3,388→1,921
our own scrape, washed
Chain branches where the manager can't sign anything. Inboxes read by a group HQ two counties away. Dead domains. A nail salon in Florida because it's also called Windermere. LeadWash strips them out — and hands you a receipt for every rejection.
Bring your own scrape, or just name a trade and a few towns — we'll source the raw rows, wash them, and hand back the receipt. Your data stays yours; we store nothing.
Verification services are excellent at one question, and it isn't the one that decides your reply rate. They ask does this address exist. We ask is there anyone behind it who can say yes. A chain branch inbox passes every verifier on the market at 100% — and still wastes your send.
| The lead | An email verifier says | LeadWash says |
|---|---|---|
| Travelodge York, reception@travelodge.co.uk | ✓ Valid · deliverable, low risk | ✗ Rejected — chain branch. The person reading this can't buy anything. |
| The West Gate, info@greeneking.co.uk | ✓ Valid · corporate domain, good reputation | ✗ Rejected — email domain ≠ site domain. Group HQ reads this, not the landlord. |
| My French Cafe, Orlando FL | ✓ Valid · nothing wrong with the address | ✗ Rejected — wrong country. You searched a UK town; this isn't in it. |
| Sotto Sotto, sottosotto@gmail.com | ? Unknown · free mailbox, can't confirm | ✓ Keep — mailbox name matches the business. Independent, owner-run. |
We're not competing with verifiers — run one after us if you like. They de-risk the send. We decide the list.
Spot-check ten rows. If the reasons hold up, trust the batch. That's the whole deal — no black-box score, no "AI confidence 87%". And anything we can only guess at gets tagged, never rejected.
A verification vendor prints one number for everything they do. You can't check it, and it doesn't break down by what actually failed. So here is ours the other way round: per rule, on labelled real rows, with the sample file shipped alongside so you can re-run it and get the same answer.
Sample size is 14 rows scraped from live Google Maps listing pages and labelled by hand. That is small, and we're saying so rather than rounding it up to a marketing number. The rules that need a website, an email or a DNS lookup aren't in this table yet, because listing pages don't carry those fields — they get measured when we have labelled data that does. Both sample files are in the repo under samples/.
No export, no upload, no API key. Search Google Maps the way you already do, scroll the results, and every listing gets marked as you go. Then take the clean list with you.
Nothing leaves your browser. The extension reads the page you're already looking at. No Places records touch our servers — the same BYO promise, enforced by architecture instead of policy.
The list view runs the cheap rules — chain names, closed listings, review volume, category and price band. It's a first pass that cuts a 3,000-row search down to a few hundred worth a closer look.
The expensive rules still need the full record. Website URL, email domain and DNS aren't on the results page — so branch-path, HQ-inbox and dead-domain checks stay in the CSV pipeline. We'd rather say that than pretend the extension does everything.
Rejecting dead rows saves your domain. But the rows we keep come back armed: we visit every website, check what's missing, and write the finding as a line your sales rep can open with — plus the freshest customer review as personalisation ammo, and the email checked down to SMTP level.
Outscraper, Apify, Places API exports — recognised automatically. Odd CSV? Map columns once.
name + address is enough to startEvery rule that can run, runs. Every rule that can't is reported with the exact count — never silently skipped.
per-rule coverage with every batchYour original columns, untouched, plus tags and a rejected list with one reason per row. Your outreach, your tools.
washed.csv + rejected.csvMost cleaners fail silently when your file is missing a column — the output looks normal and the accuracy quietly collapses. Worse, most tools fold don't know into no, which is how you end up pitching a restaurant a booking system it's had for three years. We report what ran, on how many rows, and what we couldn't check.
Tell you whether a Gmail address is live. Google throttles and blocks bulk probing, so nobody does this reliably — including the services that advertise 99.9%.
Judge whether the mailbox belongs to the business or to an individual. whitbyfishbar@ and sarahjones87@ are not the same lead.
A headline accuracy percentage. Every vendor's figure measures something other than whether your list was any good. Spot-check ten rows instead — that's a real test.
We built LeadWash for our own cold outreach to UK & Ireland small businesses: 3,388 scraped rows washed to 1,921 sendable, zero bounces since. Now we want to see it fail on someone else's data — so the first one's on us. You get back a sendable list, a rejected list with a reason on every row, and an honest note on what we couldn't check.