Sales pipeline stages for early startups

Sales pipeline stages are the shared grammar for what progress means on an early B2B deal. Fuzzy grammar produces hopeful forecasts. Clear exit contracts turn a weekly review into a short proof check.

This is a definition playbook for the columns on that map: which ones to keep, what evidence advances a deal, and how to teach the ladder to a new hire. It is not a field-freshness or permissions manual. Product labels cited once below match Leads & Sales CRM.

Why sales pipeline stages need an exit contract

A column title is not a contract. "Negotiation" can hold warm chats that lack a next proof point. YC guidance for young companies keeps returning to concrete learning from real buyers (Y Combinator Library, checked 1 Nov 2026). Exit contracts are how that concreteness shows up in revenue review.

Write three lines per column:

1. Entry condition: what must already be true.
2. Exit evidence: what proof advances the opportunity.
3. Advancing role: who may move it.

If those lines take more than a few minutes, the column is doing two jobs. Split it or delete it. Adding columns later is cheap. Living with fake motion is expensive.

Treat the map as a teaching artifact. Someone who joins next month should read the contracts and know what "Qualified" means without tribal lore. If two co-founders disagree on whether an opportunity sits in Evaluation, the wording is unfinished. Resolve it in writing the same day.

Evidence ladder for early B2B columns

Most young teams need five to seven live columns plus two finals. The ladder below is an example pattern for illustration only, not a benchmark. Swap the proof types for whatever your market actually produces.

Final columns must stay final. Moving an opportunity into a final win or loss column updates outcome status and metrics. Never park soft pauses there.

Services teams may insert "Scope agreed." Product-led teams may collapse Evaluation into product-qualified behavior. The rule stays stable: one unique exit evidence type per live column.

Exit evidence should be copy-pasteable. If the only proof lives in someone's memory of a call, it will not survive a busy week. Prefer artifacts: a quote in notes, a proposal file, a trial workspace link, a signed order. Memory-only advances are how ladders quietly become fiction.

When you coach a new hire, pick one live opportunity and walk the three contract lines together. Ask them to propose the next proof before they propose the next meeting. Meeting-first habits refill calendars without filling the evidence ladder. Proof-first habits shrink the number of soft columns you need.

If your market sells through procurement, add a single procurement-ready column only when legal or security questionnaires actually begin. Do not invent that column because a blog post about larger companies mentioned it. Empty enterprise columns teach people that titles matter more than proof.

Worked mini-example (illustration only)

Example numbers for illustration only, not research findings. Suppose a two-person founding team has 11 open customer conversations.

• 4 sit in New with no owner → they are not yet on the ladder; assign owners before debating forecast weight.
• 3 are Qualified with ICP notes but no buyer quote → they are not Discovery done.
• 2 have a priced proposal and a Friday decision day → Evaluation is honest.
• 1 has a verbal "yes" with legal still silent → keep it in Evaluation, not Verbal commit.
• 1 died on budget → Closed Lost with reason no budget, not an endless Negotiation column.

After rewriting exit contracts, the same 11 conversations might occupy five live columns instead of nine. Empty decoration columns disappear. Example outcome only: review time drops because each advance must cite a proof line.

Another illustration-only pattern: when eight of eleven opportunities lack a buyer quote, discovery quality, not pricing - is the bottleneck. Fixing quotes before renaming columns prevents a cosmetic reorg that changes nothing about learning speed.

What the middle columns should force you to learn

Qualified is a gate, not a compliment. List ICP attributes you will not waive. Friendly chats that fail fit should leave the active ladder quickly.

Discovery done exits on language, not meeting count. Same-day notes should hold a short buyer quote naming the broken workflow and the metric they already watch. That quote becomes the spine of the proposal. If you cannot write it in two sentences, the exit failed.

Evaluation requires an artifact under review: trial workspace, scoped pilot, or priced proposal. "We will think about it" fails the contract. Name the decision day and the people who must say yes. Attach one proposal version so debate points at a single file.

Verbal commit means commercial terms plus a named signer. Strong-call optimism is still Evaluation. If paperwork stalls across two review cycles, move backward or record the loss with a reason instead of protecting a vanity column.

Connect people and companies in Contacts so one buyer is not three half-filled records. Duplicate humans inflate counts without adding proof.

Buyer quotes beat adjective-heavy CRM notes. "Interested" and "hot" are not exit evidence. "We still build the weekly roster in a spreadsheet and miss two clinics every Monday" is exit evidence for Discovery done. Train the team to prefer sentences a stranger could understand next quarter.

