Where revenue slows down at a B2B marketing services company

What this diagnostic is based on

The company. A B2B marketing services company. It designs and runs campaigns for brands' marketing departments and bills by project. An in-house sales team sells directly to those brands and, now and then, through agencies. Delivery relies on an operational network with low direct cost and on the company's own results-tracking system.

Where the data comes from.

• Previously validated deliverables: client master brief, knowledge base and competitive analysis.
• No CRM connection: the client chose not to export its opportunities. Pipeline and revenue figures are aggregates from its CRM and invoicing software, reported by the client with data as of mid-month 9 of the current year.
• Our own measurement: visibility in AI answer engines (a nine-day window, 808 conversations) and web research.

How to read the labels.

Revenue figures are shown as an index: previous year = 100.

1. The starting point

2. The diagnostic's thesis

Most of the room for revenue growth comes after the campaign has been delivered.

Nobody owns the account once the campaign is over, and the third-party-verified proof isn't used to sell the next one. As a result, a large share of each year's revenue has to come from new business, and the year depends on one or two large accounts turning up.

Confidence in the thesis: medium.

• What supports it: repeat-purchase figures from the CRM, leadership's own description of how campaigns close, and the verified absence of the proof from sales communication on the website and in AI answer engines.
• What limits it: without an export, retention can't be split by segment or by salesperson, and the cause partly rests on an assumption (#20: nobody has an account target after the campaign).

3. The revenue chain today

The return loop, from a delivered campaign to the next sale, has yet to be built.

4. Bottlenecks

#1 · Critical — The campaign ends with the report, not yet with the next sale

Where

Onboarding and expansion: the step from a delivered campaign to the next proposal.

What happens

The relationship closes with the final report. 73% of last year's clients had not bought again by mid-month 9, and 64.8% of clients from the last three years ran a single campaign. The repeat business that does exist is concentrated in a few accounts and comes at the client's initiative, with no process driving it. One large account does repeat (five campaigns in three years), but its spend swings widely: close to 80% of everything it has been billed fell in a single year.

Why it happens

Growing the account isn't yet anyone's target: the sales team is geared to opening new business. There is no results review with the client and no continuity proposal. The verified data that would justify a repeat purchase is handed over as a wrap-up, not as the case for the next one. There is no framework agreement and no service-tier pricing to encourage annual planning.

Impact

More than 60% of revenue has to be won again every year through new business, the most expensive and least predictable part of the chain. ESTIMATE Winning back half of last year's clients who haven't returned, at the average deal size, would add around 25 index points a year, roughly 12% of the target, without a single new client. This is illustrative: the average deal size is skewed by the long tail.

Evidence

DECLARED repeat and retention figures from the CRM · diagnostic interview: the closing report rarely leads to a new opportunity · ASSUMPTION #5, #16 and #20.

Confidence

Medium-high on the figures; medium on the cause.

Impact × effort

High × medium.

#2 · High — The sales engine runs with little visibility

Where

Capture, qualification, opportunity development and close. It runs through the whole chain.

What happens

0% of opportunities have a recorded source or loss reason. Stages reflect the probability each salesperson estimates, with no milestones yet, and revenue forecast records sit alongside real opportunities. Value lost over 24 months is almost 3 times the value won, with not one loss reason recorded, and the average lost deal is 29% larger than the average won deal. Each campaign's operational data stays in the operations system and never reaches the CRM.

Why it happens

Nobody owns the quality of sales data: fields aren't mandatory and stages have no written definition. It isn't a tooling problem, as the company automates other processes well. It's a matter of definitions and ownership.

Impact

Nobody knows who the company loses to, why, or which price works. The fall in win rate (52.8% → 23.1% between year Y-2 and last year) may largely reflect a change in how deals were recorded (ASSUMPTION #6). Nor can anyone tell which channel or salesperson brings in the clients who come back, so growth decisions are made without data.

Evidence

DECLARED the client's written answers on CRM completeness · call with leadership · ASSUMPTION #6 and #12.

Confidence

High: the client acknowledges the gap.

Impact × effort

High × low.

