The B2B, SaaS, GTM and Marketing Playbook: Strategy, Execution, and AI Skills

Sources:

• Jeanne DeWitt Grosser (COO Vercel, ex-CPO Stripe): Lenny's Podcast — GTM as a product, the GTM engineer, AI agents replacing SDR workflows, segmentation, sales comp, PLG
• Margot (SaaSMastery): First 100 SaaS customers, 3 mistakes, 5-step GTM system, scaling beyond 100
• Henry (GTM strategist): Go to market strategy framework, pre-launch/launch/post-launch phases
• sales-skills/sales (Claude Code skills): /sales-gtm, /sales-icp, /sales-positioning, /sales-competitive, /sales-forecasting, /sales-territory, /sales-reporting, /sales-coaching, /sales-enablement
• Claude Skills for SaaS: Complete Revenue Playbook, Claude Agents for B2B SaaS

This playbook combines the best strategic GTM thinking from operators who have built world-class go to market teams at Stripe, Vercel, and dozens of SaaS companies, with the Claude Code skills and agents that execute it. Every section has both the framework and the prompt to run it today.

What go to market actually means

Most people define go to market as marketing and sales. That is the tip of the spear. The complete definition: any function that touches a customer or makes a dollar.

This includes: marketing, sales, technical sales (sales engineers, platform architects), customer success, support, and partnerships. The problem in most companies is these functions run as a Venn diagram rather than an integrated lifecycle. Marketing pursues one segmentation. Sales pursues another. Support uses a third. None of them line up perfectly.

The opportunity right now: those 17 hyper-specialised go to market roles that emerged over the last decade are starting to collapse. AI handles the repetitive workflow layers that justified specialisation. What is left is a smaller number of higher-leverage roles that each own more of the end-to-end customer journey.

For most SaaS companies starting out, GTM means:

• Seed to $1M ARR: founder sales + early PLG + marketing for awareness
• $1M to $5M ARR: first AE or two, SDR function, defined ICP, outbound beginning
• $5M to $20M ARR: sales team with process, CS, outbound engine, enablement
• $20M+ ARR: segmented motion (SMB/mid-market/enterprise), RevOps, territory planning

The theme throughout: every function that touches a customer should share one segmentation framework and one view of the customer lifecycle from first awareness to five years retained.

Sources: Margot (SaaSMastery) + Jeanne DeWitt Grosser + Henry GTM framework

The three most common mistakes that kill GTM before it starts, the five-step system to fix them, and the Claude Code skills that execute each step.

The 3 mistakes that kill GTM before it starts

Mistake 1: Targeting too broad

"We help any business be more productive" is not a positioning statement. It is a death sentence for conversion. If you target everyone you reach no one. The fix is a specific ICP definition: one problem, one type of customer, one buying signal. Avoid this and you are already ahead of 70% of first-time founders.

Mistake 2: Building before validating

Spending 6-12 months building without talking to customers is the most common way to launch to nobody. The only way to validate whether people want your product is to have conversations with them before and during the build. Proof of interest (conversations, signups, or pre-payments) is not optional. It is the evidence that you are building something the market wants.

Mistake 3: Generic messaging

"The future of productivity" and "work smarter, not harder" communicate nothing specific. Generic messaging does not speak to your customer's pain. The fix: one clear promise with proof and why it matters to them. Invoice Ocean changed their headline from features to "Generate an invoice in less than 30 seconds" and conversion rate increased 2.5x. Specific beats generic. Always.

The 5-step GTM system for the first 100 customers

Step 1: Define your ICP with precision

Not "construction businesses" or "tech companies". A working ICP has four layers:

• Firmographics: company size, location, tools they use
• Buyer personas: who has authority to buy, who influences but does not control budget
• Pain points and desires: what keeps them up at night, their goals, fears, what they are suspicious of
• Disqualifiers: who looks right on paper but never converts

When you know exactly who you are talking to, every decision becomes easier: your messaging, your pricing, your feature prioritisation. Your first 10 customers are the raw material for building this profile. What do they have in common? That is your ICP.

Step 2: Nail your messaging

Messaging that resonates speaks directly to the ICP's pain points, not your product's features. Test different subject lines, headlines, and ad copy. Track who opens, what they click, and whether conversations start. Iterate until you find what works. Only then scale the outreach.