Procurement-heavy cycles also tempt teams to add "paperwork" columns that mean everything and nothing. If the only change is that a PDF moved between inboxes, keep the opportunity in Evaluation until a signer is named. Paper motion without a signer is not Verbal commit.

Also watch for "meeting motion without proof": three discovery calls with no quote still fail Discovery done. Calendar activity is not exit evidence. The contract cares about language you can paste into a proposal.

Separate motions deserve separate ladders

Inbound product-led motion, outbound enterprise motion, and partner referrals rarely share exit evidence. The customer CRM lets each funnel keep its own columns so those motions can diverge.

Shape one ladder per motion. Draft names beside live opportunities first so standup language matches the board, then enter them with Manage Pipeline Stages (manage-stages in the actions menu). Tick the final win or loss flags only on true finals.

Keep investor conversations in Fundraising CRM. Docs describe that Investor CRM as separate from customer selling. Different buying evidence belongs on different ladders. Two products with different committees usually need two ladders; tags are labels, not a substitute for exit contracts.

Align Qualified with the ICP story in GTM Planner so the gate matches the market you claim to serve.

When a second motion appears, resist stuffing it into the first ladder with clever tags. Tags help filtering. They do not create a second exit contract. A partner motion that never "closes won" like a subscription should not borrow subscription finals.

Habits that keep the grammar useful

Run a short weekly evidence review:

1. No advance without speaking the exit proof aloud.
2. Any next action older than seven days gets a new date or a recorded-loss reason.
3. Forecast weight starts at Evaluation unless history justifies earlier columns.

Require Deal Value and Expected Close Date on live opportunities before arguing math. Missing commercial facts turn the meeting into storytelling.

Change only one major column definition per week. Renaming everything at once resets habits and invites parking opportunities wherever feels convenient. Ship New through Discovery contracts first, live one review cycle, then tighten Evaluation and Verbal commit so learning stays attributable.

Collapse mirror columns ("Demo scheduled" / "Demo completed") when they share one owner and one learning job. Skip copying a nine-step enterprise ladder when a handful of opportunities cannot fill it.

Co-founders sometimes keep extra columns as a compromise when they disagree. Compromise columns without unique exit evidence are worse than picking one definition. Pick the stricter contract for two weeks, measure how many opportunities actually meet it, then loosen only if the data says the bar was unrealistic. Example for illustration only: if zero of ten opportunities ever reach a "Security review" column, delete it until a real questionnaire appears.

Publish exit contracts where the team already reads weekly notes. In the first review after a rewrite, ask each owner to defend one advance and one non-advance. The non-advance conversation is usually where fuzzy columns surface. Capture sharper wording the same day.

Time-box disagreement. If co-founders cannot agree on Evaluation's exit evidence in fifteen minutes, pick the stricter proof for the next ten opportunities and revisit with data. Endless definition debates are themselves a stall pattern. The ladder exists to make progress inspectable, not to win rhetorical points.

Seasonality also matters. End-of-quarter calendar pressure tempts teams to advance opportunities on vibes. Write the exit contract so a stressed Friday review still demands the same artifact it demanded on a calm Tuesday. Consistency under pressure is the point of having a written grammar.

Install path (and exact product labels)

1. List every open opportunity and write the real next proof required, ignoring today's titles.
2. Collapse columns until each remaining one has a unique exit contract.
3. From the sidebar open Marketing → Leads & CRM, open the funnel, then use Manage Pipeline Stages to rename and reorder columns, then mark finals.
4. Place each opportunity where its evidence supports, including backward moves.
5. After two weeks, fill missing Deal Value or Expected Close Date fields before debating precision, and delete any column that still lacks unique exit evidence.

UI labels such as Create Pipeline, Manage Pipeline Stages, Deal Value, and Expected Close Date match the Leads & Sales CRM help article. Use that article for click-by-click detail; use this page for the grammar.

FAQ

How many columns should an early team use?

Five to seven active columns plus the two final outcome columns covers most early B2B teams. Add a column only when it has unique exit evidence and a different decision owner. That is the practical answer to how many columns you need on an early map.

What is the difference between a column and an outcome status?

A column is the process step. Won or lost is the outcome. The final win and loss columns update outcome status and metrics when you move the opportunity there.

Should product-qualified leads get their own column?

Only if PQL entry has a clear definition and a different next action than sales-qualified opportunities. Otherwise tag PQLs inside Qualified or Evaluation.

Where do I configure columns in Nautis?

From the sidebar open Marketing → Leads & CRM, open a funnel, then use Manage Pipeline Stages (or the manage-stages dialog) to add, rename, reorder, and mark the final win or loss columns.