The First Sales Hire Mistakes That Sink Startups w/ Rafa Jara Simkin, Partner @ Theory Ventures

Rafa Jara-Simkin is a Partner at Theory Ventures, the research-driven VC founded by Tomasz Tunguz.

Before investing, Rafa spent nearly a decade in the GTM trenches. First, he joined Looker as employee ~25 in 2013, and rode it from ~$1M to $100M+ ARR through the Google acquisition, building and leading sales teams in New York along the way. Post-acquisition, he did a stint inside Google Cloud running strategic enterprise accounts, then joined Monte Carlo to learn a top-down, category-creation motion under Jordan Van Horn.

He later built an executive coaching practice before Tomasz pulled him into Theory, where he now advises portfolio founders on building their first sales teams. He's a Santa Cruz surfer, a recovering economist, and deeply interested in the inner game of leadership.

2013, Santa Cruz. Rafa's reading the local paper and sees "local tech company raises millions from Redpoint Ventures" - which made no sense, because there were no tech companies in town. So he got himself into a Looker user meetup in SF, where the team assumed he was a buyer. He played the part, let them schmooze him, and met the VP of Sales.

Then the honesty play. He called the rep: "bad news - we're never going to buy from you. I think it's amazing though, and I would love to work with you guys." The head of sales wouldn't talk to him for a while because he was technically a prospect in pipeline turning into a candidate - until he accepted the deal was never real anyway. Rafa joined at ~25 employees and ~$1M ARR, moved to NYC for a "one-year stint," and ended up running 4 managers and 30-40 reps as the company scaled to 1,000 people and $100M+ through the Google acquisition.

His team-building take: "I'm going to credit at least half, if not more than half, of our success in sales to the sales engineering team at Looker." SEs partnered on every single deal, got in the weeds with champions, and made the champions love them. Not layups for the reps - but in his words, absolutely no way it happens without that team.

The bigger lesson for anyone selling to developers or data teams: there's zero tolerance for BS or a sales pitch. Devs don't want to go through you - "I'd rather just download the thing." The skill is disarming a pessimistic, hole-poking audience with genuine curiosity: "you need to geek out like a homie with them." Ego out, quota out of your head, actually care about what they're building.

The mistake that keeps happening (and one we shout from the rooftops): hiring reps to fix a pipeline problem. Rafa empathizes - you're building in a multi-unicorn space, competitors are staffing up, and a two-person team feels like falling behind. But there's a line.

His balance: "it should always be a little uncomfortable - you're hiring people before you quite have the pipeline for them" while cranking on marketing and top of funnel at the same time. What crosses the line: hiring four reps in one quarter with no repeatability, no clear PMF, and no evidence of the multiplier. His words: "a disaster waiting to happen." The related blind spot: founders treating marketing and brand as a foreign concept, when their buyer is sitting in a sea of thousands of AI tools where "default is no."

Founder-led sales until repeatability is great advice for most founders - and bad advice for some. Deeply technical founders with zero GTM aptitude (or interest) shouldn't spin tires experimenting - they need "almost like a co-founder buddy" on the go-to-market side early.

On who that hire should be: avoid the operator who's spent years running managers from a dashboard and living in internal politics. The Goldilocks profile is an incredible frontline manager who grew up under great VPs, hasn't had their shot yet, and is dying to prove it - still hungry, still in the weeds with reps. The rare exception: the overqualified exec who's a true deal junkie and physically can't stay out of the trenches. And his broader conviction applies to all of it: a great sales team doesn't transfer - drop peak Snowflake's team into Looker's glory years and "it would have failed miserably," and vice versa.

The newest theme in his portfolio conversations: open source models caught up. GLM and Kimi are performing near frontier level at a fraction of the cost and can be hosted in private data centers - and it's reorienting everybody's forecast of the entire market.

The sales-side effect: enterprise buyers are now diligencing AI-native vendors' margins. Everyone's under pressure to run negative margins betting that token costs fall, so buyers are asking the uncomfortable question: "if I buy your software and you fail because you token'd yourself out of your capital, now I have to find another vendor." If you're selling for an AI-native, expect your business model's sustainability to start showing up in deal cycles.

(09:23) Origin story: How Rafa fell into sales and snuck into Looker through the back door
(13:29) The Looker rocket ship: 25 to 1,000 employees, $1M to $100M+ ARR
(17:17) Where the Looker mafia went and what made those hires work
(19:09) Selling to developers vs. app-layer buyers: two different personas
(31:57) The mistakes founders keep making with their first sales hires
(35:43) Three types of founding sales roles (including the disaster zone)
(39:34) What's new in the portfolio: open source, inference costs, and margins
(41:45) Why enterprise buyers now diligence your token margins
(44:33) Theory's thesis and the app-layer "Goldilocks zone" vs. Anthropic
(48:52) Why generic sales advice is B.S. without context
(55:55) Executive coaching: what founders actually struggle with
(59:37) Rapid fire: AI tools, elite GTM teams, and creative outreach