Step 3: Focused outreach for early adopters

Three creative tactics that outperform cold email blasting:

Social engagement first: find your ICP on LinkedIn or X, follow them, comment on their posts, like their content. When you eventually reach out, you are not a stranger. You are someone they have seen. Frame your outreach as: "I built something for people like you and I'd love your feedback."

Value-first outreach: do not send a generic pitch. Send something that solves a specific problem they have. A competitor analysis. A mini audit of their site. A rewrite of their worst-performing post. You are proving you understand their world before asking for anything.

Personalised video outreach: record a 4-5 minute Loom with their name, their website or LinkedIn on screen, and an explanation of exactly how your product solves their specific challenge. In a world of automated text emails, a personal video is almost impossible to ignore. Tip: find two decision makers at the same company, send both a personalised video, and mention in each that you also sent one to their colleague. This creates a gossip effect that compels them to watch.

Step 4: Landing page as a customer development tool

Your landing page is not just a conversion point. It is a validation tool. Send mockups to your email list. Ask for feedback. Test what they respond to. When your ICP visits the site, it should feel like the copy is speaking directly to their exact pain. Social proof and a strong CTA are table stakes. The page should evolve as you learn more.

Step 5: Customer feedback loop

The secret skill that separates good founders from great ones: obsess over feedback in the early stages. Ask what is missing, what frustrates them, what features they wish existed. They will help shape your messaging more than any amount of internal brainstorming. Founders who think they never have to talk to customers because they are building software are building products that miss.

The 3-phase GTM framework

Phase 1: Pre-launch

Goal: build awareness and credibility before you need to sell.

How: educate your target market through content and community. Build an organic following who will beta test your product and spread the word when it goes live. For B2B SaaS targeting parents of children with learning disabilities (Henry's example), this means: blog + YouTube + Quora for search-led content, Facebook groups for community infiltration, Slack group for owned community. Target: 1,000 community members and 10 influencer partners before launch. Timeline: 3 months.

Phase 2: Launch

Goal: drive user adoption.

How: announce across all organic channels you built in Phase 1, amplify with paid advertising (search ads on Google and YouTube target in-market searchers), activate your influencer partnerships for review content, post on Product Hunt/Reddit/TechCrunch for press. Timeline: 1 month. Target: 10x the community members from Phase 1.

Phase 3: Post-launch

Goal: optimise and scale.

How: scale down on channels that are not working, scale up on channels that are, test 1-2 new channels. For B2B: test partnership channels (selling to businesses that have your ICP's consumer base, like pediatrician clinics or schools). Simultaneously: take customer feedback and start working on V2. Your first cohort is your product team.

Claude Code skills for GTM strategy

# Install
npx skills add sales-skills/sales

# Build a GTM plan for a new segment or motion
/sales-gtm

Help me build a GTM plan for [new product / new market / new segment].

Context:
- What I sell: [one sentence]
- Current ICP: [describe]
- New target: [new segment or market]
- Resources: [N reps / N budget / timeline]
- Goal: [N meetings / N pipeline / N revenue in N months]

Return:
1. ICP for this motion: titles, company types, signals
2. Channel recommendation: which channels and why
3. Message framework: what to lead with and what to avoid
4. 90-day plan: what to do in the first 30, 60, and 90 days
5. The one metric to track that tells me if this is working

# Define or refine your ICP from pipeline evidence
/sales-icp

Refine my ICP using pipeline evidence.

Current ICP: [paste current definition]

Evidence:
[paste: won deal patterns / lost deal reasons / churn data / best customer profiles]

Return:
1. Where my current ICP is too broad
2. Where it is too narrow
3. Disqualification signals I am missing
4. The 3 highest-signal buying indicators in this data
5. Revised ICP with updated criteria

# Sharpen your positioning
/sales-positioning

What I sell: [one sentence]. ICP: [describe]. Main competitors: [list 2-3].
What customers love most: [describe]. Objection I hear most: [describe].

Return: differentiated one-sentence positioning, 3 proof points, how to frame against each competitor,
the one thing to stop saying, the one thing to start saying.

Source: Jeanne DeWitt Grosser (COO Vercel, ex-CPO Stripe) + sales-skills/sales

Segmentation is not a go to market thing. It is a company thing. The best companies teach it to every new hire in week one.

How Stripe built their segmentation model

Most companies segment by size only: small, medium, large. That is a starting point, not a model. Size tells you how complex the buying process is (single decision maker vs. committee). But it does not tell you what to sell them or how to message it.

