In-House vs Agency: The Real Question Is Who Owns the Funnel

In-house vs agency is a staffing answer to an ownership question.

A £5M services firm asks whether to hire a marketing manager or sign an agency, and whichever it picks, the same thing happens twelve months later. The hire is busy and the pipeline is flat. Or the agency is busy and the pipeline is flat. Both were given tasks. Neither was given the number, and the number is the whole point of owning the funnel for named accounts.

The question that decides the outcome is who owns the funnel: the one named person who, when the board asks why qualified pipeline fell last quarter, answers without pointing at someone else.

Settle that, and in-house vs agency becomes a question about which jobs sit where. This article splits the funnel into seven jobs, says who can own each, and sets out what to put in writing before anyone is hired or signed.

The seven jobs of the funnel and who can own each

A funnel for a complex B2B service has seven jobs. Every one of them exists whether or not someone is doing it.

Three of the seven cannot leave the building: strategy, the account list, and the sales handoff. They are decisions about what the firm sells, to whom, and how sales and marketing treat each other. An agency can draft them and should push on them. It cannot own them, because it is not in the room when the firm changes its mind.

Two of the seven almost always go outside at a £5M to £20M firm: capture and the build of the measurement plumbing. They need craft hours, in paid search, site work, tracking, CRM configuration, that a one-person marketing team does not have and should not be hired for.

The remaining two, demand creation and the reading of the measurement, are the real decision. They can sit either side. Where they sit decides whether the in-house person is a marketing manager or a coordinator of suppliers.

Seven jobs, three that stay, two that go, two that decide the shape.

Why "who owns the number" decides the rest

Pick the owner of the pipeline number first and the table fills itself in.

If the owner is in-house, the hire needs job 7 and job 5 as their own: they present the board page and they can read the CRM without asking anyone. Everything outside reports to them against the same number. The agency is a supplier of jobs 3 and 4 with a stage target, which is a very different brief from "run our LinkedIn and Google".

If the owner is outside, the firm is buying full-funnel ownership, and the agreement has to say so: the agency owns demand creation, capture and measurement, reports pipeline by stage to a named director monthly, and holds the account list and handoff with sales as a shared document. The in-house role shrinks to jobs 1, 2 and 6 plus the relationship. That is a founder or a commercial director for a few hours a week, not a marketing hire.

What does not work is the middle. A marketing manager who owns "marketing" and an agency that owns "the campaigns" with the number sitting nowhere. The firm pays for both and gets activity reports from each. The Edelman and LinkedIn 2024 B2B Thought Leadership Impact Study of 3,500 decision-makers found only 29% of the organisations producing thought leadership could link any lead or win back to it. That is the signature of unowned measurement: a lot of output, nobody accountable for whether it did anything.

What to put in writing, whichever way you go:

1. The name of the person who owns qualified pipeline, and the stage definition they are judged on.
2. Which of the seven jobs each party owns, builds or drafts, using the three verbs and no others.
3. The reporting page: one page, monthly, pipeline by stage against spend, presented by the owner. The measurement contract behind it is job four of the ABM operating layer, and the owner checks the figures on it with the agency ROAS reporting audit.
4. The exit: what the firm keeps if the arrangement ends (accounts, data, access, the measurement build) and the notice period.
5. The review date at which the split is revisited, usually after two full sales cycles.

Our guide to judging a marketing engagement on a long cycle sets out what the monthly page should show at month one, three and six; the sales and marketing alignment article covers the handoff definitions that job 6 depends on.

One name on the number, in the contract, before the first invoice.

What the in-house hire is for, and what it is not

The in-house marketing hire at a £5M to £20M services firm is for the jobs that need the building: knowing the clients, sitting with sales, hearing what the founder learned on last week's call, and turning that into positioning and content nobody outside could write. If the firm wants founder-led demand creation, which for complex services is usually the strongest lever, the in-house person is the one who makes it happen every week.

The hire is not for running paid search, building tracking, configuring the CRM or designing landing pages. One person cannot be a strategist, a copywriter, a paid media buyer and an analyst, and the job advert that asks for all four produces a generalist who does each at a level the firm would not accept from a supplier.

The arithmetic, illustratively, for a firm choosing a single mid-level hire at £55,000 salary with a loaded cost of around £70,000 a year, plus a £20,000 tools and media budget they have to spend alone:

Hire: £70,000 loaded cost + £20,000 budget = £90,000 a year Hours on capture craft (paid, tracking, site) at perhaps 40% of time: 0.4 × 1,650 hours = 660 hours of work done at generalist quality

Compare the same £90,000 split as a part-time commercial owner inside (job 1, 2, 6, 7, say £30,000 of a director's time) plus £60,000 outside for jobs 3, 4 and 5 with a pipeline target. The firm now has a name on the number inside and craft outside, for the same money. The honest limit is that this only works if the director actually gives the hours; if nobody inside owns it, the outside spend becomes the unowned middle again.

