Overcharge or undercharge? Pricing your first 10 B2B customers

The short version. Undercharging hurts more than overcharging for your first 10 B2B customers. Charge a number that makes the buyer run a real approval path, then use narrow discounts only to buy speed, access, written feedback, or a longer term. Small checks create polite users and fake proof. Price is signal. A founder who cannot ask for a serious contract will also struggle to ask for implementation access and renewal intent from a busy executive who owns the budget.

Before you start:

• Err high enough to trigger a real approval path; instant yes usually means the price taught you little.
• Discount only for a written trade: speed, access, feedback, extra term, or reduced scope.
• Separate design partners from customers so a learning deal does not become your default package.
• Raise when the same buyer, pain, approval path, and budget owner repeat across early deals.

Step 1: Pick a single buyer and a painful workflow

Name the buyer. If your first 10 prospects include a VP of Sales at a SaaS company and an operations manager at a healthcare services firm, the same price will teach you almost nothing. The narrow ICP work in How to define your first ICP, uncomfortably narrow, then iterate matters here because price only becomes real inside a specific workflow.

Choose the workflow where money already leaves the building: agency fees, contractor time, support queues, enrichment tools, cloud compute, failed renewals. Pain needs a receipt. A first price should sit near an existing cost center because a buyer can trade budget from a vendor line faster than they can invent a new category for your deck. Done looks like: a named buyer, a narrow workflow, a current budget owner, and a reason to buy before the next planning cycle.

• Buyer job title
• Current vendor line
• Workflow owner
• Trigger event

Step 2: Ask, will this price force a real decision?

Make it uncomfortable. A price that clears without a manager or finance note often buys friendliness instead of commitment, and friendliness will poison your read on product value during the sales cycle. I would rather hear a buyer say the number needs approval than hear instant praise from someone who cannot move budget.

Pick a number you can say without apology and with a clean link to cost removed, revenue recovered, risk reduced, or hours returned. Discounts need fences. I allow a discount only when the buyer gives something concrete in exchange, such as faster legal review, executive access, a paid annual term, or precise product feedback in writing. The broader frame is that early pricing is an interview with finance, procurement, security, and the champion, with the spreadsheet serving as backup.

Record the approval path inside the opportunity while the call is still fresh, because pricing discipline disappears when notes become folklore. Memory lies. Done looks like: the buyer pauses, names the approver, sets a legal path, and keeps the conversation alive after the price lands.

• Annual package first
• Monthly fallback named
• Discount trade written
• Approval path recorded

Step 3: Quote the package before the discount

Quote the package. Your buyer needs to see the full contract shape before any concession, because a discount from a visible anchor teaches more than a cheap pilot with hidden future terms. When you lead with a pilot price, the buyer learns the floor before they understand the ceiling.

Use a simple order form structure: scope, start date, term, success criteria, support owner, data obligations. Clarity sells. When the buyer negotiates, keep the same package and change only the concession lever, because moving scope and price together makes the signal unreadable after procurement touches it. Done looks like: a quoted annual package, a named concession, a written trade, and a future price with no mystery.

Stripe Billing docs treat pricing model selection as a design decision across flat-rate, usage-based, tiered, and per-seat structures. Use that lightly. Your first 10 deals still need founder judgment because thin data turns every spreadsheet cell into theater when the buyer has a real workflow, budget owner, deployment risk, and deadline.

• Full price shown
• Concession reason
• Term on paper
• Success criteria attached
• Future price named

Step 4: Separate design partners from customers

Label the lane. A design partner pays less because they accept rough edges and scheduled feedback while helping you find the production package without pretending the current product is complete. This belongs beside Founder-led sales: how to land your first 10 customers, because the same founder can accidentally sell a product promise during a call and a services promise during the next.

Customers pay for an outcome now, even when onboarding still has founder fingerprints and messy documentation. Contracts draw the line. If the buyer wants roadmap influence, unusual service, special data work, or direct founder access, price those burdens inside the deal because gratitude will not cover the hours. Done looks like: every account is tagged as design partner or customer, with price logic attached in the CRM.

