CRM playbook
CRM for Agencies: Client Pipelines, Retainers, and Reporting Without Chaos
Agencies do not fail at CRM because they lack software. They fail because new business, retainers, and delivery get stuffed into one board that tries to be a project tool, a forecast, and a time sheet. The result is a colorful mess that cannot answer a simple question: what might we close this month, and which clients are at risk?
This playbook treats an agency CRM as a commercial system. Delivery can live next door. The CRM holds relationships, opportunities, retainers, and the story you will need when a client changes marketers for the third time this year.
Two pipelines, not one slush pile
Create a new-business pipeline with stages that match how you sell: intro, chemistry, proposal, verbal, won. Create a separate expansion or retainer pipeline for more work from current clients. Mixing a $12,000 website lead with a $9,000 monthly retainer renewal makes conversion rates meaningless.
Projects themselves can stay in your project tool. Sync the client company and the commercial status. When a project is late, that is a delivery risk that should update a health note on the account, not spawn a fake deal.
Accounts are the center of gravity
Agency knowledge walks out the door when an account director leaves. The company record should hold brand contacts, billing contacts, contracted scope in a sentence, renewal month, and a short history of what blew up last time. Contacts should have roles: economic buyer, day-to-day, legal, finance.
If you only store the Slack handle of the person who likes your work, you will be surprised when procurement appears.
Retainers need dates more than stages
A retainer is not a forever deal sitting in Closed Won. Put renewal date, notice period, and a health flag on the account. Ninety days before renewal, create a task for the owner. If health is yellow, the task is a recovery plan, not a “just send the invoice” reminder.
Lost retainers deserve a reason: budget cut, in-house, competitor, relationship decay. After six losses you will know whether your problem is pricing, staffing, or the first 60 days of onboarding.
New business without theater
Agency pitches die in vague stages named “warm.” Use exit criteria. Chemistry means you have met the people who can say yes. Proposal means a scoped number is in their hands. Verbal is not won. Won is a signed statement of work or a kickoff deposit, depending on how you actually start work.
Source matters. Referral from a current client, inbound from content, and a cold RFP are different machines. Do not average their win rates and then wonder why a quarter feels random.
Reporting clients will actually read
Do not use the CRM as a public reporting portal unless you have a clean, limited view. Most clients want a simple monthly story. Keep that story in the place they already look. Use the CRM to remember what you promised and who approved it so the next report does not contradict the last one.
If you sell retainers on performance, define the metric in a field. Arguments about “what we meant by leads” belong in writing on the account, not in a hallway.
An agency CRM earns its fee when a partner can see new-business risk and renewal risk on Monday without asking three directors for their personal spreadsheets.
30-day agency CRM plan
- Split new business and retainers into two pipelines.
- Put renewal month and notice period on every retainer account.
- Name economic buyer and day-to-day contact on each active client.
- Write five new-business stage definitions and enforce them in one weekly meeting.
- Stop using the sales board as a project tracker.
- Add lost reasons for both pitches and retainers.
- Create a 90-day-to-renewal task workflow.
- Archive zombie leads older than 180 days with no activity.
Frequently asked questions
Should freelancers live in the CRM?
As vendors or a separate object, not as deals. Mixing capacity planning with pipeline confuses both jobs.
Do we need a CRM if we already have a project tool?
If more than one person sells, or retainers can churn, yes. Project tools are weak at forecasts and multi-year relationship history.
How do we handle speculative pitches?
Give them a stage or a low default amount so they do not inflate coverage. Time-box them. Endless free pitch work is a process problem.
Can one owner sit on sales and delivery?
Yes in small shops. Still keep the records separate so you can hire a salesperson later without archaeology.
What about inbound from Instagram or referrals?
Capture source on the person and the deal. Referral source should be a contact you can thank.
Operating notes for agency CRM design
Keep sales work separate from delivery work. The CRM should show prospecting, qualification, proposal, commercial negotiation, and the signed handoff; project tasks belong in the delivery system unless they directly affect renewal or expansion. Mixing both motions on one board makes the pipeline look busy long after the commercial decision is finished.
Model retainers and one-off projects differently. A retainer has start, renewal, notice, and recurring value fields; a project has scope, expected close, and booked value. If both are recorded as generic deals, agency leadership loses the ability to distinguish new logo revenue from recurring base, expansion, and short-term production work.
Capture relationship structure because agency buying is rarely one-contact sales. Record the economic buyer, day-to-day client lead, procurement or legal contact, and internal champion. Make next steps explicit after every pitch. When a champion leaves, the account should still contain enough context for another seller to recover the relationship.
Build reports around capacity as well as bookings. A large projected win can be commercially attractive and operationally impossible in the same month. Connect expected start dates and service type to a lightweight capacity view so the forecast informs staffing. The CRM does not need to become resource planning software, but it should expose demand early enough to act.