Sole Trader to Large Business: A UK ICP Guide

Most ICP advice assumes you are selling to US mid-market SaaS companies. It is almost useless in the UK context, where the business population is shaped differently: roughly 5.9 million active businesses, 76% under £632K turnover, a long tail of sole traders, and a legal-entity structure (Ltd, LLP, sole trader, partnership) that maps directly to your compliance and pricing decisions.

This guide covers how to define a UK-native ICP using the actual data you have access to — revenue bands from Companies Act thresholds, SIC codes from Companies House, and practical qualification filters that don’t require guessing.

What an ICP actually is

Before the how, the what. An Ideal Customer Profile (ICP) is a description of the kind of company most likely to buy your product, value it, stay, and not be expensive to support. Key distinction:

ICP = company attributes (size, industry, stage, model).
Persona = person attributes (role, seniority, goals, pains).

Both matter. ICP decides whether to reach out at all; persona decides who to reach inside the company.

Revenue bands — the UK-native segmentation

The UK Companies Act 2006 (as amended by SI 2024/583) defines three filing thresholds:

Band	Turnover	Balance sheet total	Employees
Micro-entity	≤£632K	≤£316K	≤10
Small company	≤£10.2M	≤£5.1M	≤50
Medium company	≤£36M	≤£18M	≤250
Large	>£36M	>£18M	>250

A company must meet at least two of the three thresholds to sit in a band. These thresholds are public, self-declared at Companies House, and update within 24 hours of filing. They are the single most reliable UK B2B segmentation data point.

Why this is better than US-style “SMB/mid-market/enterprise”:

UK-specific (US bands are higher)
Public, verifiable (Companies House spot-check)
Directly maps to affordability (a micro-entity cannot pay £300/mo per seat)

Practical rule: align your pricing bands to these thresholds. It is what LeadKing does (B1 Starter for micro, B2 Team for small up to £2M, B3 Growth for small up to £10.2M, B4 Scale for medium/large). Customers understand it immediately.

Why revenue bands matter more than headcount

A lot of UK B2B prospectors default to headcount (“10–50 employees”) because LinkedIn filters on it. Headcount is a weak proxy for the two things that really decide a B2B deal: budget available and speed of decision.

Headcount can mislead. A ten-person agency and a ten-person AI startup have different budgets by an order of magnitude. Revenue separates them cleanly.
Turnover is a decision-permission proxy. A £500K-turnover company has a founder-led, single-decision-maker buying pattern. A £5M-turnover company has a small finance team and a founder, which usually means one-step approval above a threshold. A £25M-turnover company has boards, budgets, and buying windows. Revenue is the fastest way to predict which of these you are dealing with.
Balance-sheet total is under-used. Asset-heavy companies (construction, logistics, manufacturing) can have modest turnover but significant balance sheets and procurement sophistication. Tracking both turnover and balance sheet makes a better decision than either alone.

The practical move: set your primary qualification filter on turnover band, your secondary filter on employees, and keep balance-sheet as a tie-breaker for asset-heavy verticals.

SIC codes — the vertical filter

Every UK-registered company declares a Standard Industrial Classification (SIC) code on filing. SIC codes are public at Companies House.

Useful SIC patterns for B2B lead gen:

62xxx — Computer programming, consultancy, information service activities (most SaaS + software companies)
70xxx — Head office; management consultancy
73xxx — Advertising and market research
82xxx — Office administrative, office support, business support
85xxx — Education
74xxx — Other professional, scientific and technical activities

The SIC system has quirks (a recruitment agency might sit in 78xxx but also 82xxx; a fintech might be 62xxx or 64xxx depending on licence). Don’t treat a single SIC code as authoritative — use it as a filter on top of other signals.

Legal structure: the overlooked axis

Ltd, LLP, sole trader, ordinary partnership, charity, CIC. UK legal structures are not cosmetic. They change who you can email, what you can contract with, and how the company actually makes decisions.

Ltd company. Corporate subscriber for PECR. Governed by the Companies Act. Directors have fiduciary duties. Formal contracts. Typical decision cycle for sub-£20K annual spend: one to three weeks.
LLP. Corporate subscriber for PECR. Members replace directors. Common in professional services — law firms, accountancies, consultancies. Decision-making is often partner-consensus rather than CEO-led.
Sole trader. Individual subscriber for PECR. One person, personal liability. Decision cycle: hours to days. Emotional rather than committee-driven.
Ordinary partnership. Individual subscribers for PECR. Rare now but still present in some legal and farming contexts.
CIC or charity. Community Interest Companies and charities have their own governance. Procurement is often grant-linked and slow.

