Building a Sales Pipeline in a Small Business: Stages That Force a Decision

A useful pipeline is made of stages you can fail. A stage every deal passes automatically is not a stage but a label, and it produces a tidy report while changing nothing about the selling. That distinction decides whether the system helps or merely records.

What makes a stage a real stage

The exit-condition column is what is missing from nearly every hastily built pipeline, and it is precisely why deals sit in one stage for months.

Why not build ten stages?

Because a small team will not maintain ten stages. What happens in practice is that reps use three of them and the other seven become noise that makes the real picture harder to see. Five to seven is the range where you can still look at the board and understand the situation in thirty seconds.

There is one exception: a long sale with stages that depend on an outside party - an approval, a demo, a site visit, a tender. There an extra stage is justified, because it represents a genuine wait rather than an internal nuance.

What to measure, and what not to

• Deals stalled more than X days in a stage - the most useful metric in a small business.
• Stage-to-stage conversion - shows exactly where deals fall out.
• Loss reason - only if picked from a short list rather than typed freely.
• Average time to close - useful after a few months, not at the start.

What is not worth measuring at the start: total pipeline value. In a small business that figure swings on a single large deal, and it creates confidence or panic with no relation to reality.

What do you do with deals that stall?

The simple rule that works: every stage has a time limit, after which the deal must move - forward, back, or to closed-lost. Not "left open just in case".

What happens without it: a pipeline with 60 open deals of which 15 are real, and nobody knows which. That does two kinds of damage - decisions based on an inflated number, and a sense of busyness that stops the rep looking for new leads.

A deal closed as lost is not lost information. You can and should bring it back into the pipeline in six months with a recorded reason, and it is a far cheaper lead source than any advertising.

How to run the first week

1. Write the stages on paper before touching the system.
2. Sit with one rep and place ten open deals into them.
3. If a deal fits no stage, the stages are wrong - not the deal.
4. Set a time limit per stage.
5. Agree just three loss reasons to start with.
6. Meet the team once a week in front of the board, for ten minutes.

Point 6 decides whether the pipeline survives. A pipeline nobody looks at together dries up within a month, so a short standing meeting is worth more than any sophisticated configuration.

What does not belong in the pipeline

• Existing customers in ongoing service - those are not open deals.
• Ideas - "maybe we should approach them" is not a lead.
• Tasks - "send the contract" is a task, not a stage.
• Renewing subscriptions - a separate process with different logic.

Mixing those four into the pipeline is the most common reason it stops reflecting reality. Separating them connects directly to the wider decision described in ERP, CRM or accounting software.

What a ten-minute pipeline meeting looks like

This meeting is the mechanism that holds everything else together, and it stays short only if it is structured. The order that works: open the board and look at three groups only. Deals past the stage time limit - decide now where each one moves. Deals that arrived this week - has anyone made contact. Deals that closed - won or lost, and the reason.

What you do not do in this meeting: discuss the content of a deal. The moment it becomes "what he said and what I answered" is the moment it stretches to an hour and stops happening the following week. A content discussion is a separate conversation with whoever is relevant.

After a month of these meetings a pattern appears: which stage collects most of the stalled deals. That is the most valuable thing a pipeline produces, and it is usually a surprise - the bottleneck is rarely price, but the stage before it, where too little was learned to build a good proposal.

What to do with returning leads

A customer who enquired a year ago and did not buy is not a cold lead. They are a lead with history, and a far better starting point than a purchased list. The simple way to use that is to record, on closing as lost, when it is worth returning - in three months, in a year, or never - and to set a reminder.

In businesses that do this, a meaningful share of closes comes from that group at effectively zero acquisition cost. In businesses that do not, those leads are simply forgotten - not because nobody wanted them, but because no mechanism reminds anyone.

It is also worth recording the loss reason honestly. "Price" is what gets recorded when nobody asked; the real reasons are usually more specific - timing, an existing supplier, doubt about fit - and each one leads to a completely different follow-up.

Sources

• Fireberry by Powerlink - lead and sales management
• Kala CRM - sales process and quotation management