UK agency playbook
A practical Companies House prospecting workflow for marketing agencies
For UK agency owners and heads of outbound who already pull from Companies House (or know they should) — and keep losing the week between the free export and the first real conversation.
For UK agency owners and heads of outbound who already pull from Companies House (or know they should) — and keep losing the week between the free export and the first real conversation.
Companies House is free. Your Tuesday still is not.
The filings library is public. The time you spend turning rows into qualified local targets, then into emails that do not sound like spam, is not. This post is a practical workflow you can run with free tools — and a clear map of where DIY quietly dies so you can decide what to keep manual and what to productise.
Who this is for (and who it is not)
For:
- Local marketing agencies selling to UK service businesses, trades, shops, and similar
- Lead-gen agencies whose job is a steady stream of new local targets
- Operators who already open Companies House once a week and hate the afterwork
Not for:
- End local business owners looking for consumers
- People hunting “millions of free B2B emails”
- Anyone expecting a compliance or legal opinion on data use (get proper advice for your stack)
What Companies House is good at (and what it is not)
Good at:
- Public signals that a company exists, was incorporated, or changed in a way that shows up in filings
- A starting universe you did not have to buy
- Teaching juniors what a real UK company record looks like
Not a full prospecting system:
- It does not rank “best fit for our offer this week”
- It does not put a clean decision-maker email on a silver plate every time
- It does not track whether you already emailed them, got ghosted, or booked a call
- It does not write the first outreach in your agency’s voice
Treat it as a library, not a lead desk.
The working method (DIY, honest)
Step 1 — Decide the filter before you open the site
Write four lines on a sticky note:
- Geography (e.g. M25, Greater Manchester, specific postcode clusters)
- Business types you can actually help
- Signal you care about (newly incorporated, expanding footprint, director change only if it maps to your offer — do not collect noise)
- Disqualifiers (too small, pure holding company, wrong SIC cluster, already a client competitor you will not touch)
If the filter is vague, the export will be noise and you will blame “data quality” when the real issue was taste.
Step 2 — Pull a narrow list, not a heroic dump
Aim for a list you can work this week, not a trophy CSV.
- Prefer recent signals over “everything in the sector forever”
- Cap the first pass so a human can review it in under 90 minutes
- Export only fields you will use (name, number, address/region, date/signal, status)
Volume without a work plan is how folders of abandoned CSVs get born.
Step 3 — Qualify in one ruthless pass
Create five columns (spreadsheet is fine here):
| Column | Purpose |
|---|---|
| Fit (Y/N/Maybe) | Does this match the sticky-note filter? |
| Local proof | Website, Google presence, or clear local service footprint |
| Why now | One line: newco, expansion, or other timing reason |
| Contact path | Website form, LinkedIn, known email pattern, phone — whatever is real |
| Next action | Email / LinkedIn / skip / nurture later |
Rules that save hours:
- Maybe becomes No if you cannot explain “why now” in one line
- No website and no discoverable presence → skip unless your offer specifically helps brand-new entities with zero footprint
- Holding companies and pure SPVs → skip for most local agency offers
Step 4 — Enrich only what you will contact
Do not boil the ocean.
For each Yes:
- Confirm the trading name people actually use (often different from the legal string)
- Find one human or one clear inbox path
- Note one concrete local detail you can mention without stalking
- Stop. You are building a conversation list, not a data product
If enrichment takes longer than the eventual email, your process is upside down.
Step 5 — First touch within 48 hours of the signal (when you can)
Freshness decays. A “newly opened” angle lands better in week one than week six.
Simple structure that stays human:
- Why you are relevant to their local situation
- One specific observation (opening, area, service type) — not a dossier
- Soft ask (15-minute chat, useful teardown, clear next step)
- Easy opt-out tone — pressure is not a strategy
Personalisation means one true detail, not a paragraph of scraped trivia.