#3 · High — The differentiating proof hasn't reached the buyer yet

Where

Visibility and discovery; proposal and negotiation.

What happens

The company has a differentiator that's rare in its category: delivery of its campaigns is backed by third-party verification. Yet on the website that verification comes sixth out of six selling points, and no case studies are published. In questions about trust and verification, the company appears in 0.8% of AI answers, and outside questions that name it, its visibility is 11.2%, the lowest in its set.

Why it happens

The proof stays inside operations: it isn't packaged, priced or published. There is no marketing function or budget (ASSUMPTION #8). Some public claims about leadership and volume also need reviewing, as the internal systems don't currently back them up.

Impact

In acquisition, a buyer who doesn't know the brand won't find it. In the sale, objections about price and uncertain results have no evidence-based answer. In repeat business (#1), the same verified data that would justify a renewal never becomes a case study. HYPOTHESIS Some of the largest losses are "do nothing" decisions that verified proof could tip. Falsifiable criterion: if salespeople put the last ten losses down to capacity, timing or product rather than trust or expected results, the hypothesis falls.

Evidence

MEASURED visibility in AI answer engines · MEASURED review of the public website · trade press and the verification published by the third party.

Confidence

High on the asset existing and being absent from communication; low on its effect on conversion.

Impact × effort

High × medium.

#4 · Medium — Pricing and packaging, still to be formalised

Where

Proposal and negotiation.

What happens

Price is set proposal by proposal, with no price list or discount policy (ASSUMPTION #17). The year's largest invoice is about 600 times the smallest. A fixed price per campaign, the majority, sits alongside per-unit billing on a minority of invoices. 100% of invoice lines are billed under a single campaign item: tracking and verification are included, with no price of their own.

Why it happens

There's no package that makes the proof count, and the proof is underused because it isn't packaged: each problem feeds the other. A package would open the door to proposing an annual plan.

Impact

Margin depends on each negotiation, price can't be defended with the extra value delivered, and there's no tiered annual plan to make repeat business easier (#1). HYPOTHESIS Unit prices vary widely between clients, judging by how public volume claims square with the per-unit rate. Low confidence.

Evidence

DECLARED invoices from the invoicing software and the client's written answers · pricing HYPOTHESIS with low confidence.

Confidence

Medium on the facts; low on the impact.

Impact × effort

Medium × medium.

#5 · High — Revenue is concentrated in a few accounts and a few people

Where

Close, and the financial year as a whole.

What happens

A single account makes up 24.7% of the current year, with no confirmed repeat yet. The top five add up to 54.6%. 57% of deals won over the last two years are recorded under one person on the team; on another base given by the same source it would be 43%, unreconciled, and the record owner isn't always the person who sold. As of mid-month 9, the current year is 50% ahead of last year at the same point. The large account's invoice explains about three quarters of that gap. Without it, the year would still close 8–22% above last year (index 108–122). The part that could be sustainable improvement is 12% over last year.

Why it happens

It follows directly from #1: without a base that repeats, the difference between a breakout year and a modest one is a large account nobody planned for. And with no shared process or data (#2), knowledge about closing stays with whoever closes.

Impact

When one unplanned large account weighs this much, forecasting the year gets harder: revenue is volatile and the forecast loses reliability. There is also key-person risk at the close.

Evidence

DECLARED account concentration from the invoicing software and record owners from the CRM · ASSUMPTION #15 and #19.

Confidence

High on account concentration; medium-low on concentration in one person.

Impact × effort

High × high.

5. Causal pattern

How the bottlenecks connect.

1. A sales engine with little visibility (#2) hides the real problem: the fall in win rate is read as the first symptom, when the figure that best describes the business is repeat purchase (#1), and that one is in the system.
2. Without repeat business, the year has to be rebuilt from new business. With no marketing and no visible proof (#3), that acquisition is expensive, slow and dependent on hiring salespeople.
3. Price is negotiated case by case (#4) because there's no package that makes the proof count, and the proof isn't used because it isn't packaged.
4. The overall result is dependence on one or two large accounts and on whoever closes them (#5).

Root causes.

Confidence: medium. It rests on the call with leadership, assumptions #16 and #20, and system figures.

