Counted, not claimed.

How we work, shown on our own markets. No client result is claimed on this site.

Our own campaigns

Our own campaigns, 1 August to 16 October 2026

Before we run anyone else's, we run our own. These are counted, not rounded up.

14,947

emails sent

167

bounced, about 1%

owners and managers replied with real interest, counted by hand

We emailed commercial roofing, commercial cleaning, landscaping, mechanical contractors, pharma validation, welding cells, machine vision, water treatment, overhead cranes and packaging machines.
14,947 emails sent. 167 bounced, about a 1% bounce rate.
56 owners and managers replied with real interest in what we offered. We counted them by reading every reply, one at a time, and set aside the out-of-office and wrong-person replies and the suppliers who wrote back to say they sell the same thing.

Figures are from our sending platform as of 16 October 2026, with the interest count made by hand. They count emails sent and replies from people, not clients. We publish no reply rate: the campaigns ran in ten different markets, with different lists and different copy, so one rate would tell you nothing about yours. Calls agreed and calls held are not printed here, because we only print what we have confirmed.

One market, start to finish

One market, start to finish: independent forklift dealers

What we do before a single email goes out, shown on a market we tested in October 2026 for ourselves.

The buyer is written as one sentence first: independent US dealers that sell, rent and service lift trucks, not owned by the manufacturer.
The exclusions are written next: manufacturers, parts-only sellers, auction and used-equipment resellers, warehouse operators, crane companies, vehicle and trailer dealers, and dealers owned by a parent group.
Then a test before any build. We pulled 40 companies and scored each one against that sentence, reading its own website. The rule was set in advance: build only if at least half are the buyer and the market holds at least 40 firms.
It passed. 25 of the 40 were clearly the buyer, 12 were not and 3 were unclear.
We ran the same rule on four other markets that week. Three failed and one needed a hand check, and we did not build any of the four.
A hand check of the passing list cut it again, and not every company had a decision-maker email we could find and verify. The list we send to is smaller than the market, and we say so.

As of 18 October 2026 this campaign had not sent its first email, so no result is claimed for it. It is here to show the method.

Replies

How each kind of reply is handled

Asks to be removed: suppressed at once, for every campaign we run, and not emailed again.
Out of office, left the company, autoresponder: kept apart from real replies and never counted as one. An address that has left the company comes off the list.
Not interested: marked, and nothing more is sent to that person.
Wrong person: marked as wrong person, never counted as interest.
A question about what we sell: answered from what is written in the agreement. Nothing is promised that is not on paper.
Asks for a quote, a demo, a proposal, a scope call, a meeting or a call: passed to you with the whole thread, the company, the person and their title, the business day it arrives.

Every rate we judge a campaign on uses contacted as the bottom number, never replies, because a single dead mailbox can auto-reply on every send.

Data

Where the data comes from and how it is checked

Companies: public registries and directories for the trade, such as licensing lists and trade association members, and business listings.
Fit: each company's own website is read and scored against the one-sentence description of your buyer, before anything is paid for.
People: the job titles that decide this purchase, found in contact databases, and only for companies that passed the fit check.
Email: every address goes through two independent verifiers. An address either one marks invalid is never loaded.
Duplicates: a campaign that would email someone already emailed is paused automatically.
The order is cheapest check first, so money is spent on the companies that already look right.

Honesty

What we do not claim

We publish no client case study, client name or client logo.
We print no reply rate, because a single dead mailbox can auto-reply on every send. A campaign is judged against companies contacted, never against replies.
We do not print calls held. We print what we have counted.

Contact Breezy

See what it would look like for your market.

Start a conversation Read the full deck