Step 6 — Put every Yes in a pipeline home the same day
Minimum stages that work for a small agency:
New → Researching → Contacted → Replied → Meeting → Won/Lost
Rules:
- Spreadsheet is allowed until it lies to you
- “Contacted” only when something actually left the building
- Review the board weekly or the board becomes a graveyard
Step 7 — Run it on a fixed weekly cadence
Example for a 2–10 person agency:
| Day | Job |
|---|---|
| Mon | Pull + filter (45–60 min) |
| Tue | Qualify + enrich Yeses (60–90 min) |
| Wed–Thu | First touches + light follow-up |
| Fri | Pipeline tidy + note what language got replies |
Protect the block like a client workshop. Prospecting that “happens when we have time” does not happen.
Where DIY Companies House workflows die
- Export theatre — big pull, zero outreach
- SIC cosplay — codes treated as perfect ICP labels
- Enrichment addiction — research as procrastination
- No owner — junior dumps a sheet; nobody books the send block
- No memory — same companies reappear; nobody knows who was emailed
- Tool sprawl — CH + Sheet + CRM + mail tool + sticky notes, none of them the system of record
- Stale pride — working a three-month-old “new business” list and wondering why replies died
Free data does not equal free pipeline time. The tax is human attention.
A one-week walkthrough (anonymised)
Agency: 5-person local marketing shop, North West England. Offer: websites + local lead gen for trades and home services. Filter: new or expanding service businesses within a defined drive-time radius.
Monday: 40-minute pull focused on recent signals in target SICs/areas — not a national dump. Tuesday: 70 minutes qualifying; 18 Yes, 11 Maybe→No, rest skipped. Wednesday–Thursday: 18 first touches with one local detail each; all logged as Contacted. Friday: 3 replies (mix of “not now” and “send more”); 1 meeting booked; language from replies noted for next week’s openers.
No miracle metrics claimed here — just a loop that finishes the week with conversations instead of a colder spreadsheet.
How FoundryHQ fits (without replacing your brain)
FoundryHQ is built for UK local marketing and lead-gen agencies that want this loop as a weekly habit:
- Fresh newly established and expanding local business signals (UK-first)
- A visual lead desk and pipeline so targets do not die in a sheet
- One-click AI first outreach drafts you still edit and send as a human
It does not argue Companies House is useless. Companies House is the library. FoundryHQ is the operator layer that turns signals into a desk you can work every week.
Agency plan: £79/month. No free trial.
We will not claim exclusive data, guaranteed reply rates, or that you must throw away a CRM you already love. The test is simple: does Tuesday get shorter, and do more fresh locals reach Contacted?
Next step
If you want the end-to-end picture — signals, desk, outreach — start here:
https://foundryhq.agency
When you are ready to work this week’s list in a real desk:
https://app.foundryhq.agency/login?next=/leads
Related reading (cluster)
- How UK local agencies find newly opened business clients without living in spreadsheets
- Fresh local leads vs big stale lists: what actually fills an agency pipeline
- (Next) The first outreach email to a newly opened local business
FAQ
Is Companies House enough on its own?
Enough to start. Rarely enough to run a calm weekly outbound system without extra process, enrichment, and pipeline memory.
How “new” should a company be before I contact them?
Sooner is usually better for a newco angle. If the signal is expansion rather than incorporation, time the message to the expansion, not an arbitrary calendar rule. Stale “new” lists underperform.
Can I email every company I find?
You are responsible for lawful, decent outreach in your jurisdiction and channel. This article is workflow guidance, not legal advice. Prefer relevance and restraint over spray.
Should juniors own the full workflow?
Juniors can pull and first-pass qualify with a tight filter. An owner should still define ICP, review edge cases, and protect the send block — otherwise you industrialise noise.
What is the first process upgrade if I only change one thing?
Same-day pipeline home for every Yes, plus a protected weekly pull block. Tools help after the habit exists.
Next step
Put the weekly loop somewhere it survives Tuesday
FoundryHQ is built for UK local marketing and lead-gen agencies: fresh newly opened and expanding business signals, a simple lead desk, and one-click AI first outreach. Growth Toolkit £79/month or Enterprise / Agency £1,000/month. Growth Toolkit includes a 7-day trial.