Eleven of a UK consultancy's last 14 closed deals came from accounts the partners had picked by name. The firm, 12 people, had spent three months emailing 400 broad contacts a week at a 1.8% reply rate and seven booked calls a month, and that list had produced just three won deals. So we paused it and built an account based marketing strategy around 15 target accounts, with account-specific copy for each and a LinkedIn connection step before the email sequence. Three months later the reply rate was 14.3%, from a list one-twentieth the size.
ABM vs broad outbound for a 10-person UK team
The honest trade-off is time-per-contact versus quality-per-contact. Broad outbound costs roughly five minutes per contact once you have a running list and template. ABM costs 45 to 90 minutes per target account for the initial research and copy layer, dropping to 20 to 30 minutes for subsequent cycles once the account profile exists in your CRM.
What matters is which approach generates the deals you actually want to work on.
| Broad Outbound | ABM (15 Accounts) | |
|---|---|---|
| Contacts per month | ~1,600 | 15–30 contacts |
| Time per contact | 5 min | 45–90 min (first cycle) |
| Typical reply rate | 1–3% | 8–18% |
| Pipeline quality | Mixed, off-ICP common | High, pre-qualified |
| Time to first booking | 2–4 weeks | 5–8 weeks |
| Deal value | Variable | Typically higher |
| Right for | Testing ICP, filling pipeline fast | Targeting specific named accounts |
Once we isolated won deal data by source, broad outbound had generated deals, just not the engagements the partners wanted. ABM replaces only the portion of outbound that fills the calendar with the wrong conversations.
For a 10-person firm doing account based marketing for B2B sales, the practical limit is 20 to 25 active target accounts. Beyond that, per-account research quality degrades and the copy starts reading as generic as the broad list it replaced.
Six signals that make an account worth targeting
If you are wondering how to start account based marketing, start with the list. Use these six signals to score and shortlist target accounts. An account needs at least four to justify the per-account investment.
- Revenue range match. The account sits within your ICP revenue band. Use Companies House filings, not self-reported data from LinkedIn or press releases.
- Active hiring in the relevant department. Look for live postings in the department you sell to. Job postings are a buying signal two to three months ahead of a project starting.
- Technology fit. The account uses tools you integrate with, or has a tech stack that creates the problem you solve. Check LinkedIn's company overview section or use BuiltWith for web-visible stack data.
- Growth indicator. Headcount growth of 15% or more in the last 12 months via LinkedIn, or a recent funding event visible on Companies House or Crunchbase.
- Trigger event. A new executive hire, a recent acquisition, a published report, or a live tender on Contracts Finder.
- Network adjacency. A first-degree connection at the firm, or a shared reference client in the same sector.
Score each account from zero to six. Four or above enters the active sequence. Two or three goes to a 90-day watch list. This scoring step is what separates an ABM programme from a slightly shorter cold list.
Account intelligence gathering without a research team
The research task splits into three layers: public signals, trigger monitoring, and contact-level intelligence.
Public signals take 15 to 20 minutes per account: Companies House for financial data and director changes, the LinkedIn company page for headcount trends and department structure, and six months of Google News for press coverage and contract wins.
Trigger monitoring can be largely automated. Set a Google Alert for the company name plus relevant keywords: "management consultancy X procurement", "X new operations director". Clay's enrichment workflows can pull Companies House change events, new LinkedIn job postings, and G2 review activity automatically, updating your CRM when an account crosses a threshold. The B2B intent data signals post covers the full enrichment stack in more detail.
Contact-level intelligence means knowing who to reach before writing the first line. For a firm targeting operations leads, the first contact is typically the COO or Head of Operations. LinkedIn Sales Navigator's decision-maker filter gets you there in under five minutes. Their last three LinkedIn posts and any published articles are the raw material for the personalised opening line. The LinkedIn lead generation systems guide covers Sales Navigator filtering and connection workflows in more detail.