• Design-partner terms
• Customer terms
• Extra service priced
• Feedback calendar
• Production package named

Step 5: Test whether the price survives procurement

Procurement is useful pressure. If the prospect accepts instantly but misses legal deadlines and keeps finance away, you may have a pleasant sandbox rather than a customer who will fight for deployment. I care less about applause on the demo than the buyer's ability to pull a purchase order through their building.

Ask for the approval path early and listen for nouns: budget owner, procurement portal, security questionnaire, legal redlines, implementation lead, renewal date. Nouns beat vibes. A buyer who can describe the path usually knows where money sits, while a buyer who only praises the demo often cannot protect your price internally. The operating plan in How to build your first GTM strategy from scratch (seed to Series B) should carry those pricing signals.

Keep the price conversation connected to risk removal, because procurement has a cleaner job when the contract mirrors the business case. Paper matters. Done looks like: price has survived a real approver, a paper trail, an implementation owner, and a start date before you celebrate the close.

• Approver named
• Paper trail started
• Security path known
• Deployment owner visible
• Renewal date captured

What goes wrong

Fear gets expensive. A cheap customer with unlimited Slack access, custom reporting, data cleanup, weekly strategy calls, and emergency fixes can consume the same founder hours as a serious account while teaching less. Undercharging often hides inside the word partnership, especially when the buyer knows you want proof more than cash.

Confidence has a ceiling. If onboarding depends on manual data work and the champion has no budget authority, a giant quote creates a dead opportunity that looks like ambition in the CRM. Overcharging becomes theater when the product still needs founder labor for every deployment and the buyer lacks internal power.

Never sell apology. A vague founding-customer discount trains the buyer to wait and your team to cave while leaving finance with no reason to respect the renewal price later. The habit protects the next deal more than the current signature. Done looks like: discounts become rare, written, conditional, and easy to defend.

• Free pilot creep
• Custom work hidden
• Discount without trade
• Champion without budget

Step 6: Raise after a repeatable win

Patterns earn price. After your first 10 conversations start producing similar triggers, approval paths, implementation needs, and budget owners, the next quote should rise before your calendar becomes the constraint. For AI products, the pricing logic in How to price an AI product: VC consensus on seats versus usage versus outcomes matters when usage can outrun the contract.

Public API pricing pages from OpenAI and Anthropic are a reminder for AI founders: model cost sits under every usage-heavy contract, even when the buyer only sees a neat annual package. Margin is oxygen. A contract that sounds bold can still punish you if data volume, user behavior, workflow frequency, or model choice expands faster than the price inside the account.

Raise in clean jumps, with legacy terms protected for customers who gave early learning and kept their side of the bargain. Do not sneak. Tell new prospects the package has changed because the product, onboarding, support model, and risk removal have changed, then watch whether win quality improves inside the same segment. Done looks like: the next cohort pays more, complains less, implements faster, and renews from budget discipline.

• Repeated buyer pain
• Shared approval path
• Margin guardrail
• Clean price jump

Common questions

How much should I charge my first customers?

Charge the highest annual package you can defend against a current pain and a current budget line. If the buyer accepts without involving an approver, the price is probably too low to teach you much. If the buyer refuses before discussing value or deployment, the package may be ahead of proof. For the first 10, I care about approval-quality learning.

Should I give my first customers a discount?

Yes. Give a narrow discount only when the buyer gives something concrete back: faster legal review, executive access, a longer term, written feedback, or reduced scope. The discount should expire in the contract and in your mind. A founding-customer label with no trade becomes a permanent anchor.

Should I run a free pilot?

Rarely. Free pilots attract champions who like experiments and avoid budget conversations. A paid pilot can work when it has production data, a success criterion, an owner, and a path into the annual package. If the buyer cannot fund even that, the buyer may lack the power to make your first 10 meaningful.