Adsome · UK trades pilot

Pipeline

Generated 2026-07-31
node scripts/build-dashboard.mjs

32verified leads
15sites live
17A grade
4/28no site / bad site
24with an email
33median reviews
0cards posted
0QR scans

Where the leads are

Sourced by trade and by town, so no two neighbouring firms get the same pitch.

landscaping
15
tree surgery
9
fencing
8
Cheltenham
6
Stonehouse
2
Longlevens
2
Gloucester
2
Quedgeley
2
Stroud
2
Upton St Leonards
1

Every lead

Ratings read off Google directly. Site problems verified by opening the actual site. 21 of 32 are limited companies with filings we can check.

BusinessTradeSiteRatingGradeEmailStatusThe problem we fix
Five Valleys Arbor LtdMinchinhampton tree surgery BAD 5.0116 reviews A yes builtopen very thin stale site, footer still reads 2023
AW Tree Surgery & Garden Care LtdMinsterworth tree surgery NONE 4.995 reviews A no builtopen Facebook page only, no website at all
Treeation ArboricultureEbley tree surgery BAD 5.089 reviews A yes builtopen no HTTPS at all, so browsers warn visitors the site is not secure
Willow & Oak LtdNailsworth tree surgery BAD 5.085 reviews A yes builtopen low-end site builder, homepage title is literally Home, no enquiry form on any page, no responsive CSS
Tree Surgeon Gloucester LtdGloucester tree surgery BAD 5.073 reviews A no builtopen their old web domain now serves an Indonesian gambling spam site, so anyone who finds them online lands on that instead
I Nixon LandscapesWinchcombe landscaping BAD 5.050 reviews A yes builtopen site was last touched in 2013, built on table layout with no mobile support and no enquiry form
K. Collier Landscapes LtdLonglevens fencing BAD 5.048 reviews A yes builtopen The site has no H1 heading and only about 2,400 characters of text in total, so Google has almost nothing to read and rank. Their work is far better than what the site says about it.
Dowdeswell Forestry LimitedLower Dowdeswell tree surgery BAD 5.046 reviews A yes builtopen No quote form anywhere on the site, no phone or address on the homepage, and their published email address has a typo with a space in it. Every enquiry has to work around the site.
Golden Valley Landscape Gardening LtdBussage landscaping BAD 5.042 reviews A no builtopen the domain does not resolve at all, so the site linked from Google is dead
Moe's GroupStonehouse landscaping BAD 5.038 reviews A yes builtopen Site is badly broken on a phone: text overflows the screen, the logo is cut off, and there is no phone, email or address anywhere on the page. Most customers look on a phone.
BRH Landscapes LtdCheltenham landscaping NONE 4.932 reviews A no builtopen no website at all, the Google listing points only to an Instagram profile
Severn Way Tree CareSlimbridge tree surgery NONE 5.027 reviews A yes builtopen Google itself shows 'Add website' on their profile. Facebook page only. Reviews describe large tree removals, so the work is high value but invisible online.
Elm GloucestershireInnsworth landscaping BAD 5.019 reviews A no builtopen site is completely dead, https refuses connection
Marcus Hannaford FencingUpton St Leonards fencing NONE n/a A yes builtopen No Google Business Profile at all and their old domain is dead, despite 225 Checkatrade reviews. Customers who hear his name have nothing to find.
Stephen Bassett Complete GardensLonglevens landscaping BAD 5.022 reviews B yes builtopen no quote form, blank placeholder images instead of photos, logo asset dated 2016
M.D Fencing and Landscapes LtdGloucester fencing BAD 4.833 reviews A yes queued no forms at all, WhatsApp link instead of a quote form, copyright stuck at 2024
KJP Fencing LtdCheltenham fencing BAD 5.023 reviews A no queued no quote or enquiry form on any page including contact, phone only
Marshalls Driveways & Sons LimitedRudford landscaping BAD 4.518 reviews A yes queued no quote form anywhere, thin template page, no copyright year
Greaterscapes LtdStroud landscaping BAD 5.065 reviews B yes queued no quote or enquiry form anywhere, not even on the contact page
T R Lane Fencing and Fabrication LtdStonehouse fencing BAD 5.042 reviews B yes queued two domains: Google links the abandoned placeholder at trlanefencing.co.uk, while their real site at trlanefencingfabrications.co.uk is current and responsive but has no enquiry form at all
Town & Country Trees & GardensBishops Cleeve tree surgery BAD 5.041 reviews B yes queued single-page brochure site with no service pages
Rowan Tindale Landscape DesignStroud landscaping BAD 5.037 reviews B yes queued no enquiry form, and no phone or email published anywhere on the site
Tec GardensChalford landscaping BAD 5.035 reviews B yes queued no quote or enquiry form on any page, newest content dated 2023
CPH Contractors LtdCheltenham fencing BAD 5.033 reviews B yes queued single-page site with anchor-only navigation and no service pages
B E RudmanCheltenham landscaping BAD 4.933 reviews B yes queued one-page site with anchor navigation and a yahoo contact address
Cheltenham Tree Services LtdCheltenham tree surgery BAD 5.023 reviews B no queued homepage has almost no text and no quote form, the form is buried on a separate page
Cheltenham Building & LandscapesCheltenham landscaping BAD 5.021 reviews B yes queued no quote or enquiry form on any page
Pattern Stone Landscape & DevelopmentsTuffley landscaping BAD 5.019 reviews B yes queued no form, footer admits stock images only, leftover roofing copy and a misspelt town name
J A Ford Fencing and LandscapingQuedgeley fencing BAD 5.016 reviews B yes queued Wix site with copyright 2018, zero forms, gmail contact address
Designer Driveways LtdQuedgeley landscaping BAD 5.014 reviews B yes queued no quote form, contact email sits on an unrelated domain, no company details
Charlton Kings Fencing LtdCharlton Kings fencing BAD 5.013 reviews B yes queued template landing page with barely a thousand characters of text and no quote form
Aspen GroundcareTewkesbury landscaping BAD 5.010 reviews B no queued no viewport meta tag so it does not scale on mobile, no form, contact details only inside an image

