# Mark Roberge

Mark Roberge writes, speaks, and invests in companies building and scaling go-to-market and sales teams, with a focus on how AI is changing that work. His posts discuss his experience building HubSpot’s CRM and his work with founders through Stage 2 Capital.

- Profile: https://b2binfluencer.marketing/creators/mark-roberge/
- LinkedIn: https://www.linkedin.com/in/markroberge
- Median of last 8 posts: 190 reactions, 32 comments (checked 2026-11-02)
- Topics: artificial-intelligence, sales, startup, ai-in-sales, go-to-market-strategy, customer-retention, founder-led-sales
- Based on 12 dated posts, latest 2026-09-30

## Summary

Mark Roberge is most useful for B2B SaaS go-to-market and sales team advice, especially how AI changes roles and retention.

- Best for: A B2B SaaS company rethinking its go-to-market and sales organization for the AI era.
- Not for: A consumer brand seeking sponsored posts, product demonstrations, discount codes, or reach and engagement targets.
- Check first: Confirm recent LinkedIn audience data and any documented survey results for the roles, industries, or seniority you need.

### Takeaways

- Retention is the best measure of product-market fit, so find an early leading indicator.
- AI go-to-market prompts fail when the ICP ignores post-sale outcomes.
- Sales roles are moving back to full-cycle sellers, with the future CRO closer to RevOps.

## Main ideas

- **Retention is the best measure of product-market fit, so find an early leading indicator**: Roberge argues revenue can hide churn and surveys can produce false positives, so retention is the better test. Because renewal data takes a year, Roberge uses a leading indicator of retention: a behavior visible in a customer's first week or month that predicts whether they stay. The formula Roberge gives is "P% of customers do E event every T time." Roberge treats it as a "forever metric," not a one-time speed-to-value milestone, and says teams should know whether 30% or 90% of customers hit it before scaling.
- **AI go-to-market prompts fail when the ICP ignores post-sale outcomes**: Roberge identifies three mistakes founders make when using AI to build GTM foundations: defining the ICP only by close rate, sales cycle, and ACV; treating the leading indicator of retention as a one-time speed-to-value number; and ignoring early customer interviews when writing the first sales playbook. Roberge argues the first 20 interviews, including with people who turned out not to be a fit, hold the richest inputs for qualification and discovery, connecting The Mom Test to Sales 101.
- **Sales roles are moving back to full-cycle sellers, with the future CRO closer to RevOps**: Roberge argues AI that can do 2-10x a human's work ends the SDR/AE/CSM specialization era and brings back the full-cycle athlete. Roberge expects the future chief revenue officer to look more like today's revenue operations leader and questions functionally siloed tech stacks. The conversation also covers whether everyone still has a job in five years, building a virtual workforce, whether CRMs are enough, and the problem with AI SDRs.
- **This is the best time to be a fast follower, and vertical AI will sell the whole stack**: From an NYSE interview, Roberge argues AI development tools have made copying products faster, while many first movers lack moats and cannot cut prices after raising at high valuations, opening the door for efficient fast followers. Roberge expects vertical AI software to sell an entire tech stack rather than a point solution, raising ACV and lowering cost. Roberge also argues change management beats product right now because user and organizational resistance is harder than proving the AI is best, and says MBAs are undervalued as deciding what to build matters more than building it.
- **Humans stay in sales because they own outcomes, not only because of what they can do**: Roberge summarizes a HubSpot Unbound conversation with Michael Hartman, a Stage 2 Capital LP and head of enterprise Americas at Anthropic. Hartman's argument is that humans remain in the loop because they must be accountable for decisions and outcomes, and that AI will eliminate human-only jobs rather than human jobs. Other points from Hartman include not waiting for an employer or school to teach AI, Anthropic's hybrid of traditional SaaS and AI-native tools, investing in counterbalancing operations to manage spend and spot redundancies, and weighting prior domain experience.
- **Stage 2's operator network is the support model: named LPs help with compensation, hiring, call reviews, and positioning**: Roberge describes Stage 2 Capital's support model through Haize Labs' acquisition by Beacon Software. The LP base includes more than 1,000 go-to-market operators from companies such as Anthropic, OpenAI, Databricks, Cohere, and Salesforce. Roberge credits specific operators with specific work: Graham Collins of QuotaPath on the GTM system, compensation model, AE hiring scorecard, and candidate assessment; Matt Braley, former CRO of EliseAI, reviewing sales calls weekly to sharpen discovery and qualification; and Meini Heuberger, former head of sales at Cohere, on positioning, messaging, and buyer journey.
- **Scaling lessons from a portfolio company: invert R&D, price for commitment, hire two reps**: In a Science of Scaling conversation with Dan Lee, co-founder and CEO of Nooks, Roberge lists five practices behind Nooks's growth from Stanford to a nine-figure business. Nooks spends 70% of effort studying customer usage, prospect sentiment, and market signals to decide what to build and 30% building, inverting the traditional ratio. Free is rarely right on day one, but profitability is premature, so Nooks priced for commitment. Roberge compares copying early LLM and DevOps distribution to the 2012 PLG wave, while Nooks chose sales-led growth for its context. Nooks's initial north star was delivered customer outcomes, not contracts or revenue, and it hired two salespeople rather than one so a struggling rep could be diagnosed as a bad hire or a systemic problem.
- **Founder-led sales: the first playbook is already there, and admin work can be automated**: In an SF Tech Week fireside chat with Andrew Smith, then VP of global sales at Cursor, Roberge covers defining an ICP for accelerated scale, finding first design partners in half the time, running executive first meetings that close at twice the rate, doubling selling time by automating pipeline management and forecasting, and the claim that founders have already built a first sales playbook without realizing it. Roberge also raises the unsolved question of what happens when the buyer is an agent.