Stripe added two more axes:

Axis 2: Growth potential. Stripe is consumption-based. A company growing at 200% year-on-year is worth more investment than one growing at 8%. They weighted their targeting and resources accordingly.

Axis 3: Business model. B2B (wires, recurring billing), B2C (consumer payments, Apple Pay), platform (Stripe Connect), marketplace (Connect). Each business model implied a different product, different value prop, and different persona.

Vercel's model:

Axis 1: Size (SMB to enterprise)

Axis 2: Traffic (Crux rank from Chrome data). A company with 3,000 employees that is a top-25 traffic site goes into enterprise, not mid-market.

Axis 3: Workload type. E-commerce (product listing pages, order management systems), crypto (AWS-native stack), enterprise SaaS (migration-heavy). Each requires different language, different reference customers, different sales motion.

The result: when Vercel analysed their segmentation data, they found strong penetration in e-commerce but weak penetration in enterprise SaaS. Not surprising in hindsight: enterprise SaaS companies built their stack before Vercel existed and migrating 2 million lines of code is a big lift. That insight directed the new motion: target enterprise SaaS companies adding AI-native features to existing apps, where the barrier to entry is lower.

The three-attribute rule: beyond three segmentation attributes, you cannot reason about it practically. Five sellers cannot cover 25 different segments. Start with three: something about size, something about buying behaviour or growth, something about what they actually buy.

Segmentation as a company-wide framework

Segmentation fails when it is only known by sales. Every new hire at Vercel receives a segmentation briefing in week one. Every product manager knows which segment their new feature targets. When product and sales use the same language, you stop getting the endless "sales asked for this feature" versus "product deprioritised it" conflict. They are having the same conversation in different rooms.

KYC (know your customer) is a company value at Vercel, not a sales competency. That is the shift to make.

Segmentation Claude Code skills

# Build or audit your segmentation model
/sales-territory

Help me build a segmentation and territory model for my sales team.

Team: [N reps]. Market: [describe TAM]. Current segmentation: [describe or say none].
Rep capacity: [N accounts per rep].

Return:
1. Segmentation model: 2-3 axes with criteria per axis
2. How each segment maps to a different product, value prop, or sales motion
3. Account tiering within each segment: tier 1/2/3 criteria
4. Coverage rules per tier
5. The attributes most correlated with high ACV in my current win data

# Competitive teardown to find your segmentation wedge
/sales-competitive

Run a competitive teardown on [COMPETITOR].
Pull: their homepage, pricing, G2 and Capterra reviews from last 12 months, recent LinkedIn messaging.

Return:
1. Their positioning and which segments they own
2. The segments they are weak in (evidence from reviews and complaints)
3. The segment I can own that they cannot credibly claim
4. 3 landmine questions for discovery that surface their weaknesses
5. The one claim I can make that they cannot make back

Source: Jeanne DeWitt Grosser (COO Vercel) + existing Claude Agents for B2B SaaS lead mag

The GTM engineer role is not new. Jeanne tried to build it at Stripe in 2017 with a data scientist and a systems admin. It did not work because AI did not exist. Now it does. One GTM engineer spent 25-30% of their time over 6 weeks and moved 10 SDRs to 1.

What the GTM engineer actually does

The GTM engineer's job: go through each function in go to market (SDR, AE, CS, support, marketing), map every workflow they run, and determine where AI is better placed than a human to do that task. Then build the agent that does it.

The operating principle: shadow the highest performer in each function first. Watch every tab they have open, every tool they switch between, every judgment call they make. That is the workflow you are encoding. You are not guessing at a generic process. You are capturing the best-practice version.

What works and what does not:

• Start with inbound, not outbound. Inbound qualification is the most legible workflow: it is replicable, mostly deterministic, and produces structured output (qualified / not qualified, message draft). Outbound is messier because personalisation requirements are higher, especially in enterprise.
• Start with SMB, not enterprise. A single decision maker at an SMB means less customisation required. Enterprise deals involve multiple stakeholders, multiple layers of an org chart, and signals that require genuine human judgment.
• Always keep a human in the loop for first messages. The agent makes the call (qualify / not qualify, draft the message), a human reviews and approves, a human hits send. The approval rate is the feedback signal that trains the agent over time.
• Track the same KPIs you would hold an SDR to. Lead-to-opportunity conversion rate. Number of touches to convert. Time to convert. If the agent matches human performance on these, you are ready.