The operational rule: build your ICP with a legal-structure field and use it for both compliance routing and outreach tone. A sole-trader consultant responds to a different email than the CFO of a Ltd accountancy.

Regional patterns that actually matter

UK regional data is noisy, but two patterns are robust enough to use in ICP definition.

London-registered companies behave differently. London postcodes over-index on SaaS, fintech, media, and professional services. Budgets are 20–40% higher than equivalent-turnover businesses outside London, per publicly reported benchmarks. Decision cycles are marginally longer because more stakeholders are involved.
Regional clusters exist and are worth respecting. Manchester and Leeds for digital and media; Edinburgh and Glasgow for financial services; Cambridge, Oxford, and Bristol for deeptech; Birmingham for industrial; Belfast for cyber; Cardiff for public sector. Selling into a cluster means competing for attention with every other vendor that has noticed the same cluster.

The mistake to avoid is assuming “UK is one market”. It is not. The mistake to also avoid is stereotyping too hard; plenty of exceptions exist. Use region as a weighting factor, not a hard filter.

Common UK ICP mistakes

1. Too broad

“UK SaaS companies” is not an ICP. It’s 30,000+ companies. Your outreach capacity is 20–50 per week. The maths does not work.

Narrow until your list is 500–2,000 companies. That is the actionable range.

2. Too narrow

“UK-based fintech with a regulated consumer-lending licence operating in retirement planning founded after 2023 with a female CTO” is not an ICP. It’s a specific company. If you accidentally win them, you have no second customer.

3. US transplant ICP

“£1M–£10M ARR companies” does not map to the UK population. £1M ARR is rare in the UK until Series A. Use revenue-band thresholds, not US-style ARR tiers.

4. No refresh

ICPs drift. What worked at your first 10 customers may not work at customer 50. Re-derive your ICP quarterly from actuals — which companies converted, which churned, which paid top-of-band — not from assumptions.

5. ICP derived from hopes, not close data

Writing an ICP from “who should buy this” rather than “who has bought this” is the most common founder mistake. Aspirational ICPs feel visionary and are almost always wrong. Until you have at least ten closed-won customers, your ICP is a hypothesis. Treat it that way: write it down, test it, revise it from actual sales data every six weeks.

6. No disqualification layer

A strong ICP does not just describe who is in — it describes who is out. Common disqualifiers for UK B2B:

Dormant companies (filed dormant accounts in the last 12 months)
Insolvency or winding-up filings at Companies House
Recent director resignation without replacement (often signals a failing business)
Public-sector-heavy turnover if your product is not procurement-cleared
Regulated sectors where you lack specific compliance track record (healthcare, education)

Writing down disqualifiers saves enormous time. A list of 2,000 ICP-matching companies minus 200 disqualifiers is a much cleaner list of 1,800 than a list of 2,000 with silent disqualifiers you discover mid-conversation.

UK-specific ICP dimensions worth tracking

Dimension	Why it matters
Turnover band	Maps to affordability and pricing
SIC code cluster	Vertical / substitute signals
Legal structure (Ltd, LLP, sole trader)	PECR + contract decisions
Region (registered office)	Sales cycle + in-person vs remote
Incorporation year	Maturity / budget stage
Director age + background	Buying authority
Hiring velocity (12 months)	Growth stage signal
Funding history	Budget-unlock signal
Web presence (has website? uses X stack?)	Maturity signal

A worked UK ICP example

Take a hypothetical payroll-automation product priced at £150 per month. A credible UK ICP for this product might read:

Primary segment: Ltd-registered UK companies, turnover £632K–£10.2M (small company band), 10–50 employees, operating for at least eighteen months.
Verticals: SIC codes 62xxx (software), 70xxx (management consultancy), 73xxx (advertising), 82xxx (business support). Exclude 85xxx (education — different procurement).
Region: weighted equally London, Manchester, Bristol, Edinburgh. Under-indexed regions (Northern Ireland, rural Wales) kept but de-prioritised.
Decision-maker persona: CFO, head of finance, head of people, or founder-operator acting in those roles. Not the IT team.
Signals that trigger outreach: first HR hire in the last 60 days, first finance hire in the last 60 days, recent director change at CFO level, outgrowing Xero complaints within the last 30 days, Companies House filing transition from micro-entity to small company.
Disqualifiers: insolvency filings, sole-trader status, over-50 employees (outgrown band), pre-existing payroll contract with a known competitor we cannot displace.

