01 / Start close
Why should the founder sell the first customers?
Early sales is not only a revenue task. Each conversation tests whether the product, language, price and buying process make sense to a real person. When a founder delegates every first conversation, the business may gain activity while losing the evidence that should shape the offer.
Y Combinator's lessons on first customers encourage founders to do the initial work themselves. Use that principle with care: the founder does not need to perform every administrative step forever. They do need direct exposure to the buyer's objections and the moments when a supposed problem turns out not to be urgent.
After each call, write what the buyer actually said. Separate the pain they described from the solution you suggested. A later team member should be able to read the record and understand why the opportunity exists, not merely that the founder had a promising chat.
02 / The record
What should every opportunity record contain?
A useful record is short enough to update immediately after a conversation. It needs a named company and contact, the buyer's problem in their words, current workaround, consequence of inaction, decision participants, proposed value, open objections and one dated next action. If there is no next action, the opportunity is either waiting on a clear event or it is not active.
Do not turn the record into a biography or a place to store every possible field. More fields can create the appearance of discipline while making the system harder to maintain. Begin with the information that changes a decision. Add a field only when someone can name the action it will support.
Salesforce's pipeline guidance describes stages and exit criteria for moving an opportunity forward. The practical lesson for a small team is that a stage should mean something observable. 'Interested' is a mood; 'buyer agreed to a paid pilot with a date and decision owner' is a testable state.
03 / The pipeline
Which stages are enough at the beginning?
Five stages are usually enough for an early operating review: identified, discovery complete, fit confirmed, offer under decision and committed or closed. A separate status should record why an opportunity is paused or lost. Define the evidence for each transition so that moving a card is a decision, not a gesture.
The example below is an operating design, not a Salesforce requirement. Adapt it to the buying journey. A household consumer purchase, an enterprise contract and a manufacturing partnership will not share the same steps.
| Stage | Evidence to enter | Next question |
|---|---|---|
| Identified | Named buyer and plausible problem | Can we reach the actual user or decision maker? |
| Discovery complete | Buyer described current process and cost | Is the problem frequent and funded? |
| Fit confirmed | Buyer accepts a test or specific proposal | What must be true to buy? |
| Offer under decision | Price, scope, owner and date documented | Who can approve, and what remains open? |
| Committed or closed | Payment, signed agreement or explicit no | What did the outcome teach us? |
04 / Follow-through
How should a follow-up work without sounding automated?
Send a short recap after a substantive conversation. State the problem as you understood it, the agreed next step, who will do it and when. If the buyer corrected your understanding, keep the correction. The purpose is to prevent the relationship from becoming dependent on one person's memory.
A reminder system can be simple: every open opportunity has a next-action date, a named owner and a reason for the action. Review items due today before starting new outreach. When a buyer is not ready, agree on a real trigger for reconnecting rather than sending generic check-ins. If no trigger exists, mark the opportunity dormant.
Salesforce's Trailhead module explicitly recommends summarising a sales call and reiterating next steps. That supports the recap practice. The exact number of follow-ups should come from the buyer's context, not an arbitrary universal sequence.
05 / Weekly review
What should the founder examine every week?
Review the pipeline in this order: overdue promises, opportunities with no next action, decisions waiting on the buyer, deals stuck in a stage, and newly lost conversations. Ask what evidence changed rather than who sent the most messages. The best review may reveal that the offer is wrong or that the wrong buyer is being approached.
Track a small set of measures: qualified conversations, proposals based on a stated problem, time between meeting and promised next step, conversion to paid commitment, and reasons for loss. Counts alone can reward activity. Read a handful of actual customer records alongside the numbers so the team can see what the metric hides.
Amazon's public principles of customer obsession and diving deep are useful reminders that numbers and anecdotes must inform each other. This is an analogy, not a claim that Amazon uses this exact founder sales process. The operating system should make customer reality more visible to a growing team.
- Do the oldest promised action before adding new leads.
- Ask which stage has the longest wait and why.
- Review lost deals for repeated objections, not just a lost count.
- Write one change to the offer, process or qualification rule each week.