Vercel's result: 10 SDRs running inbound qualification became 1 SDR doing QA on the agent. The 9 freed up moved to outbound. The agent cost $1,000 per year to run. Previous salary spend: over $1M. Lead-to-opportunity conversion held flat. Touches to convert decreased because the agent responds instantly where humans let leads sit overnight.

The profile of a great GTM engineer

Vercel's first three GTM engineers were sales engineers who had been front-end developers before moving into sales. That combination matters. You need someone who:

1. Has GTM experience. They need to know what good looks like. When Jeanne reviewed her lead agent's output, she saw responses she would not have sent. The agent was modelled on the best SDR on the team, who had two years of experience. The GTM engineer reviewing QA output needs 20 years of judgment, not two.
2. Can write code. These agents are being built on Vercel's workflow SDK. They write actual code. Vibe coding tools help but someone needs to architect the system.
3. Geeks out on sales even if they come from engineering. If they are not going to read sales books and learn what best practice looks like, they will encode mediocre workflows into agents that run at scale. Mediocre at scale is worse than mediocre manually.

The alternative profile if you cannot find a technical sales engineer: a technically-minded AE or SDR who is willing to learn to build. They already know the workflow. You are teaching them to encode it.

Building your first GTM agent: the lead qualification agent

The inbound lead qualification agent is the highest-leverage first build for most SaaS companies. It is legible, repeatable, and the ROI is immediate.

Architecture:

# Stage 1: Trigger and enrich
# Workflow triggers from CRM webhook when new inbound lead arrives
# Agent enriches via Apollo or ZoomInfo: company size, tech stack, revenue, job title

# Stage 2: Score against ICP
# Claude receives enriched data + ICP system prompt from CLAUDE.md
# Returns JSON: {icp_score: N, reasoning: "...", tier: "hot|warm|cold"}

# Stage 3: Draft response
# If score 70+: draft personalised qualification email referencing specific signals
# If score below 70: draft nurture email or disqualify

# Stage 4: Human review
# Human sees: lead data, score, reasoning, and drafted email
# Human approves or edits before anything sends

# Stage 5: Log and learn
# Every approved/rejected/edited draft becomes training signal
# Agent improves over time as rejection patterns accumulate

Build using the n8n-workflow-builder skill:

"Build a lead qualification workflow in n8n.
Trigger: new inbound form fill in HubSpot.
Enrich: Apollo API for company data and contact data.
Score: Claude Sonnet with this ICP prompt: [PASTE ICP CRITERIA]
Output: JSON with score, tier, reasoning.
Route: score 70+ to Slack #inbound-hot + draft email. Below 70: nurture sequence.
Human review gate before any email sends.
Log everything to Google Sheets."

The dealbot: run agents against your Gong transcripts

Vercel's most powerful internal agent: the dealbot. It reads every Gong transcript, Slack interaction, and email for an opportunity and surfaces insights the AE missed.

What it found on Vercel's biggest Q2 loss: the AE logged the loss as "price". The agent read every interaction and concluded the actual reason was failure to engage the economic buyer and failure to demonstrate ROI convincingly. Two completely different problems requiring two completely different fixes.

Now it runs in real time. Every deal channel in Slack gets a feed of insights: you are this far into the sales process and have not talked to the economic buyer; your last call with the economic buyer did not land well, here is what to consider for follow-up.

# Build the dealbot using the automation-builder agent
@automation-builder Build a dealbot that reads all Gong transcripts, Slack messages,
and emails for each open opportunity.

For each deal, surface:
1. Whether the economic buyer has been engaged
2. Whether ROI/TCO has been discussed and whether the prospect seemed to buy it
3. Whether a next step with a specific date was agreed
4. The top risk factor in this deal
5. One recommended action for the AE this week

Post insights to the deal's Slack channel once per day.
For lost opportunities: run a retrospective. Why did we actually lose?
Compare to what the AE logged as the loss reason.

Also use it for enablement:

"Run the dealbot across this week's calls after we launched [NEW FEATURE].
Find: where did reps struggle to explain it? Where did objections go unanswered?
Diagnose the gaps as bugs in our GTM process.
Return a prioritised fix list: content to add to the objection handling guide,
concepts to add to the discovery guide, changes to the demo."