Under UK PECR and GDPR, B2B outbound to corporate email addresses is permitted under legitimate interests, but every message needs a clear opt-out. The ICO's direct marketing guidance covers the electronic mail rules. Read it before sending the first account sequence.
The content layer: landing pages, case study references and email copy written for one firm
In practice, personalised means a content template with four to six account-specific variables, plus one genuinely bespoke observation that could only apply to that firm, rather than a different email for every account.
The content layer had three pieces:
Account-specific email copy with four personalisation slots: company name, a reference to something the decision-maker had recently published or said publicly, a specific operational challenge visible in their job postings, and a case study reference from a client of similar size and sector with a named result (not "a firm like yours").
A single landing page per account, not a custom website, but a standard case study page with the account name in the headline and a one-paragraph "here is what we would build for you" section. Hosted on a path like /approach/[account-slug]. HubSpot CMS makes this straightforward with smart content tokens.
A LinkedIn message drafted for the connection request and first follow-up.
Here is a sample Clay enrichment table configuration that feeds these slots automatically:
{
"table": "target_accounts",
"enrichment_columns": [
{
"name": "companies_house_sic",
"source": "companies_house",
"key": "sic_code"
},
{
"name": "recent_job_postings",
"source": "linkedin_jobs",
"filter": "posted_within_days: 30"
},
{
"name": "decision_maker_last_post",
"source": "linkedin_posts",
"contact_field": "coo_linkedin_url",
"max_results": 3
},
{
"name": "tech_stack_crm",
"source": "builtwith",
"fields": ["crm", "erp", "marketing_automation"]
}
],
"output_to_hubspot": {
"object": "company",
"property_prefix": "abm_",
"update_existing": true
}
}
The abm_ prefix keeps ABM enrichment separate from standard CRM fields. The LinkedIn AI SDR build shows the full enrichment-to-sequence handoff, including how enriched CRM fields map to personalisation tokens. We build this kind of setup for clients under sales and outbound automation.
ABM outreach sequence design across LinkedIn, email and voice over 6–8 weeks
The sequence uses three channels in order: LinkedIn first, email second, voice third. Sequence order matters: a cold email referencing a LinkedIn connection that does not yet exist is inconsistent. A LinkedIn message following an email lands with more context than one sent cold.
Week 1 2 3 4 5 6 7 8
| | | | | | |
LI-1 EM-1 EM-2 LI-2 EM-3 VM-1 EM-4
(conn) (intro)(case)(post)(fup)(vml)(final)
- LI-1 (Week 1): LinkedIn connection request with a brief, genuine note referencing their work. No pitch.
- EM-1 (Week 3): First email once connection is accepted. References the decision-maker's post, a recent hire, or a live tender.
- EM-2 (Week 4): Case study email. One named client, same sector, specific result. Not "we helped a firm like yours": a real example with a real number.
- LI-2 (Week 5): Comment on a post the decision-maker published. Genuine engagement, no selling.
- EM-3 (Week 6): Short follow-up. "Did the case study resonate?" Two sentences maximum.
- VM-1 (Week 7): Voicemail. Ten to fifteen seconds. Name, company, one sentence on why you reached out specifically.
- EM-4 (Week 8): Final email with a specific proposed time for a 20-minute call. No reply: move the account to a 90-day nurture list.
For the email execution layer (deliverability, sending infrastructure, and warm-up), our multi-channel outbound sequence post covers the technical setup in detail.
CRM setup for ABM tracking at account level
Standard CRM setups track at contact level. ABM requires account-level tracking. If HubSpot or Salesforce is not aggregating contact activity to the company record, you can only tell whether a single contact opened an email, not whether the account is engaged.
The minimum CRM configuration for ABM:
| Object | Required Properties | Notes |
|---|---|---|
| Company | abm_tier, abm_status, abm_owner, abm_sequence_start |
Tier 1 = active, Tier 2 = watch, Tier 3 = disqualified |
| Contact | abm_sequence_step, abm_last_touch, abm_reply_status |
Roll up to company score on every update |
| Deal | deal_source_type, target_account_name |
Required for pipeline-per-account reporting |
Configure a HubSpot company score that increments on: email open (+1), email reply (+5), LinkedIn connection accepted (+3), website visit (+2), account landing page visit (+5), booked meeting (+20). When a company score crosses 15, move the account status to "engaged" and send a Slack notification to the assigned AE.