Built site quality

Google Lighthouse, run against the live URLs. SEO sits low on purpose: previews carry a noindex tag so they cannot compete with the client's own site before they say yes.

SitePerformanceAccessibilityBest practicesSEO
dowdeswell forestry 94889658
k collier landscapes 878810058
marcus hannaford fencing 878810058
moes group 789410058
severn way tree care 978810058

Unit economics

Every card price below came back from Stannp's own API in test mode, not an estimate. UK second class is what we actually pay; the international rate applies only to the test cards going to Stockholm.

£0.98per card, UK second class
£1.04on 350gsm silk, +6p
£2.02per card to Sweden
£18.74post all 15 + 2 to Carl
£0hosting, Cloudflare free tier
0 tokensper extra site
£310one customer, per month
1first payments to cover the postage

Revenue

Live from Stripe. Aggregates only: this page is public, so no client names or emails are ever written here.

2active subscriptions, all Adsome
15,156SEK monthly recurring
0pilot conversions so far
£310each pilot win adds

Refresh with node scripts/stripe-snapshot.mjs. Pilot wins are matched on subscription metadata against the lead slugs, so a UK customer appears here without being named.

What I think is going on

Hypotheses, not conclusions. Nothing here has been tested with a real prospect yet.

Bad sites may beat no sites

A firm with a broken site has already decided a website matters and has paid for one before. A firm with nothing has to be convinced the category matters at all. 28 of our 32 are bad-site leads, so wave one should tell us which converts.

The dead-domain leads are the sharpest pitch

Two leads have sites that fail outright: one refuses connections, one redirects to an unrelated beauty salon. That is demonstrable in five seconds and impossible to argue with, unlike "your site looks dated".

Email coverage is the real bottleneck

Only 24 of 32 have a findable email. The postcard sidesteps that entirely, which is the strongest argument for post over cold email here, but follow-up needs enrichment.

Risk: the QR is the whole funnel

If the printed code does not scan cleanly on 300gsm, response goes to zero and we will not know why. This is why the test cards to Stockholm come before any real send.

Risk: numbers on print go stale

Review counts change between printing and delivery. Cards now carry only durable facts, and the build refuses to print a card containing a rating. Numbers live on the sites, where they can be updated.

Open question: which front converts

Two front variants exist per lead, brand-pure and gold-accent. With five leads there is no signal. Split roughly 25/25 across wave one and the QR tracker will answer it per card.

What I would do next

1. Print test first

Top up about £11, post all five to Stockholm, check the QR scans and the pine prints true. Five days, and it protects the other 45 cards.

2. Build ten, not five

Ten sites costs no tokens and gives a real spread of trades and site problems to compare. Pick the ten sharpest problems, not the ten biggest firms.

3. Enrich emails before follow-up

Postcards do not need email, but the follow-up sequence does. Worth one Apollo pass across the whole list rather than per lead.

4. Keep the sheet as the mirror, not the master

leads/*.json stays the source of truth because the build reads it. The sheet is for you to read and annotate; one command re-syncs it.