell at a16z for the shape of it.

Build a spreadsheet of candidate countries with a few core characteristics: language, market size for your product, competitive landscape. The countries on it will mostly be the obvious ones. Think at country level rather than region, because the legal implications attach to countries.

For Handshake the UK was the clear first choice, for two reasons. Language, and model. The UK still has a vibrant university-led early talent ecosystem, whereas France, for example, does not have career centers at its universities, which makes the model far less transferable.

The best signal, though, was demand. Handshake was getting pulled into the UK by employers with relationships on both sides of the Atlantic. His rule: when customers are telling you they want your product somewhere, that is the market.

He notes an alternative pattern, where companies pick a small English-speaking market like New Zealand for quiet product launches, and a general preference for starting in English-speaking countries and using them as beachheads.

On China he is direct, drawing on two years working there. He would not go, and would wait considerably longer if you must. It is an extremely hard market for Western technology companies and has become harder. Investing in Chinese companies may be the better play than trying to build your own.

The operating system, and honest words about OKRs

The biggest change to how Handshake runs itself was the pandemic. They had been almost entirely in a San Francisco office, with a new New York sales office not yet filled and a small Denver engineering team.

The other driver was simply decision-making capacity. There was a period when the CEO, or the CEO and Stull, made the decisions. Building a bigger team with strong leaders means empowering them, which requires structure that goes well beyond meetings into documents, cadence and how work moves.

Handshake uses OKRs, and Stull's endorsement is refreshingly qualified: he hears reasonable complaints about them, and considers them the best of a bad set of options. What matters, whatever you use, is clarity on the vision, the goals and the metrics you track. Michael points out that John Doerr is an investor.

On structure, Handshake is largely functional with business units aligned to customer groups, since selling to universities differs substantially from selling to employers. Engineering, product and design stay consistent across both while marketing and sales were split, and a new CMO was being hired to bring them together.

The UK story is instructive. It began as its own team under a single leader, and they found it becoming siloed rather than embedded, so they moved toward aligning people to their global functions while keeping a UK overlay for coherence.

His conclusion is that any technology company reaching a certain size ends up in some form of matrix, and the work is being clear about which line is solid and which is dotted.

Building a board beyond investors

Handshake added Dr. Michael Lomax, president and CEO of the UNCF, to its board.

Stull's argument for independent directors is one earlier-stage companies should hear. Early boards are investors. As you grow you need independents, eventually for legal reasons ahead of an IPO, but sooner because it changes the dynamic. Investors have their own interests even while honoring fiduciary duty, and collectively tend to share a narrow set of experiences.

Their first independent director was Margo Georgiadis, formerly president of Google Americas and CEO of Mattel and Ancestry. Stull says the difference showed in her first meeting, where the questions were higher level and harder: are these the right goals, are they right for now, how is this evolving.

The search for Dr. Lomax took about a year, deliberately looking for someone deeply embedded in higher education rather than a career academic, and committed to the mission. UNCF has funded more scholarships for Black and Latinx students than any other organization and is the primary scholarship function for HBCUs. His description of what Lomax brings includes something you rarely see in a board write-up: he is warm and funny, gives you a kick when you need one, and makes you feel he has your back while being someone you can hold accountable.

Three levels of a stated commitment

Handshake's approach to access shows up in the product, the team and the capital structure.

In the product, employers use the network to build early-career pipelines, because in many fields the diversity they are seeking does not exist to hire at mid-career. A recruiter can search for, say, a computer science major with Python experience who wants to work in a specific city, and build a relationship rather than wait for an application.

On the team, Stull is candid. The company was founded by three white men from Michigan, and he was an early hire and a white man from California. He describes fits and starts, real investment, accountability through OKRs, and genuine pride in where the team and leadership are now, while noting the irony of a company selling this capability having to do its own work on it.

