Case
How Mark Roberge Built HubSpot’s Sales Engine
Mark Roberge, the head of sales at HubSpot, wasted hours posting ads across job boards while hiring for his team. He got hundreds of applications, completed 50 phone screens and took dozens of in-person interviews. Despite this, he hired zero candidates.
Back in September 2007, this was critical for Roberge. He had just joined HubSpot as its first sales hire and had no experience running a sales team. As he put it himself: “I’ve never done sales before. I’m a mechanical engineer by trade. I wrote code for the first years of my career.” Yet he would go on to serve as SVP of Worldwide Sales and Services, scaling the company's revenue from $0 to over $100 million and expanding the team from one employee to four hundred and fifty. An MIT-trained engineer by background, Roberge brought a metrics-driven, process-oriented approach to sales that would eventually earn him recognition as one of Forbes’ Top 30 Social Sellers in the World. HubSpot had grown revenues more than 6,000% since 2007 and by 2011 was recognised as the second fastest growing software company and 33rd fastest growing company overall by Inc. 500. The company went public on the New York Stock Exchange in October 2014, raising $125 million in its initial public offering. But that was all later. In those early days, none of that was guaranteed. First, he had to figure out how to actually hire salespeople who could sell.
Roberge eventually realised that truly great salespeople never have to apply for a job. They would always have multiple offers at all times, even if they were not actively applying. In other words, unlike most job seekers (who had to actively look for their next role), truly great salespeople were ‘passive candidates’. They never needed to go into job-hunting mode.
In his book, The Sales Acceleration Formula, Roberge observes:
Their old bosses are calling them, probably quarterly. “Can I take you to a ball game?” “How is the new gig?” “Are you still happy?” “Are you making money?” “Did they change the compensation plan on you?” “You'll never believe how good things are going over here.” “You have an open invitation to be on my team.”
Clearly, Roberge’s traditional recruiting methods were not attracting great talent. It was time to change strategy. He would have to adopt a ‘passive recruitment process.’ That is — one suited for passive candidates.
His first thought was to simply hire a recruiting firm. In the first year, he worked with around 10 firms. The results were average. These agencies charged between 15% and 20% of the base salary of the candidates. Each firm wanted him to work with them exclusively so that the candidates were not contacted by multiple firms. He ignored this request and always had two or three firms going at once. If a firm presented a handful of candidates who did not make it through the early stages of the process, he stopped working with that firm and moved on. He realised, in an industry with multiple, similar competitors, it was important to move on quickly if you found yourself unimpressed with the product or service delivery.
For a while, he was able to tolerate average agency results. But relying on outside resources for the most important drivers of his success proved difficult. What if HubSpot wanted him to triple the pace of sales hiring? Scaling the external agency model simply wouldn’t be good enough or predictable enough for their needs.
Then, he reached a turning point:
I received the best advice I've ever gotten on candidate sourcing.
Don’t hire a recruiting agency. Don’t build a corporate recruiting team. Build a recruiting agency within your corporation.
The problem with recruiters was that they were not working exclusively for you. Even though they were incentivised with performance-based packages, there was no guarantee that they would pitch your company first. They would naturally prioritise the company that would pay them the highest commission.
On the flip side, internal corporate recruiters were aligned with the company but under-motivated. They made less money because they had a fixed salary, with no commission. As a result, they just weren’t interested in cold-sourcing candidates, nor were they particularly hungry to win high-quality hires. However, they did a far better job at launching job ads, at directing inbound resumes to hiring managers and ushering candidates through the hiring process. These recruiters typically hired agencies to outsource the cold outreach ‘dirty work’.
Roberge adopted a hybrid of these two models. He built a recruiting agency within HubSpot. He found a talented agency recruiter who was thinking about starting her own firm and said to her, “Why not start the firm within HubSpot?” He paid her and her team like they were external, agency recruiters. Instead of a flat base salary, they got a lower salary with meaningful performance bonuses that amounted to a higher overall earning potential. They were prohibited from using external agencies. Their performance bonuses were based on the fill rates, timing, and long-term success of the hires they made.
He took it one step further and made sure the team was measured like a sales team. How many outbound candidates did they ‘touch’ this week? How many touches led to connections? How many connections led to a phone screen? How many phone screens led to interviews with HubSpot's hiring managers? How many interviews led to a hire?
Now, Roberge was looking at a predictable, scalable process. He had access to the metrics and could iterate on them quickly.
This structure was still not enough to find great salespeople. He had created an in-house team of hunters, but how would they go about finding exceptional sales talent? There were two ways to go about finding passive candidates: search on LinkedIn or get referrals from your team. Roberge went with both.
