A US B2B company in workplace safety was established in its market and marketing was running. What was missing was not activity but structure: no segmentation, no lead scoring, no content or demand-generation strategy holding it together. The diagnosis was that the growth constraint sat in an asset the company already owned.
The company held more than 13,000 contacts in HubSpot, built up through its own activity over time. On paper that is a substantial asset. In practice the database was not segmented in a way that allowed anyone to act on it, and there was no structured scoring model to distinguish a compliance manager at a manufacturing site from an individual buying hearing protection for personal use.
Marketing was running. Content was being produced, paid acquisition was live, outbound was being set up. Each part worked on its own terms, and none of them fed the others. The commercial question was therefore not how to generate more demand, but how to make the demand and the data that already existed commercially useful.
The clearest symptom was lead quality. A product category that serves both workplaces and individuals attracts both audiences through the same search terms and the same landing pages. Personal-use enquiries entered a B2B pipeline, consumed sales attention and distorted the signals that paid acquisition was learning from.
Filtering that at the form was never going to be enough. The fix had to operate at three points at once. In acquisition, through compliance-specific keywords, negative keywords, qualifying ad copy and landing pages written for a workplace buyer rather than an individual one. In the CRM, through scoring that recognises the difference. And in the feedback loop, by telling the ad platform which enquiries became real business instead of letting it optimise towards form fills.
I designed the 2025 demand architecture as a closed system: attract the right audience, capture and qualify demand, score and route it correctly, nurture what is not sales-ready, and feed the commercial outcome back into acquisition.
Four pillars built on the buyer's actual problems: compliance pressure, workplace noise hazards, employee engagement with hearing protection, and the cost and risk of hearing loss. Formats ranged from articles and whitepapers to webinars, tools, video and email nurture.
A practical way to divide a broad safety and compliance buying group by industry relevance and responsibility, rather than by a single job title that does not exist consistently.
A model combining firmographic fit, engagement behaviour and intent signals, with negative scoring for personal-use indicators so poor-fit demand is filtered rather than routed.
How CRM data, enrichment, email outreach and LinkedIn would fit into the wider system, defined as requirements rather than as a standalone volume channel, with the existing contact base as the starting point.
Analytics, search and reporting connected so that qualified commercial outcomes, not form submissions, become the signal acquisition optimises against.
Re-engagement logic for the existing contact base, so the first source of pipeline is the asset the company already owns, supported by website, partnerships and organic visibility on the attract side.
Industry benchmarks, not client figures. Decay rate: MarketingSherpa, used as the standard HubSpot benchmark. Revenue loss and lost opportunities: Validity, State of CRM Data Management 2025, based on 602 respondents. Gartner separately estimates the annual cost of poor data quality at around USD 12.9 million per organisation.
These figures are not this client's results. They describe why the diagnosis mattered commercially. A contact base that sits unsegmented is not stable inventory waiting to be used later: roughly a fifth of it becomes unreliable each year while the decision to do something with it is postponed.
That reframes the priority. Buying more reach adds to a pool that is already leaking. Segmenting, scoring and re-engaging what exists addresses the part of the problem that is losing value fastest, and it costs a fraction of an equivalent acquisition budget.
Deciding what counts as a good lead forces a company to state who it actually sells to, which industries it can serve credibly and which enquiries it is willing to leave alone. That conversation is uncomfortable and it is usually postponed, because adding a channel feels like progress while defining a qualification threshold feels like a constraint.
Without it, every downstream system inherits the ambiguity. Sales works an undifferentiated list, nurture speaks to everyone in the same voice, and the ad platform optimises towards whichever audience submits the most forms. The scoring model was the point where the rest of the architecture became possible.
Most outbound programmes are designed as a separate machine: buy a list, write a sequence, send volume. Here the starting point was different, because the company already held the audience it needed. The design used the CRM as the source of truth, enrichment to complete and verify records, and segmentation to decide who was worth contacting in which order.
Email and LinkedIn were specified as two layers of the same motion rather than two campaigns, and the messaging was tied back to the four content pillars so that outreach could reference something useful instead of opening with a pitch. Sender infrastructure was treated as part of the design: separate domains and mailboxes, warm-up before volume, and a planned capacity corridor of up to several hundred sends per day once the infrastructure was established. That corridor was a design target for the execution team, not a figure that was reached during the mandate.
The stack was defined as a set of roles in the system and evaluated against cost, overlap and the realistic capacity of the team. Where two tools did similar work, one was removed: Reply.io was assessed as a combined email and LinkedIn option and deprioritised on cost and fit, in favour of a dedicated sending layer with a separate LinkedIn motion.
Where this case ends is as relevant as what it contains. The architecture, the workflows and the scoring logic were designed and documented during the mandate. Implementation of the integration layer continued after handover, and hands-on operation of the enrichment and LinkedIn tooling was not part of this engagement.
The engagement produced a GTM and demand-generation roadmap, a segmentation and scoring model, an outbound architecture defined as requirements, and a measurement design that closes the loop between commercial outcomes and acquisition. When the initial technical implementation stalled, I stepped into the first configuration and sender warm-up far enough to unblock it, then handed execution back to the delivery team.
The mandate was approximately two months and ended at handover. Ongoing campaign execution and performance optimisation continued with others, so no post-handover results are attributed here. What the case demonstrates is the part I was engaged for: turning an existing data asset and a set of disconnected activities into one system somebody else can run.
The strategy, segmentation, scoring logic, outbound architecture and measurement design were developed during the mandate. Implementation of parts of the stack continued after handover and is not claimed as delivered work.
No conversion, pipeline or revenue figures are presented, because none can be reliably attributed to this engagement.
The instinct when pipeline is thin is to buy more reach. In this case the more valuable move was to make the existing asset legible: segment it, score it, and connect it to content and outbound that speak to a defined buyer. Volume added to an unstructured system produces more of the same noise.
The engagement also ran entirely remotely from Europe into a US market with its own regulatory vocabulary. That is a constraint worth naming: a demand strategy for an OSHA-driven buyer cannot be transposed from European B2B practice, because the compliance pressure, the job titles and the language buyers search with are specific to that market. The first work was learning it well enough to build content pillars a US safety manager would recognise as written for them.
This is also why the engagement was deliberately short. Designing the operating model is a different job from running it, and the handover is the point at which the design has to survive contact with the team that owns it.
How I work →If you are generating leads but cannot tell which of them are worth working, the Marketing System Audit shows where qualification, data and acquisition are pulling against each other. Diagnose the constraint before you scale it.