Sized correctly, this ICP resolves to roughly 3,000 UK companies, of which 100 to 200 show meaningful intent in any given month. A team of one or two operators can work this list properly with weekly discipline.

The exercise of writing the ICP down in this concrete form is half the value. It makes trade-offs explicit. It tells your tool what to filter on. It lets you reject a shiny-but-off-ICP lead without guilt.

How LeadKing builds and refines ICP per customer

Our approach:

User describes offer and buyer in plain English. No form fields with dropdowns — a paragraph.
AI translates that into a structured ICP (turnover band preference, SIC code cluster, legal structure, signals that matter most).
User confirms or edits the structured ICP. You stay in control.
Weekly runs produce candidates scored against the ICP + signals.
Feedback re-weights the ICP. If you consistently mark “too senior” or “wrong vertical”, next run’s ICP shifts.

After 5+ marked outcomes, the system auto-retunes. You see a weekly trend card: “hub-X signal down-weighted”, “Y SIC code up-weighted”. No black box.

Frequently asked questions

How often should we redo our ICP? Formally, every quarter. Informally, every time you close or lose a deal that surprises you. The formal review averages the surprises; the informal review catches them before they distort your quarter.

What if our customer base is too small to derive an ICP from data? Below ten closed-won customers, treat your ICP as a hypothesis and be explicit that you are exploring. Over-confidence in a small-sample ICP narrows your focus too early and can kill otherwise-good markets. Write your best guess, but flag the assumptions for retesting.

Can one business have multiple ICPs? Yes, but not many. Two is typical, three is stretch, four starts to mean you do not really know who you serve. When a second ICP emerges clearly — different size, different vertical, different pain — maintain it as a separate track with its own signals and messaging.

Is revenue data from Companies House accurate? For small companies and above, yes — accounts are audited or at least independently filed and are legally reliable. For micro-entities, the figures are self-declared and less comprehensive. Treat micro-entity revenue as directional, not precise.

How does ICP interact with product-market fit? A clear ICP is evidence of approaching fit; a fuzzy or constantly-shifting ICP is evidence of not yet having it. The two are tightly related. Investing in ICP discipline is, in effect, investing in fit.

Self-check: is your ICP working?

Red flags that your ICP is off:

Most qualified leads come from your own network, not inbound — your ICP’s too narrow or too different from your actual close pattern
Your top-of-funnel is large, conversion is low — your ICP’s too broad
You find yourself “making exceptions” for customers outside ICP — time to re-derive, your ICP is stale
New reps cannot describe the ICP in one sentence — it’s not operationalised

From ICP to persona to message

An ICP is a door. Behind it are people. Mapping the ICP to the right person and the right message is where most outreach fails even when the ICP is good.

For a UK B2B Ltd company in the small-company band, the persona map typically looks like this:

Founder or MD: owns everything under about £20K annual spend. Busy. Replies fast if the pitch is relevant and specific. Tolerates one follow-up; ignores more.
Head of finance or CFO: owns financial stack, cost-cutting, efficiency tooling. Slower to reply. Responds to ROI-first language and specific financial claims.
Head of sales or revenue: owns pipeline tooling, CRM, and outbound. Replies to peer-language (“we sell to teams like yours”) and concrete numbers.
Head of people: owns HR, payroll, onboarding, L&D. Relatively receptive. Prefers soft framing and proof of compliance.
Head of operations: owns process, IT adjacencies, internal tooling. Responds to specificity about time saved and integration ease.

Pair the signal that triggered the lead to the persona: a “first SDR hire” signal means the head of sales or revenue is the right person, not the CFO. A “director change at CFO level” signal means the CFO’s first 90 days is the context, and outreach should show awareness of that.

Messaging should change by persona. A six-line email that works on a sole-trader consultant falls flat with a CFO. A three-paragraph ROI pitch loses a founder. Matching register to role is a quiet skill, and most teams get it roughly right after about fifty sent emails and a notebook of what worked.

What to do next

If you want the broader context, read UK B2B Lead Generation in 2026.
For the signals side of qualification, see What are buyer intent signals?.
If you want to see how LeadKing builds ICPs per customer, see how it works or join the waitlist.