To identify candidates on LinkedIn, the first step was to use the platform’s advanced search feature to narrow down good fit candidates. Roberge was hiring in Boston, so he narrowed it to his zip code. In the job title field, he looked for labels like “sales” and “account executive”. He would screen for intelligence based on schools. Candidates from quality undergraduate schools were likely to be more intelligent. Lastly, Roberge was aware of local companies with large sales teams that trained their teams well. He used this to identify more viable candidates.
The most effective way to find people turned out to be from other heads of sales in Boston with whom Roberge had a relationship. Since he had been networking for years, many of them “owed him a favour”. Once he identified such a sales leader from his network, he would look for salespeople who used to work in that leader’s company. Within 15 minutes of searching, Roberge could typically come up with a solid list of people he could consider. He would then ask the sales leader if they recommended any of the candidates as standout performers. They would always respond with recommendations and would add a few more names based on what HubSpot was looking for.
With a quick screen of their LinkedIn profiles, he was able to eliminate low-quality candidates. He looked for several key signals.
First were the obvious indicators of sales excellence: rankings on their team, consistent quota attainment, President's Club attendance, and similar accolades.
Longevity mattered, but only in the right context. Anyone could last a year at a company, even if they were underperforming. Mediocre salespeople might make it to two years. But when he found someone who had survived three, four, five years or more at a company known for its high-performing sales team, that was a strong signal. These environments did not tolerate dead weight for long.
He also looked at buyer context alignment. Were they selling to large enterprises or SMBs? Was the product a commodity or something complex? Was the sales process relationship-driven or transactional? A mismatch here wasn't a dealbreaker, but alignment meant a shorter learning curve. Low-hanging fruit, in other words.
School and college major served as proxies for intelligence and prior success, both of which he’d identified as predictors of sales performance. The quality of the institution, the difficulty of the major, and academic performance all correlated with these traits. Interestingly, Roberge found that second-tier schools often produced the best candidates. He’d hired plenty of successful people from MIT and Harvard (some eventually became top leaders at HubSpot), but many grew bored in sales roles and wanted career progression faster than the organisation could provide.
Finally, LinkedIn profile quality itself mattered, though less than the other factors. A weak, photo-less profile was a massive red flag. Conversely, a polished profile with a professional photo, 500+ connections, and glowing recommendations from high-level executives made a strong positive impression. He wouldn’t get carried away with this signal, but candidates at either extreme of the profile quality spectrum definitely influenced his assessment.
Now, he had a workable shortlist of potentially great salespeople. The next challenge was getting through to them.
The best scenario was when he shared a mutual connection with the candidate. A good friend or fellow employee could make an introduction, and that almost always got him access. It was an enormous advantage when it was available.
But second-degree connections weren’t always there. When he had no mutual contacts, Roberge would take a different approach. He’d guess at the person’s corporate email format and send them a direct message. He never bothered with LinkedIn's InMail feature. The email inbox, he figured, was more valuable real estate to occupy.
His messages followed a specific structure. Take this example he sent to a Yahoo! salesperson who’d recently graduated from Boston College:
Subject: Yahoo!/Boston College
Body: John,
Congrats on all your success! I run the sales team over here at HubSpot. Our current team can't keep up with the inbound lead flow so we are expanding the team. Your background is similar to those of our current top performers. Are there any folks in your network who are in the job market and have a background similar to yours?
Best,
Mark Roberge
SVP of Global Sales
Mobile: 123-456-7890
The approach had a few things working in its favour.
The subject line was designed to be opened. Something simple usually worked — like the person’s current company and alma mater. Most people would at least open that out of curiosity.
The message itself stayed short. Roberge wasn’t trying to sell the opportunity at this stage. These were people who probably weren’t looking for jobs. He just wanted to create a tiny bit of doubt in their minds. What if they were missing something life-changing?
The request itself was indirect. Rather than asking John directly about his interest, he asked if John knew anyone who might be looking. This framing made the whole interaction feel less aggressive. The prospect didn’t have to feel uncomfortable about engaging. And the question was sincere. He was interested in John, sure, but he was just as interested in anyone from John’s network with similar credentials who happened to be actively job hunting.
When emails went unanswered, he’d call the next day. One advantage of recruiting salespeople versus, say, engineers: they all had desk phones, and they generally answered them. Getting someone on the line wasn’t particularly difficult.
The second tactic to reach folks that Roberge used was to get a “forced referral”. The idea was to leverage his existing team’s networks to find strong candidates, but in a much more structured way than the typical referral request. This turned out to be his best technique for finding talent, though it only really worked once the team reached a certain size. With just one or two salespeople, there wasn’t much to work with. But once the team started scaling, the method became remarkably effective.