6. The lever to pull: repeat business brings the target closer

To reach the target (index 217) from the current year's projected close, new business would need to grow by:

ESTIMATE The gap between the two scenarios is 20 points of retention on the projected close, roughly 27–30 index points of new business a year. Acquisition works: between last year and this year, won deals grow by 48% and opportunities by 29%, at an annualised rate. Neither lever is enough on its own; the one with the most headroom today is repeat business.

Base: projected close of 135–149, with 61.6% of revenue from new clients. Medium confidence.

7. What isn't the problem

In the only comparable years (last year and this year, ASSUMPTION #24), won deals grow faster than opportunities: this year's improvement comes more from converting better than from creating more volume.

8. Prioritisation: what to do first

Confidence in the prioritisation: medium. It's the consultant's judgement on impact and effort, based on declared data.

9. Recommended sequence

HYPOTHESIS Many brands plan the following year's launches in the last quarter, which makes it the moment to propose continuity. It's worth confirming with clients when they sign off the relevant budget line.

10. Implications for Design

The engagement covers Diagnostic only. If the work continues into Design, this is what each piece would solve:

Recommendation: Essentials as a minimum, plus the sales process and the measurement framework. CRM instrumentation doesn't need to wait: the team can set it up now. The process by account quadrant (deal size × repeat rate) is part of turning campaigns into accounts.

11. For the checkpoint: what we can decide now

1. Whether repeat business is the missing lever. Hiring salespeople sustains acquisition; if retention also rises, the growth needed in new business drops from 73–98% to 41–65%.
2. Who owns the account after the campaign. And who manages the highest-value accounts.
3. Whether third-party verification becomes the central promise. Along with the review of public claims that this implies.

Confidence in the diagnostic: 3 out of 5. It would rise with the year's actual close, knowing whether the large account repeats, and a post-mortem of the most recent losses.

12. Gaps, assumptions and hypotheses

Missing information

• MISSING DATA Opportunity export. The client chose not to share it. Without it, we can't quantify the effect of recording practice on win rate, split retention by segment and salesperson, or measure the cycle of lost deals.
• MISSING DATA Loss reasons. They can't be recovered from the system; salespeople's recollection stands in for them (ASSUMPTION #21).
• MISSING DATA Internal sales materials. They exist but haven't been shared, so we don't know whether proposals use the verification and the case studies.
• MISSING DATA When brands sign off the relevant budget line. This sets the timing of renewals.
• MISSING DATA Voice of the client. There's no structured post-campaign feedback.
• MISSING DATA Why the recurring large account keeps coming back: who decides, which budget line pays, and what role the closing report plays. That's what would need replicating with this year's large account.

Assumptions, with their falsifiable criterion

Consultant's hypotheses, with their falsifiable criterion

• HYPOTHESIS The website and contact form bring in little demand. It falls if salespeople attribute most of this year's new clients to inbound or referrals.
• HYPOTHESIS Some high-value losses are "do nothing" decisions that verified proof could tip (#3). It falls if the last ten losses are explained by capacity, timing or product.
• HYPOTHESIS Price weighs on high-value losses (#4). Without loss reasons, this can't be confirmed.
• HYPOTHESIS Brands plan their launches in the last quarter. It falls if clients sign off the budget line at another time of year.
• HYPOTHESIS Opening new markets before fixing repeat business would multiply the same leak.

Contradictions found

• Dependence on individuals reported as none, against 57% of won deals recorded under one person.
• Two different bases for won deals in the same client answer (57% against 43%), unreconciled.
• "We work little with agencies", against an agency being the year's top client.
• "No marketing tools", against a visitor-identification tool listed in the privacy policy; a web form, against zero opportunities with a web source.
• Public claims against system figures.

External dependencies

• If this year's large account repeats, the no-intervention baseline rises and the gap to target narrows, though concentration (#5) gets worse.
• If the year's actual close falls outside the projected range, the roadmap's starting point changes.
• If brands keep shifting activation budget towards retail media, competition for the budget line grows, whatever the company does.
• If a competitor obtains comparable third-party verification, the #3 differentiator becomes track record and verified volume, not exclusivity.