How to set up a sales pipeline for an agency: stages from lead to retainer

Why an agency pipeline is different from a product pipeline

Agency sales have three features that a generic sales template ignores.

Deals are projects, not licences. A won deal becomes a scope, a team and a delivery timeline. If the pipeline does not capture scope and start date, the handover to delivery happens by email, and the first week of the project is spent finding out what was promised.

The first sale is rarely the valuable one. A $6,000 website build that turns into a $2,500 a month retainer is worth more than a $15,000 one-off. The pipeline needs a way to separate project revenue from recurring revenue, or the forecast is wrong in both directions.

Most leads come from referrals and inbound. Few agencies run outbound at volume, so the top of the funnel is narrow and each lead matters. Speed to the first call is the stat that moves most, and a pipeline should make it visible who has been waiting.

The five stages, and what to record at each

A stage should mean the same thing to everyone, and moving a deal into it should require specific information. Here is the set we use, with the entry condition for each.

Two things about this table.

First, there is no "contacted" or "qualified" stage. Qualification happens in discovery; a stage for it just lets deals sit. If you want to track leads you have not yet called, use a filter or a label on the Lead stage.

Second, "Won" means signed, not "verbal yes". A verbal yes is Negotiation with an expected close date in the next week. Agencies that count verbal yeses as wins overstate revenue every quarter.

Lost deals leave the pipeline at any stage. Record the reason in one of five or six fixed options (price, timing, chose another agency, went in-house, no response, no budget) rather than free text, so you can count them later.

How to set up the pipeline, step by step

1. Write down the five stages and their entry conditions. Put them in a page everyone can see. The entry conditions are the part people skip, and they are what makes the pipeline comparable between salespeople.
2. Decide the fields. For each deal: company, main contact, source, stage, owner, value, type (project or retainer), expected close date, next action, next action date. Ten fields. Resist adding more until a report needs them.
3. Separate project value from retainer value. Two number fields, not one. A deal can have both: a $6,000 build and a $2,500 a month retainer.
4. Give every deal one owner and one next action with a date. A deal with no next action is a deal nobody is working. Review the list of those weekly.
5. Put a booking link on your discovery stage. The fastest way to lose an inbound lead is six emails to find a time. A public booking page with your availability turns the lead into a discovery call in one click. In Atrix, a booking link from scheduling sits on the discovery stage and the call lands on the deal record and the calendar.
6. Agree a weekly pipeline review. Thirty minutes. Go stage by stage, starting with Negotiation. For each deal: what is the next action, is the date realistic, has anything moved since last week. Move or close anything that has not changed in three weeks.
7. Move won deals into delivery with a handover. Scope, price, start date, named lead. In Atrix this means creating the project from the deal; in a spreadsheet it means a second tab and a Slack message. Either way it should be a step, not an afterthought.

The numbers worth watching

You do not need a dashboard for a pipeline under 50 deals. You need five figures, checked weekly.

• Deals per stage. If Proposal is holding more than a third of your open deals, you are writing proposals for people who were not qualified.
• Win rate from Proposal. Proposals sent in the last 90 days that were won, divided by proposals sent. For a referral-led agency this is often 40 to 60 percent; for inbound it is lower. The absolute number matters less than the trend.
• Days in stage. How long the average deal sits in Discovery and in Proposal. When Proposal stretches past two weeks, the follow-up cadence is the problem.
• Weighted pipeline. Each deal's value multiplied by a stage probability (we use 10, 30, 50, 80 percent for the four open stages). It is rough, but it is a better forecast than the raw total.
• Recurring revenue won this quarter. Retainers signed, in monthly value. This is the number that decides whether the agency grows.

When a spreadsheet stops working

A spreadsheet is the right first pipeline. It costs nothing, everyone can edit it, and the fields are whatever you decide. It stops working at a predictable point: around 20 deals open at the same time, or when a second person starts selling.