Source: Jeanne DeWitt Grosser + Margot (SaaSMastery) + sales-skills/sales + Claude Skills for SaaS

When to hire your first salesperson, how to think about PLG vs sales, what segmentation tells you about your motion, and how to build a comp plan in a moment when annual planning is harder than it has ever been.

When to hire your first salesperson

$1M ARR is the widely cited threshold and it is roughly right. But the real signal is not the number. It is repeatability.

Founders are exceptional salespeople. They convinced investors to give them money. They will convince early customers to buy on vision. But if you reach $1M ARR and your customers look nothing like each other, you still have an evangelist sale. You do not have a sales process. You have a charismatic founder doing something that no salesperson will be able to replicate.

You are ready to hire your first salesperson when:

• You can write down your ICP in one sentence
• You know which questions reveal the buying decision fastest
• You know the 3 most common objections and how you handle them
• You have at least 3-5 customers who look like each other

When you do hire: actually hand over the reins. Enable them with your content, your discovery questions, your objection handling. But do not hand over entirely. Stay connected to customers. You still have R&D to do to understand where the product resonates next and where you are getting stuck as you scale.

PLG vs sales: the real answer

PLG is not a replacement for sales. It is a starting motion that has a ceiling.

When PLG makes sense: you are targeting developers or individual contributors who can trial and buy independently. Your product delivers value in a self-serve flow. Your ACV is low enough that the economics of a human-assisted sale do not work.

When it does not: you are going after global 2000 companies from day one. PLG will not unlock eight-figure deals. Sierra (the enterprise AI company) would not use PLG. They are selling to procurement committees.

The failure mode is not PLG itself. It is waiting too long to add sales. Building a replicable sales process takes time. Adding outbound on top of inbound takes even longer to turn into a predictable engine. Companies hit walls when they ride PLG past the point where it can grow them and do not have the sales infrastructure in place to pick up the slack.

Jeanne's take: there is probably no $100B company that is PLG only. Every company at scale has a sales function. The question is when you add it and how you orchestrate it with your PLG motion.

For pricing: Vercel's pricing evolution is instructive. They started with flat-rate SaaS-like pricing, moved to consumption-based, then unbundled enterprise features. They discovered that 50% of their enterprise SKU users were startups, which meant enterprise features that startups wanted were locked behind a sales-assisted motion. Moving those features to self-serve drove PLG growth.

Sales comp in a world where you cannot plan 12 months out

Jeanne's hot take: traditional sales comp plans are built on 12-month commitments at a moment when most companies cannot see 12 months ahead with confidence.

Vercel wrote sales plans for the year without knowing the AI Cloud was coming. The AI Cloud launched mid-year. Every plan was written around the wrong product.

The tension: sales comp motivates through pay-for-performance which requires committing to what you value 12 months in advance. But the thing you value most may not exist yet at the start of your fiscal year.

Design principles that hold regardless:

• Simple: a rep should be able to calculate their own commission in under 2 minutes
• Aligned: what earns commission should be what the company actually needs
• Accelerated: above-quota performance should pay disproportionately more to drive upside
• Protected: below a floor (usually 50% of quota) no variable pay, to focus effort on the right opportunities

Hiring philosophy: mix traditional sales profiles with consultant or banking backgrounds. Sales professionals teach the skill (discovery, objection handling, pipeline discipline). Consultants teach the analytical lens (P&L thinking, TCO analysis, executive stakeholder management). They learn from each other and the combined team operates like general managers, not just quota carriers.

Claude Code skills for sales ops and comp

# Design a sales comp plan
/sales-compensation

Help me design a sales compensation plan.

Role: [SDR / AE / AM / Sales Manager]. OTE: [$N]. Revenue target: [$N/year].
Deal size: [$N average]. Sales cycle: [N days]. Stage: [seed / A / B / growth].

Return:
1. Recommended base/variable split and why
2. Commission structure: rate, when it kicks in, how calculated
3. Accelerators: what triggers them and how much
4. Spiff ideas for specific behaviours to incentivise
5. What to avoid: structures that cause sandbagging or gaming

# Hire your first salesperson
/sales-hiring

Help me hire a [SDR / AE / Sales Manager].

Role context:
- Company stage: [describe]
- What success looks like in 90 days: [describe]
- Must-have: [describe]
- Red flags: [describe]

Return: JD, interview scorecard, 5 interview questions, take-home exercise, reference check questions.