He’d connect with all his salespeople on LinkedIn, including recent hires. Then he’d wait. About a month or two into their tenure, once they’d settled in and felt comfortable, he’d ask them for referrals.
That part wasn’t novel. Plenty of companies ask for referrals and offer bounties for successful hires. What made this different was how he structured the ask.
Rather than simply saying “We offer $2,500 for new hire referrals – do you know anyone?”, he’d schedule a 20-minute meeting for the following day. Then he’d spend that evening going through their LinkedIn connections himself. If someone had 275 connections, he’d comb through all of them, looking for salespeople in Boston who seemed like potential fits.
The next day, he’d show up to the meeting with a list. He shortlisted, say, 18 names that matched his criteria. Then he’d let the employee do the filtering. Which of these people were actually top performers? Which ones would they feel comfortable introducing him to?
It required more upfront work than a generic referral request, but the results made it worth the effort.
The final step was to map out the entire sales landscape in Boston. He wanted to know everything. How large were the sales teams at other companies? What were their people making? Were they running inside or outside sales? What did their training look like? Had they recently changed their compensation plans?
As he worked through LinkedIn profiles, he started building a comprehensive list of every company in Boston with an inside sales team. Before long, he’d interviewed at least one person from each organisation. Sometimes he’d take interviews with candidates who looked mediocre on paper, just to learn more about their company's sales operation.
The interviews became intelligence-gathering sessions. Roberge had a standard set of questions he’d work through. What did the company pay its salespeople? How were the comp plans structured? What was the buyer context like? Were deals transactional or complex? Enterprise or SMB? Outbound or inbound leads? How many reps worked there? What were the different sales roles? How was the team organised?
He’d ask about sales training. Did they use a formal methodology? Did they bring in outside trainers or have full-time staff dedicated to development?
Then he’d probe for instability. Had there been any major changes that might cause top performers to leave? A new commission structure? Leadership turnover?
Finally, he’d ask about the top salespeople at the company. If the person he was interviewing wasn't the top performer, he’d ask who was. Sometimes, salespeople would blame their territory. He’d ask which territory the best salesperson worked in, then find a way to network with them. He never actually used this last tactic himself, but he knew peers at other companies who’d done it successfully.
During the interviews, Roberge would look for traits that best predicted success at Hubspot. Knowledge of these traits were hard won. In the early days, he scored new hires on 12 different criteria. As time passed, he’d go back to the 12 scores to see which quantitative measures were most predictive of success. Eventually, Roberge had hired enough people that he could run a regression analysis — he narrowed his list of criteria down to five characteristics. These five measures were then turned into a repeatable process that were used throughout Hubspot’s sales org.
The first was coachability. He’d test this through role-play exercises. After the candidate finished, he’d ask them to self-assess, then coach them on one improvement and ask them to redo it immediately. Most stumbled on the second pass, but he wasn’t looking for perfection. He was looking for effort. In over a thousand interviews, maybe five people crushed the second attempt. Those five became rock stars.
Curiosity was also important. Great salespeople asked questions that felt genuinely interested rather than interrogative. He’d test for this before the interview even began. He’d observe if they asked questions when they met in the lobby. Then he would check again if they asked the right follow-up questions during the role-play exercise.
Prior success was easiest to measure. He wanted the top 10% performers. For candidates without sales backgrounds, he’d look at academic performance, sports, leadership roles, or exceptional achievement anywhere. His sales team ended up including an Olympic gold medalist, a symphony cellist, and a comedian featured on Comedy Central.
Intelligence mattered because the industry was evolving rapidly. He’d test this by sending training materials after the first screen, then referencing those concepts in the next role-play. How well had they absorbed it? How clearly could they communicate it back?
Work ethic was the hardest to evaluate. He’d observe behaviours throughout: how quickly they returned calls, turned around deliverables, pushed the pace. During reference checks, he’d ask former supervisors to rank the candidate’s traits from strongest to weakest.
The whole system Roberge built came down to a simple insight: hiring great salespeople wasn’t fundamentally different from selling itself. He stopped waiting for candidates to come to him and went out to find them. He built an in-house team structured like a sales team, measured like a sales team, and compensated like a sales team. He mapped the competitive landscape, used LinkedIn searches like lead lists, and crafted cold emails with the same care his team put into prospect outreach. None of this required special genius. It required a change in mindset. Great salespeople don’t apply for jobs. So if you want to hire them, you have to sell to them.
Sources:
The primary source for this case was The Sales Acceleration Formula.