## Work to read

- [Rebuilding sales roles for the AI era, with the co-builder of HubSpot's CRM](https://www.linkedin.com/posts/markroberge_crm-thescienceofscaling-activity-7501645532600164353-_fT_)
- [Three mistakes founders make when prompting AI to build go-to-market foundations](https://www.linkedin.com/posts/markroberge_worlds-collided-on-thursday-afternoon-i-activity-7508220517372346368-t7FC)
- [Four market theses from an NYSE interview](https://www.linkedin.com/posts/markroberge_nyse-activity-7500249399101411328-3Ae-)
- [The leading indicator of retention, explained with a Slack example](https://www.linkedin.com/posts/markroberge_lir-startups-activity-7519068897128648704-HSgI)
- [Lessons from incubating Nooks out of Stanford](https://www.linkedin.com/posts/markroberge_scienceofscaling-activity-7512971710611136513-yDtv)
- [What a frontier AI lab is actually doing with go-to-market](https://www.linkedin.com/posts/markroberge_this-month-at-hubspot-unbound-i-sat-down-activity-7511175885702688770-JhT9)
- [What Stage 2 Capital actually does for a portfolio company](https://www.linkedin.com/posts/markroberge_congratulations-leonard-tang-and-the-entire-activity-7508865477146972161-6ghu)

## Brand relationships

- **Stage 2 Capital** (investment)
- **HubSpot** (mention)
- **QuotaPath** (disclosed_partnership)

## Partnerships

Mark Roberge's partnership activity in the reviewed posts runs through investing work at Stage 2 Capital and media produced with HubSpot For Startups. Companies appear alongside Roberge as podcast guests, as co-hosts of on-stage conversations, as portfolio companies supported by go-to-market operators, and as tool partners. None of the posts publishes sponsorship inventory, pricing, editorial terms, or a booking route, so the formats below describe the shape of existing collaborations rather than a published offer; terms would have to be confirmed directly.

## Alternatives

- Yannick Kpodar: Choose Kpodar if the brief is specifically about making an early sales hire—not redesigning the whole sales organization around AI. In a post about a Full Stack CEO episode, Kpodar lays out a useful decision: hire to discover the sales motion, repeat a proven one, or build a sales team. The examples distinguish a self-sufficient seller who can replicate a proven playbook from a more senior operator who can own a territory like a “mini CEO.” That gives a hiring-focused session a direct decision framework.
- Austin LaRoche: Choose LaRoche if the company’s main obstacle is getting an existing HubSpot setup adopted and used well. LaRoche describes two projects where over-engineered CRM instances left teams struggling, and argues that implementation should fit the client’s capacity and build new habits at a suitable pace. A session or advisory engagement built around simplifying CRM configuration and improving adoption would be more directly grounded in this material than Roberge’s strategy-led AI-era org redesign.
- Daniel Gamboa: Choose Gamboa if the brief is centered on customer retention, expansion, or improving adoption of HubSpot Service Hub. One post contrasts acquisition spending with net revenue retention and describes health scoring, automated check-ins, and expansion triggers; another presents a client case involving CRM restructuring and training, with reported adoption and sales results. That makes Gamboa a more specific option when the desired campaign outcome is practical retention or CRM adoption work.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/