Hire for what needs the building; buy what needs the craft.

If engagement is climbing and the CRM is not moving, the operating layer is the first place to look, so book a call and we will trace it together.

What an agency is for, and what it is not

An agency is for the jobs that need craft hours and a view across many accounts: capture, the measurement build, and increasingly demand creation where paid channels carry it. It is also for the one thing a hire cannot give you, which is the option to stop. A retainer ends in a notice period; a hire ends in a redundancy process.

An agency is not for owning what the firm sells or who it sells to. Any agency that will take your positioning off your hands is taking your margin with it, because positioning is the decision that sets your price. And an agency is not a substitute for the inside owner. If nobody at the firm can read the pipeline report and challenge it, the report will be written to be read by nobody.

Gartner's 2025 CMO Spend Survey, 402 marketing leaders mostly at companies over $1bn revenue in North America, the UK and Europe, found budgets flat at 7.7% of company revenue, 59% of CMOs saying the budget is insufficient, 39% planning to cut agency spend and 22% saying generative AI had reduced their reliance on agencies for creativity and strategy. Those are very large companies and the percentages do not transfer to a £5M firm. The direction does: the work agencies are losing is the drafting and the decks. The work they keep is the craft and the accountability for a number.

The commercial version of this question, which agency, at what fee, on what terms, belongs on the service pages. Actualyse takes funnel ownership in two shapes: the growth marketing programme, where capture, measurement and reporting run as one engagement with a pipeline target, and the account-based marketing programme, where the account list and the committee work are added. The case studies show the firms each has suited. Neither shape works without a named owner inside, and we say so before signing.

Buy craft and accountability; never outsource the decision about what you sell.

What does not settle the question

Four comparisons get made in this decision and none of them decides it.

Cost per hour. An agency's blended rate against a salary divided by 1,650 hours tells you what an hour costs and nothing about what an hour produces. A paid-media buyer's hour and a generalist's hour on the same task are different products.

Awards and logos. They tell you the agency can win awards and has had clients. Neither predicts whether it will take ownership of your number or produce reports about activity.

A shared Slack channel. Responsiveness feels like ownership and is not. A supplier who replies in four minutes and owns nothing is the most expensive supplier you will have.

The number of people on the account. Six names on an org chart does not mean six people doing the work, and three excellent people with the number in their contract beat six without it.

A client-side executive at a large software firm was asked in a trade podcast whether agencies are still worth paying for. It is the right question asked of the wrong party. An agency is worth paying for when a job needs craft the firm lacks and someone inside owns what the craft is for. It is not worth paying for when it is bought to avoid deciding who owns the funnel.

The falsifiable claim: a firm that names one owner of the pipeline number, inside or outside, and writes the seven-job split into its agreements will produce a monthly pipeline-by-stage page within one quarter. If it is still presenting channel activity at month four, the ownership was not real, whatever the contract said.

A fast reply is not an owner; a name on the number is.

FAQ

Should a £5M B2B services firm hire a marketing manager or an agency first?

Name the owner of the pipeline number first. If that owner is a director inside giving real hours, buy capture and measurement outside and skip the generalist hire. If nobody inside can give the hours, a senior hire who owns jobs 5 and 7 comes before any agency.

Which jobs should never be outsourced?

Strategy and positioning, the account list and ICP, and the sales handoff. An agency can draft all three and should challenge them; the firm owns the decision.

How long before we review the split?

Two full sales cycles. On a nine-month cycle, that is eighteen months, with a monthly pipeline page in between so the review is not the first time anyone looks.

What does "own the number" mean in a contract?

A named person, a stage definition for qualified pipeline, a monthly one-page report by stage against spend, and an exit clause that says what the firm keeps. Without all four it is a sentiment.

Is a hybrid model just the unowned middle?

Only when the number has no owner. A hybrid where an inside director owns reporting and the handoff while an agency owns capture and measurement against a stage target is the shape that works most often for firms at this size.

Summary

• Decide who owns the qualified pipeline number before deciding whether to hire or sign.
• Keep strategy, the account list and the sales handoff inside; buy capture and the measurement build outside unless you have the craft in-house.
• Write five things down: the owner's name, the seven-job split using own, build or draft, the monthly one-page report, the exit terms and the review date.
• Do not hire one generalist to do four crafts; do not sign an agency to avoid naming an owner.
• Review the split after two full sales cycles, and treat a month-four report that still shows activity instead of pipeline as the trigger to change it.

Actualyse runs account-based and full-funnel programmes where one owner connects spend to pipeline. Book a call to talk through where yours stands.