The symptoms are the same at every agency we have talked to:

• Nobody trusts the "last updated" column, so the weekly review becomes a status meeting where each deal is re-explained from memory.
• Contacts and companies are typed in by hand, so the same client appears three times with three spellings and the history is split across them.
• Follow-up dates live in the salesperson's head or inbox, not in the sheet, and leads go quiet.
• There is no link between the deal and the discovery call, the proposal document or the signed agreement, so the context is in four tools.

At that point you need three things a spreadsheet cannot do well: records for companies and contacts that deals attach to, a history of what happened on each deal, and a permission model so the right people see values.

Choosing a CRM for an agency

There are two sensible directions. A dedicated sales CRM, which does pipelines deeply and nothing else; or a workspace that includes a CRM alongside projects, docs and scheduling, so the won deal becomes a project without export.

Prices are per vendor pricing page in October 2026. Pipedrive has a fourth plan between Growth and Ultimate; we did not verify its price, so it is not listed. HubSpot's promotional rate is described on its own page as limited-time and for new customers only, so check what applies to you.

Choose Pipedrive if sales is a separate function with two or more full-time people, you run multiple pipelines (new business, upsell, partnerships) and you want reporting on each.

Choose HubSpot if marketing automation matters as much as the pipeline, and you are prepared for the cost of the hubs you add later.

Choose Atrix if the agency is 2 to 15 people, the people who sell also deliver, and you want the deal, the discovery call, the proposal page and the delivery project in one place. The limits are real: one pipeline with five fixed stages, no multi-pipeline reporting, and no marketing automation. The CRM covers companies, contacts, deals and notes; the agencies page shows how the pieces fit together; and the deal pipeline glossary entry defines the terms we use here.

Who should see the money

Deal values are the most sensitive number in an agency. Account managers need to see them. A junior designer brought into a discovery call does not. In a spreadsheet the only control is a second sheet; in most CRMs it is a permission set on a higher plan.

Whatever tool you use, decide the rule before the pipeline fills up: who can see values, who can edit them, and who can mark a deal won. In Atrix this is the crm.value.view permission, granted per member. A member without it opens the same deal, sees the stage, the contacts and the notes, and sees no price.

FAQ

What are the stages of an agency sales pipeline?

Lead, discovery, proposal, negotiation and won. Each stage has an entry condition: a discovery call booked, a proposal sent, a client reply with changes, a signed agreement. Lost deals leave from any stage with a fixed reason code.

How many deals should an agency have in its pipeline?

Enough that the weighted pipeline covers next quarter's revenue target. As a rough rule, if your win rate from Proposal is 40 percent, you need proposals out worth about two and a half times the revenue you want to sign. The spreadsheet-to-CRM threshold is around 20 open deals.

Should a retainer be a separate deal from the project?

Record them on the same deal with two value fields: project value and monthly retainer value. Create a separate deal only when the retainer is sold later as its own decision, with its own proposal and close date.

What is a good win rate for an agency?

It depends on the source. Referral-led agencies often win 40 to 60 percent of proposals; inbound and outbound are lower. Track your own number over 90-day windows and watch the trend rather than comparing with other agencies.

When should an agency move from a spreadsheet to a CRM?

Around 20 live deals, or when a second person starts selling, whichever comes first. The signs are duplicate contacts, follow-ups missed because dates live in inboxes, and a weekly review that re-explains each deal from memory.

Does Atrix replace Pipedrive for an agency?

For a 2 to 15 person agency with one pipeline, yes: companies, contacts, a 5-stage pipeline, booking links and permission-gated values are included in every plan. For a sales team running several pipelines with stage-level reporting, Pipedrive is the deeper tool.

Next step

Write the five stages and their entry conditions on one page today, and put a booking link on the discovery stage. When the spreadsheet reaches 20 open deals, the Atrix CRM is built for exactly this pipeline, and the agencies page shows how a won deal becomes a project in the same workspace.

Sources

Prices checked in October 2026 on the vendor pages: Pipedrive (plan prices also stated on Pipedrive's own CRM guide), HubSpot Sales Hub pricing and the HubSpot Starter page.