# Build a revenue forecast
/sales-forecasting

Target: [$N]. Current pipeline: [$N, by stage if possible]. Avg deal: [$N]. Avg cycle: [N days].

Return: bottoms-up forecast, gap analysis, coverage ratio, top 3 deals, pipeline needed this month for next quarter.

Source: Jeanne DeWitt Grosser (Stripe and Vercel GTM design) + Henry GTM framework + Margot (scaling tactics)

Go to market as a product means mapping the full customer journey from first awareness to five-year retained and designing experiences at every step rather than just optimising individual touchpoints.

The buying journey as a product

Jeanne's thesis, formed while watching software commoditise: when technical differentiation narrows, the buying experience becomes the differentiator.

80% of customers buy to avoid pain or reduce risk. 20% buy to increase upside. Startup founders love to sell the art of the possible. That resonates with other founders. For everyone else (especially enterprise), the motivating question is: what risk am I avoiding by buying this? What happens if I do not buy this and my competitor does? What does it do to my revenue target next quarter?

The flip from feature-selling to risk-reduction selling: instead of "here is what our product does" lead with "here is the specific risk you are running right now and here is how you quantify it". Vercel's outreach includes real data on a prospect's Core Web Vitals performance relative to peers. They leave with an insight they did not have before, regardless of whether they buy. That builds trust that accumulates into a purchase decision years later.

The Stripe whiteboarding session: instead of starting the first call with discovery questions (which feels like being quizzed), Stripe would ask prospects to draw their payments architecture. Through that collaborative session, Stripe learned everything they would have learned from a structured discovery call, but the prospect felt like a partner building something rather than a subject being interviewed. They left with an asset. That is experience design.

Principle: add value at every touchpoint regardless of whether the customer buys. If you miss a buying cycle, you plant a seed. Jeanne saw customers at Stripe that lost a deal and then came back four years later and bought. The quality of that earlier interaction is what brought them back.

Delivering unique insight as the outreach hook

The most effective outreach Vercel does is not about their product. It is about the prospect's performance.

They pull Core Web Vitals data and benchmark the prospect against peers. They share what they find. The prospect gets a specific and actionable insight about their current state before they have had a single conversation about pricing. That makes Vercel associated with knowledge and expertise, not with selling.

Two other insight-driven outreach approaches Vercel uses:

• MCP server education: explaining when a company's product would benefit from an MCP server, tailored to their specific architecture. Most companies have not thought through this yet.
• AEO (Answer Engine Optimisation): helping prospects understand how their SEO translates to AI search visibility. Tangential to Vercel's core product but deeply relevant to what their ICP is thinking about right now.

The pattern: find the insight that is genuinely useful to your ICP, that you are credibly in a position to deliver, and that connects in some way (even loosely) to what you sell. Deliver it before you ask for anything.

Best practice guides over generic docs

Stripe had the best marketplace implementation knowledge on the planet. Lyft, Instacart, DoorDash all ran on Stripe. When a new marketplace came to evaluate Stripe and a rep said "check our docs", the prospect pushed back: every marketplace runs on Stripe, I do not want generic docs, I want you to tell me the best way to set up payments for a marketplace specifically.

So Stripe built best practice guides: not generic documentation but opinionated, prescriptive playbooks for specific business model types. If you are a marketplace, here is exactly how to set up Connect. If you are a B2B SaaS, here is how to implement billing. These are not feature descriptions. They are implementation roadmaps built on accumulated pattern-matching from working with thousands of similar companies.

For SaaS companies at any stage: think about what you know about your ICP's specific situation that they do not know about themselves. What does your best customer do that your worst customer does not? That knowledge, packaged as an opinionated guide, is a better asset than a feature list.

Claude Code skills for marketing content and enablement

# Build your content strategy
/sales-content

Help me build a content strategy for [blog / newsletter / lead magnets / LinkedIn].

ICP: [describe]. Core topics: [list 3]. Goal: [pipeline / thought leadership / SEO / email list].
Posting frequency: [N per week/month].

Return:
1. Content pillars: the 3 topics and angles to own
2. Format mix: which types in what ratio
3. Content calendar for next 30 days
4. The 3 pieces to create first and why

# Build a landing page optimised for conversion
@landing-page-optimizer

Audit our [homepage / product page / landing page].
Conversion rate: [N%]. ICP: [DESCRIBE].
Rewrite the hero section and give me 3 headline variants.
Formulas: [Outcome] in [Timeframe] for [ICP] / Stop [pain]. Start [outcome]. / [Customer] got [result] using [product].

