How to Build a Sales Pipeline for an Agency (Stage by Stage)

A step-by-step guide to building an agency sales pipeline: the stages to use, entry and exit criteria, follow-up rules and the metrics to track.

TL;DR: An agency pipeline works when every stage has a clear entry condition, every open deal has a dated next step, and you review stalled deals weekly. Use six to nine stages that mirror your real process: lead, contacted, discovery booked, discovery done, proposal sent, negotiation, won, then handoff.

Step 1: write the process before the software

List what actually happens from first contact to signature. Each real milestone becomes a stage.

Step 2: define stages with entry criteria

StageEnter whenRequired next step
New leadRecord is createdMake first contact
ContactedA real conversation was attemptedBook discovery
Discovery bookedCall is on the calendarHold the call
Discovery doneNeeds, budget and decision makers notedSend proposal
Proposal sentProposal deliveredFollow up in 2 to 3 days
NegotiationBuyer is discussing scope or priceAgree terms
WonSignedHand off to delivery

Step 3: set probabilities

Assign a win probability per stage from your own history, so forecasts reflect reality. Start with estimates and adjust after a quarter. Our pipeline calculator and pipeline coverage guide help size it.

Step 4: attach next steps

No deal without a dated next step. This one rule prevents most leakage.

Step 5: review weekly

List open deals with no activity in the last 7 to 14 days. Move them forward or close them as lost.

Step 6: track the few metrics that matter

Conversion between stages, time in stage, source-to-win and average deal value.

Real numbers to sanity-check your pipeline

A worked example: sizing a pipeline

Illustrative scenario. The example below is a composite built to show the mechanics, not a customer case study. The figures are made up for the arithmetic.

An agency wants 4 new retainers a quarter at an average first-year value of $24,000, i.e. $96,000 in new revenue.

Suppose its own history shows these stage-to-stage conversion rates:

StageConversion to next stageDeals needed to start the stage
New lead to Discovery done25%64
Discovery done to Proposal sent70%16
Proposal sent to Won35%11
Won-4

Working backward: to win 4, you need about 11 proposals; to send 11 proposals, about 16 discoveries; to hold 16 discoveries, about 64 leads. Over 13 weeks that is roughly 5 new leads a week. That is the number to manage. Replace these rates with your own after a quarter.

The probability you assign per stage should come from the same history; for example 10% at Discovery done, 35% at Proposal sent, 60% at Negotiation. Use our sales pipeline calculator for your figures.

Building it in Pixelwand

  1. Settings, Statuses, Deal statuses: add each stage, pick a type (Open, Ongoing, On Hold, Won, Lost) and a health score from 0 to 100, and drag to set order.
  2. On the Deal page, use the Kanban board grouped by status; dragging a card updates status, probability and close dates.
  3. Moving a deal to Lost prompts for a reason; review reasons monthly.
  4. Only workspace owners and admins can create statuses; members can use them.

Curious how this looks with your own pipeline?

15-minute walkthrough, no pressure, cancel anytime.

Book a demo
Keep reading

Frequently asked questions

How many stages should an agency sales pipeline have?

Six to nine. Fewer hides where deals stall; more makes people stop updating stages accurately.

What stage do agency deals stall at most?

Commonly between discovery and proposal, and after the proposal is sent. Give both stages dated follow-up tasks and review them weekly.

What should be the entry criteria for each stage?

Write a single observable condition per stage, such as 'discovery call held' for Discovery done or 'proposal document sent' for Proposal sent. That keeps stage data honest.

What metrics should an agency track in its pipeline?

Leads by source, stage-to-stage conversion, time in stage, win rate, average deal value and pipeline coverage against your revenue target.