How to Build an Agency CRM and Sales Pipeline From Scratch

Most CRM setup advice treats agencies like any other sales team. That misses the hard part. An agency CRM has to manage the sale and the handoff into delivery, because the deal is only healthy if the team can service it profitably.

A usable agency CRM tracks lead source, qualification, proposal status, follow-up, contract or payment status, onboarding and account ownership. If it only stores contacts, it will become a shared address book with optimistic revenue numbers attached.

This guide is for building a first working CRM from scratch. It is not an enterprise reimplementation plan, and it does not assume a sales ops team. The aim is a clean system that agency owners, freelancers and small teams can run every week.

Start by mapping how your agency actually sells

Before choosing software, write down how leads arrive and how they turn into paid work. Common agency sources include referrals, inbound forms, outbound campaigns, partnerships, paid ads, events and existing-client referrals.

The upside of this exercise is that your CRM reflects reality from day one. The downside is that it exposes messy sales motions, especially if every founder or account lead sells in a different way.

Next, define the deal types you need to track. Most agencies have a mix of retainers, one-off projects, audits, strategy sessions, upsells, renewals and reactivation deals.

Do not create a separate pipeline for every service line on day one. Start with one pipeline unless the steps are genuinely different, because too many pipelines make reporting harder before they make management better.

A paid ads retainer and a web build may need different delivery plans, but they often share the same sales path: qualify, book discovery, send proposal, negotiate and close. Split pipelines only when stages, owners or close probabilities differ enough to matter.

What fields should an agency CRM include?

Your core CRM objects should be contacts, companies or accounts, deals, activities, notes, proposals, contracts, invoices or payment status and onboarding projects. Some tools call these objects different names, but the job is the same.

The benefit of defining objects early is cleaner reporting and easier handoff. The trade-off is setup time, because someone has to decide where each piece of client information lives.

For agencies, the most useful fields are service line, lead source, monthly retainer value, project value, target start date, decision maker, ICP fit, urgency, next step and owner. Add expected close date only if the team will keep it current.

Keep mandatory fields tight. If every deal needs 20 required fields, salespeople will guess, skip updates or avoid creating deals until late in the process.

Add lost and disqualified reasons at the start. Good options include no budget, poor fit, no authority, wrong timing, chose competitor, unresponsive, too small and service not offered.

This feels administrative, but it pays back quickly. Lost-reason reporting tells you whether the problem is lead quality, qualification, pricing, proposal strength or follow-up speed.

How should you build the first agency sales pipeline?

A simple agency pipeline should track buyer progress, not internal hope. A workable first version is New Lead, Qualified, Discovery Booked, Discovery Completed, Proposal Sent, Negotiation, Verbal Yes, Contract/Payment Sent, Closed Won and Onboarding.

This structure gives you enough detail to see where deals stall. The limitation is that it may be too detailed for a solo freelancer with a five-deal pipeline, so merge stages if updates become a chore.

Define Closed Won tightly. For one agency it might mean a signed agreement; for another it might mean the first invoice is paid or a kickoff is booked.

The key is consistency. If one person marks deals won after a verbal yes and another waits for payment, your revenue forecast will be fiction.

Create separate lost and disqualified states rather than letting bad-fit deals sit in active stages. A bloated pipeline looks encouraging, but it hides the real amount of work that can close.

Pipedrive’s own pipeline guidance includes stages such as prospecting, lead generation, qualification, engagement, nurturing, closing, implementation and onboarding. That is a useful reminder for agencies: the sale should connect to delivery, not stop at a signed proposal.

What are stage-exit criteria, and do you need them?

Yes, you need stage-exit criteria if more than one person touches sales. They define what must be true before a deal can move forward.

The upside is a cleaner pipeline and fewer fake opportunities. The downside is that reps and founders may resist the discipline at first, especially if they are used to moving deals based on gut feel.

For example, a deal should not move from Qualified to Discovery Booked unless budget range, service need, lead source and meeting date are known. If those are missing, the deal is not really booked for a useful discovery call.

A deal should not move to Proposal Sent unless scope, decision maker, pricing model and next follow-up date are entered. Otherwise the proposal is just an estimate with no clear buyer process behind it.

Use exit criteria to protect forecasting. If every stage has a clear rule, pipeline value becomes a management tool rather than a mood board.

Keep the rules short enough to remember. A one-page sales operating guide is more useful than a complex CRM manual that nobody opens.

What should you automate first?

Automate repetitive work that prevents missed revenue. Good first automations include lead assignment, task creation, meeting reminders, proposal follow-up tasks, stale-deal alerts and onboarding handoff tasks.

The benefit is speed and consistency. The catch is that automation magnifies poor setup, so fix stages and fields before building complex nurture flows.

Start with failure points. Create an alert when a proposal has no follow-up date, a discovery call has no next step, or a Closed Won deal has no onboarding task.

Do not automate every possible email sequence in the first week. If your qualification rules are still changing, the automation will keep sending the wrong message to the wrong leads.