The third level is the one I had not seen before. Handshake took investment from Base10's Advancement Initiative, a Black-led and Black-owned growth fund with significant capital from under-capitalized HBCU endowments, which waives fees and half its carry on that capital. As Handshake's valuation grows, those returns fund scholarships and HBCU endowments directly.

The green machine

Michael's standing question produced a good one. Around 2016, with a team of about 20, Handshake attended the NACE conference in Chicago and saved money by staying in DePaul's dorms. Stull shared a room with the CEO and slept on what students call a green machine, the waterproof mattress designed to survive an undergraduate's worst night.

The contrast is what makes it. Shortly before, at Flipboard, he had been flying to Korea to work with Samsung, in business class, staying in hotels. Now he was FaceTiming his wife and children from a dorm room on a plastic mattress, wondering whether he had made the right call.

The 5 things I took away from this conversation

1. Public constraints can be worth more than optionality. Promising never to charge for job postings closed off an obvious revenue line and bought a decade of trust in a market with several suspicious constituencies. Most companies keep every option open and wonder why nobody believes them.

2. Be immovable on the vision and negotiable on everything else. The reviews rollout is the model. Career centers raised every objection, Handshake took them seriously, and never reopened the question of whether reviews should exist. That combination is what makes customers into partners rather than obstacles.

3. A third party can solve a two-sided cold start. Universities were not just a sales channel, they were the seeding mechanism that brought students and employers on simultaneously. If you are staring at a chicken and egg problem, the question worth asking is who already convenes both sides.

4. There is no moment when you switch from generalists to specialists. It happens function by function, triggered by something breaking, and it never finishes. The practical answer is talking about role evolution constantly in one-on-ones so nobody is surprised when their scope changes.

5. Get one independent director earlier than you think. Stull's account of the first non-investor board meeting, where the questions moved from detail to whether these are the right goals at all, is the clearest argument I have heard for doing this before an IPO makes you.

FAQ

What does a good strategy implementation example look like in practice? Handshake set a small number of non-negotiable principles, students first and a single branded network rather than white-labeled software, then made public commitments about what it would never do, including charging for job postings. It negotiated tactics openly with customers while holding those principles fixed, which built the trust the marketplace depended on.

How do you solve a three-sided marketplace? Find the party that already convenes the others. Handshake sold to universities, which then brought both students and employers onto the platform, so the student and employer marketplace never had to be seeded from scratch. Network effects then took over as employers consolidated onto one login instead of dozens.

When should a startup replace generalists with specialists? When something specific breaks, not on a schedule. Stull's approach is to look at where problems are appearing and how severe they are, then hire depth there. He treats it as a continuous process that is still underway at 500 people, and stresses talking about role evolution regularly so changes are never a surprise.

How should a company choose its first international market? Score candidate countries on language, market size and competition, and weight heavily toward wherever customers are already asking for you. Handshake chose the UK because of language and because it shares the university-led early talent model, whereas markets without university career centers made the model far harder to transfer.

Why add independent board members before you have to? Because investor directors, however diligent, share relatively narrow experience and hold their own interests. Handshake's first independent director immediately raised the altitude of board discussion from operational detail to whether the goals themselves were right, and its second brought deep customer-side knowledge of higher education.

Also mentioned

• Handshake and its higher education, employer and student network
• Lookout and Flipboard, Stull's earlier product and business development roles
• Glassdoor, LinkedIn and Indeed, the platforms Stull positions Handshake against
• Coursera, a partner in the reskilling side of the vision
• Molly Graham on giving away your Legos, the essay Handshake circulates about scaling roles
• UNCF and Dr. Michael Lomax, and Base10's Advancement Initiative, the fund routing returns to HBCU endowments
• NACE, the annual conference where Handshake once slept in university dorms
• Jonathan Stull on Twitter at @JonStull

Listen to the full episode

Jonathan Stull on Between Two COO's

Between Two COO's is hosted by Michael Koenig. Subscribe on Apple Podcasts, Spotify, or wherever you listen.