More detail is in the CRM enrichment and ICP scoring post; the ABM version adds the account-level aggregation layer on top.
Why pipeline per account beats reply rate as the main metric
Reply rate is a useful diagnostic but a poor success metric for ABM. A 14% reply rate from 15 accounts is two replies. That number tells you the copy is working. It does not tell you whether the programme is generating the commercial outcomes that justify the per-account investment.
Track three metrics instead:
- Pipeline per account. Total deal value from each target account in the current quarter. Target: 3 to 5 times the cost of one sequence cycle.
- Sequence-to-meeting rate. Percentage of accounts converting to at least one booked meeting. At 15 accounts, aim for four to six per cycle, roughly 30%.
- Meeting-to-deal rate. Percentage of ABM meetings generating an active deal. Higher than broad outbound because accounts were pre-qualified before the first touchpoint.
For the consultancy: pipeline per account in the first ABM quarter was £47,000. Broad outbound pipeline per contact had been £190. The difference comes from how precisely the accounts were selected.
Forrester's B2B buyer research offers a counterpoint worth noting: ABM's advantage depends on alignment between whoever selects accounts and whoever runs sequences. Where that breaks (marketing picks accounts, a junior SDR runs the sequence without context), reply rates can climb while pipeline-per-account stays flat.
What changed in 2025–2026: AI-generated account personalisation in Clay and HubSpot
Clay's AI column feature, which shipped in late 2024 and was updated significantly through 2025, changed the economics of account personalisation. You can now generate account-specific first lines, synthesise job posting data into a one-sentence challenge statement, and pull recent news into a "why now" hook, all inside the enrichment table. By Q1 2026, teams were generating account-specific opening paragraphs at 200 accounts per hour.
HubSpot's dynamic personalisation tokens in Marketing Hub Professional (updated in late 2025) let you inject enriched CRM property data into email body text at send time, using the same abm_ prefixed fields populated by Clay. The combination removes the largest time cost in the original ABM model: writing the personalised first line for each account by hand.
The programme structure does not change. The content creation bottleneck (previously 45 to 90 minutes of manual research per account) largely disappears. A human reviews each output before sending, taking 10 to 15 minutes. The research work itself becomes the bottleneck, not the writing. The AI cold email personalisation post covers the deliverability and output-review workflow for AI-generated account sequences.
Three account based marketing examples and their pipeline per account
Three real programme designs from clients, with their pipeline-per-account results over one quarter.
Good: The 15-account focused programme
One founder, 15 named accounts, HubSpot configured for account-level engagement scoring, sequences running in Instantly. Each account had a bespoke opening paragraph from Clay, reviewed by the founder before sending. Sequence-to-meeting rate: 33% (5 of 15 booked a meeting). Pipeline per account: £51,000. Total quarterly pipeline: £255,000 from a programme costing roughly £2,800 in tools and eight hours per week.
Bad: The 75-account "ABM" programme
Same structure, account list expanded to 75 to "generate more pipeline." Research quality fell: six personalised first lines contained generic or incorrect observations. Reply rate: 4.2%. Two meetings booked. Pipeline per account: £8,400. The team had mistaken a longer account list for a broader outbound list and ended up with the worst of both.
Ugly: ABM without account-level CRM alignment
A team that ran well-built sequences but tracked everything at the contact level, not the company level. A CFO at a target account replied positively. The assigned AE followed up and booked a discovery call. Two days later, a second SDR sent a cold email to the CEO at the same account, because there was no account-level status flag to prevent it. The CFO saw the CEO's forwarded email, concluded the firm did not know what it was doing, and cancelled the meeting. One missing CRM property cost £90,000 in pipeline. Configure account-level status flags before you send the first message.