New-business systems inside marketing agencies

Marketing agencies are the cobbler's children of B2B: they sell pipeline generation and rarely run any for themselves. The cause is structural, not hypocritical — client work is billable and urgent, agency marketing is neither, so it loses every scheduling contest. The only fix that survives contact with a busy month is a new-business system that runs whether or not anyone feels like doing it.

Agencies are one sub-vertical among several with distinct pipeline patterns; the full map is in Growth Systems by Industry. This piece is the agency-specific version.

Why do agencies, of all firms, neglect their own pipeline?

Because the feast-and-famine loop is self-reinforcing. When the agency is busy, new business stops, because every senior person is on delivery. When a big client leaves — and agency client concentration makes this a matter of when — the pipeline is empty precisely because the preceding months were good. Then follows a scramble of discounted pitching, which wins poorly-fitted clients, which makes the next delivery period worse. The pattern repeats every 12–18 months and gets explained as "the nature of agency life". It is not. It is what an unowned pipeline does.

There is a quieter cost too. An agency that relies on referrals and reputation cannot choose its clients; it serves whoever arrives. If most of your revenue has always arrived by word of mouth, the maths of referral-only pipelines is worth sitting with — the numbers are less comfortable than the feeling.

What does "a system" mean, concretely?

Something that produces conversations on schedule with no discretionary effort. The mechanism we install looks like this. When the agency defines one target sub-vertical per campaign — not "brands we'd love", but a category it has proof in — then a list is built and verified from sources like Companies House and Sales Navigator. When the list loads, then sequences of 4 emails over 14 days go out at 25–40 emails per day per inbox, every working day, regardless of how busy delivery is. When a positive reply lands, then it is routed the same day to one named owner — not "whoever is free", which in an agency means nobody. When months pass, then scaling happens by adding campaigns for new sub-verticals, never by rotating or blasting the same list harder. A positive-reply rate around 4% is the working expectation; below 3% means the proof or the targeting is wrong.

The point of the machinery is that it removes the scheduling contest. Nobody has to choose between client work and new business, because the system does not ask.

Who owns new business when everyone bills?

This is the real constraint, and job titles do not solve it. A part-time "head of growth" who also runs accounts will always be an account person first. The workable pattern for a 5–50-person agency: automate or buy the top of the funnel (list building, sending, scheduling), and reserve human effort for the two things that cannot be automated — the proof-led copy and the sales conversations. That reduces the founder's role from "doing the outbound" to a few hours a week of replies and calls, which is the difference between a system and a resolution.

Ownership also means measurement. Pipeline created per month, conversations per campaign, proposal win rate — visible weekly, not reconstructed for the away day. Agencies track this rigorously for clients and almost never for themselves.

What about positioning — doesn't outbound need it?

It needs proof more than positioning. The agencies whose outbound converts lead with a specific, evidenced claim in a specific category — which is easier when services are packaged with fixed scope and price. That connects to a larger question of what productising agency services does to the pipeline: productised offers make lists targetable and emails concrete. Vague full-service positioning produces vague campaigns.

How does this compare across sub-verticals?

Recruitment agencies have the two-pipeline version of this problem — client development dies whenever the desks get busy. MSPs face the long-cycle version, where the system's job is staying present until a contract renewal opens a door. The agency version is the starkest, though, because the skill is in-house. The gap is never capability. It is that nobody built the machine that runs when everyone is busy.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.