# Generate insight-led outreach (the Vercel model)
@analytics-engineer

Build a pre-outreach data brief for [PROSPECT COMPANY].
Pull:
- Their Core Web Vitals score from public sources
- Their Lighthouse performance score
- Any public AEO data (do they appear in AI-generated answers for [THEIR CATEGORY]?)
- Benchmark against 3 direct competitors

Return: one-page brief I can reference to lead my outreach with a specific data-led insight,
not a generic pitch.

# Build a best-practice guide for your top ICP use case
"Build a best-practice implementation guide for [BUSINESS MODEL TYPE: marketplace / B2B SaaS / e-commerce] using [YOUR PRODUCT].

Based on our experience with [N] [BUSINESS MODEL] customers, write an opinionated playbook:
1. The 3 most common setup mistakes and how to avoid them
2. The recommended architecture for this business model
3. The specific features to configure first and why
4. What the best implementations have in common
5. A checklist for getting to value in under 30 days"

Source: Margot (SaaSMastery) + Jeanne DeWitt Grosser + sales-skills/sales + Claude Agents for B2B SaaS

Once you have the first 100 customers, how you scale is determined by how well you have built the foundation. Social proof, referrals, content, and a brand that attracts instead of chases.

Scaling beyond 100 customers

Three tactics that work at scale:

1. Build a library of success stories and lead with them

Do not say "our client got to $700k MRR". Say "one of our clients in your industry was struggling with exactly the same problem. Here is what they did and here is what happened". Storytelling with specifics (industry, problem, result) outperforms abstract stats. Social proof is the strongest sales tool you have at this stage. Use it in outreach, on your site, in your discovery calls.

1. Build a referral programme into the onboarding process

Make it easy and incentivised. A simple automated email sent after 90 days: "Love the product? Share it with a friend and you both get X." Do not wait for customers to refer spontaneously. Build the ask into the journey.

1. Position through thought leadership

Content is the long game. Posts, videos, short-form, newsletters: all of it should solve your ICP's pain, not talk about your product. Build authority. In the long term, inbound content drives down CAC and scales without proportional headcount growth. This is where PLG and content intersect.

The metrics that matter at each stage

Stage	Primary metric	Secondary metrics
0-10 customers	Conversations per week	Feedback quality, repeat engagement
10-100 customers	ICP conversion rate	Time to close, objection patterns
100-1k customers	MRR growth rate	NRR, CAC payback, churn rate
1k+ customers	NRR	Pipeline coverage, win rate by segment

Key metric insight from Jeanne: outbound is only a predictable engine after 12-18 months of investment. Companies underestimate how long it takes to build this into a reliable revenue source. If you are starting outbound now, model it as a Q4 or Q1 next year revenue driver, not this quarter.

Claude Code skills for scaling and reporting

# Build your email nurture and referral sequences
/sales-email-marketing

Help me build an email nurture sequence for [leads / trial users / churned customers].

Audience: [describe]. Goal: [call / trial conversion / reactivation].
List size: [N]. Platform: [Mailchimp / HubSpot / other].

Return: sequence structure, email copy per touch, segmentation rules, success metrics.

# Build your sales reporting framework
/sales-reporting

Help me build a sales reporting framework.

Audience: [myself / team / board]. CRM: [HubSpot / Salesforce / Apollo].
Key metrics: [list]. Current problem: [what I cannot see or what takes too long].

Return:
1. The 5 metrics that matter most for my motion
2. Dashboard structure by level (rep / manager / exec)
3. Weekly / monthly / quarterly report templates
4. How to set these up in my CRM
5. The one leading indicator that predicts revenue 30 days out

# SaaS metrics health check
python3 finance/saas-metrics-coach/scripts/metrics_calculator.py \
  --mrr [N] --customers [N] --churned [N] --json

# Board-ready metrics brief
"Using saas-metrics-coach:
ARR: [$N]. Growth rate: [N%]. NRR: [N%]. CAC payback: [N months]. Burn multiple: [N].
Return: all 7 core SaaS metrics vs benchmarks, what is healthy, what is a warning sign,
the narrative to lead with in the board meeting."