For an agency, the most valuable automation often sits between sales and delivery. When a deal closes, create the onboarding project, assign the owner and carry across the scope, start date and client context.

This handoff is where margin is protected. A strong close followed by a vague kickoff creates rework before the client relationship has properly started.

Which dashboards should agency owners set up?

Set up dashboards that show both effort and revenue impact. Core reports include leads by source, discovery calls booked, proposal-to-close rate, average deal size, pipeline value, close rate by service, sales cycle length and lost reasons.

The upside is faster diagnosis. The downside is that dashboards only work if the underlying fields are kept current, so reporting cannot replace pipeline hygiene.

Separate activity metrics from outcome metrics. Calls, emails and meetings show effort, while close rate, revenue, cycle length and retention-related handoff data show commercial result.

Use dashboards to decide what to fix next. Low discovery volume points to lead generation, while high discovery volume and weak close rate points to qualification, offer fit, pricing or proposal quality.

For retainers, track monthly retainer value separately from one-off project value. Mixing them creates misleading average deal sizes and makes capacity planning harder.

If you sell audits or strategy sessions as entry offers, tag them clearly. They may close quickly, but they should not be reported like long-term client retainers.

Which CRM should you start with: HubSpot, Pipedrive or Close?

Choose the CRM around your sales motion, team size and workflow complexity. Headline price matters, but it is only one part of the cost because seats, add-ons, usage and limits can all change the real bill.

HubSpot is the strongest fit if you want CRM plus broader marketing, reporting, service and scaling workflows. It is ranked above Pipedrive and Close in our index, starts free, then from $15, and its free tools include deal tracking, live chat and meeting scheduling.

The limitation is complexity and pricing shape. HubSpot moved new customers to seat-based pricing in March 2024, and paid portals can include Core Seats plus free View-Only Seats. Sales Hub Starter includes up to 2 deal pipelines, Professional up to 15, and Enterprise up to 100.

Pipedrive is the better fit if you want a focused visual sales CRM that people will actually update. It starts from $14, offers a 14-day free trial with no credit card, and its current plan names are Lite, Growth, Premium and Ultimate.

The limitation is that Pipedrive becomes less simple as you add more around it. Add-ons include LeadBooster, Projects, Campaigns, Web Visitors and Smart Docs, and they are charged per company rather than per user. Prices are VAT exclusive.

Close is the right option if outbound is central to how your agency sells. It starts from $19, offers a 14-day free trial with no credit card and no contracts, and puts calling and email activity inside the CRM.

The limitation is usage cost and fit. Close pricing depends on seats, calling and SMS usage, and AI credits; most outbound calls are around $0.02 per minute, and phone numbers start around $1 per month. Solo is limited to 1 user and 10,000 leads.

Close can also involve extra phone and SMS costs. Premium phone numbers are listed at $19 per month per line, Call Assistant at $50 per month per organisation plus $0.02 per minute, and A2P 10DLC SMS compliance fees can apply in the US.

If the agency is building an all-in-one growth engine, start with HubSpot. If it needs sales discipline without a wider suite, start with Pipedrive. If the team lives on outbound calls and rapid follow-up, start with Close.

How do you keep the CRM clean after launch?

Assign one owner for pipeline hygiene and one decision maker for CRM changes. Without ownership, every useful field slowly becomes optional and every optional field slowly becomes clutter.

The benefit of governance is a CRM people can trust. The downside is that someone has to say no to extra fields, custom stages and one-off automation requests.

Run a weekly stale-deal review. Look for deals with no next step, no recent activity, unrealistic close dates or values that no longer match the likely scope.

Every active deal should have four things: an owner, a next step, a next-step date and an expected value. If any are missing, it is not an active opportunity in a useful sense.

Document the sales-to-delivery handoff. Decide who creates the onboarding project, where client context is stored, what scope is passed over and when the account owner takes responsibility.

Review fields and stages after 30 days, then again after 90 days. Remove anything nobody uses, and add only the fields that improve qualification, forecasting or handoff.

What is the best starting path for a small agency?

For a simple sales-first setup, start with one pipeline, a small set of required fields and Pipedrive if you want fast adoption. The risk is outgrowing it if you later need broader marketing, service and reporting in the same system.

For an all-in-one growth setup, start with HubSpot if you want CRM, marketing and reporting to share the same contact record. The risk is paying for more system than you use, especially as seats and pipeline needs increase.

For outbound and call-heavy selling, start with Close if calling, email and rapid follow-up are the centre of the sales motion. The risk is that usage fees and SMS compliance costs matter more than the base subscription price.

Whichever tool you choose, build the process before configuring the software. A CRM cannot fix unclear qualification, vague next steps or a weak sales-to-delivery handoff.

The best first version is boring on purpose: clear stages, few required fields, defined exit rules, basic follow-up automation and a weekly review. That is enough to make the pipeline useful before the agency adds complexity.