Every Claude Code skill and agent from this playbook in one place.

Install all GTM skills

# Core GTM skills (sales-skills/sales)
npx skills add sales-skills/sales

# Marketing and growth skills (coreyhaines31)
npx skills add coreyhaines31/marketingskills

# AI sales team (5 parallel agents)
curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/ai-sales-team-claude/main/install.sh | bash

# GTM agents (67 plugins)
/plugin marketplace add gtmagents/gtm-agents
/plugin install sales-prospecting
/plugin install sales-pipeline
/plugin install account-management
/plugin install content-marketing
/plugin install email-marketing

# Analytics and RevOps
/plugin install marketing-skills@claude-code-skills
/plugin install finance-skills@claude-code-skills

Strategy skills quick reference

/sales-gtm     → Build a GTM plan for a new segment, motion, or market
/sales-icp     → Refine ICP from pipeline evidence
/sales-positioning  → Sharpen differentiation and competitive framing
/sales-competitive  → Competitive teardown from G2, Capterra, LinkedIn, website
/sales-forecasting  → Bottoms-up revenue forecast, gap analysis, coverage ratio
/sales-territory    → Segmentation model, account tiering, quota allocation
/sales-compensation → OTE structure, base/variable, accelerators, spiffs
/sales-hiring       → JD, scorecard, interview questions, take-home exercise
/sales-onboarding   → 30-60-90 ramp plan, milestones, certification gates
/sales-reporting    → Dashboards, KPIs, board metrics, leading indicators
/sales-coaching     → Call review, MEDDIC scoring, coaching frameworks
/sales-enablement   → Playbooks, battlecards, talk tracks, one-pagers
/sales-linkedin     → Connection requests, DM sequences, engagement ladders
/sales-content      → Content pillars, calendar, format mix
/sales-email-marketing  → Nurture sequences, segmentation, success metrics

Agent quick reference

# Full prospect research (5 parallel agents)
/sales prospect https://[target-company.com]

# Pre-call brief
/sales prep https://[target-company.com]

# Lead qualification agent (build with n8n-workflow-builder)
"Build a lead qualification workflow. Trigger: HubSpot form fill.
Enrich: Apollo. Score: Claude with ICP prompt.
Route: 70+ to Slack + email draft. Below 70: nurture.
Human review gate before any send."

# Dealbot (build with automation-builder)
"Build a dealbot that reads Gong transcripts, Slack, and emails for each open opportunity.
Surface: economic buyer engagement, ROI discussion quality, next step quality, top risk.
Post daily to deal Slack channel."

# Pipeline health check
/sales-pipeline
[PASTE PIPELINE DATA]
Flag stalled deals, top 5 to focus, top 3 to drop, next action per priority deal.

# SaaS metrics health check
python3 finance/saas-metrics-coach/scripts/metrics_calculator.py --mrr [N] --customers [N] --churned [N] --json

# Landing page conversion audit
@landing-page-optimizer
Audit [URL]. Conversion rate: [N%]. ICP: [DESCRIBE]. Rewrite hero. 3 headline variants.

# Insight-led outreach data brief
@analytics-engineer
Build a pre-outreach data brief for [COMPANY].
Benchmark their performance against 3 competitors. One specific insight to lead my outreach with.

The daily GTM routine with Claude Code

# Morning (10 minutes)
"Run my daily signal digest.
Pull: job change alerts from Sales Nav, LinkedIn engagement from yesterday's content,
funding news for ICP-matching companies.
Return: prioritised list, top 3 to act on today, drafted connection notes for Tier 1 signals."

# Pipeline review (Monday, 20 minutes)
/sales-pipeline
[PASTE PIPELINE EXPORT]
Flag no activity 7+ days. Top 5 to focus. Top 3 to consider dropping. Next action per deal.

# Weekly content (30 minutes)
/sales-linkedin-content
Topic: [FROM WEEKLY BRAINDUMP]. ICP: [DESCRIBE]. Post type: [insight / story / list / contrarian].
Hook must be a specific statement. No em dashes. Under [N] words.

# Friday review (15 minutes)
"Review this week:
- Email campaign: open rate, reply rate, positive replies
- LinkedIn: which posts got the most engagement from ICP-matching accounts
- Outreach: connections accepted, conversations started
- Pipeline: deals moved, new opportunities
The one thing to do differently next week."