Monday 3 August to Sunday 9 August 2026. The second biggest sending week of the engagement, and the first with two calls booked in seven days: Karan Virmani, CFO of customs broker Profreight, and John Moran at Rock Solid Logistics, who handed over his mobile number and was on the phone with us the same evening. Two leaks surfaced during the week, a copy mix-up and a dead booking link, and both were fixed the same day they were found. Both are documented in full below.
Sunday's call set the direction: double the volume and consolidate everything behind the winning angle. That happened. The new brokers registry campaign went out to 2,634 new people this week, built from the official CBP and FMC broker registries rather than scraped lists, and it produced 27 of the week's 29 replies and both booked calls.
The two calls are exactly the kind this engine was pointed at. Karan Virmani is the CFO of Profreight, a licensed US customs broker and forwarder in New Jersey trading since 1989, and his first reply was a technical question about Section 301 notation, which is a question only a buyer asks. John Moran at Rock Solid Logistics chased us with his mobile number after a colleague passed our email up the chain, and Jordan had him on the phone within two hours of his message.
Two things went wrong this week, and both are stated plainly rather than buried. First, the broker campaign started its run carrying importer copy, the exact mismatch the Week 5 report flagged, and 1,375 brokers received it before Adrian caught it on Wednesday. It was paused, rewritten to the broker frame, and relaunched the same afternoon. Second, a follow-up email carried a booking link that led to a dead page, and Karan hit it while trying to book. Adrian flagged it at 7:52am Wednesday and the corrected link was out nine minutes later. Karan booked within the hour. Both incidents, and what now prevents each, are in their own section below.
Open tracking stays switched off on all your campaigns, because open-tracking pixels hurt deliverability, so there is no open rate here and there will not be while we send this way.
| Metric | Result | Read |
|---|---|---|
| Emails sent | 5,186 | Second only to last week's 5,786, and this time behind one consolidated angle |
| New people reached | 2,634 | Almost all of it the new brokers registry campaign |
| Replies received | 29 | 1.10% of people reached, up from 1.03% last week |
| Calls booked | 2 | Karan Virmani (CFO, Profreight) and John Moran (Rock Solid Logistics), full cards below |
| Bounces | 85 | 3.2% of people reached, above the 2% line we hold ourselves to. Read on this below |
On the bounce rate: the registry lists are built from official CBP and FMC filings, which is why the targeting is so accurate, but registry contact data is older than scraped data and many of the small brokerages run domains that cannot be fully verified before sending. Every address still went through verification, the misses are concentrated in those unverifiable domains, and the bounces are spread thinly across 134 sending inboxes, so no single inbox took meaningful damage. A recovery lane is re-verifying the parked addresses before any of them are used.
| Campaign | People reached | Replies | Calls booked | Verdict |
|---|---|---|---|---|
| Brokers registry (new this week) | 2,634 | 27 | 2 | The whole week's result lives here. Both booked calls, the qualified pipeline below, and the technical questions all came out of this lane. |
| Classifier decision makers (follow-ups) | 0 new | 2 | 0 | Sequence finished its run. Its remaining value is the open conversations it already produced. |
Both campaigns finished their sequences over the weekend, which means sending is at zero until the next batch loads. That batch is staged and is the biggest item in the plan below.
Cumulative since outbound began: 15,945 emails to 8,553 people, producing 75 replies and three booked calls. Nearly two in five of every reply this engagement has produced arrived in the last seven days.
Twenty-six people replied for the first time this week, and the other three messages were second replies in threads already open. This is the honest split of the twenty-six.
| Outcome | Count | Read |
|---|---|---|
| Booked a call | 2 | Karan Virmani, John Moran |
| Qualified, in conversation | 1 | Shunan Liu, President of Umax Shipping, asked for Ripple's profile |
| Interested, being worked | 2 | Amanda Aykroid (O.T.S. Astracon), Leann Godfrey (Air Sea America) |
| Neutral: out of office, soft no | 4 | Two auto-replies, two polite passes |
| Not interested | 12 | Mostly one-line passes. Several trace to the importer copy that went out before Wednesday's fix |
| Asked not to be contacted | 5 | All five suppressed the same day, across every channel, permanently |
LinkedIn sent nothing again this week. The Week 5 report said the reload was on the list, and it did not happen, because the copy incident and the volume push took the week. That is the honest reason. The channel's history says it earns its slot back: a 26% acceptance rate is strong for cold LinkedIn, and the reload is item five in the plan.
| Metric | This week | To date |
|---|---|---|
| Connection requests sent | 0 | 418 |
| Connections accepted | 0 | 107, a 26% acceptance rate |
| Messages sent | 0 | 152 |
| Replies | 0 | 25 |
2,634 people reached produced 29 replies, which produced 7 live conversations, which produced 2 booked calls. Last week the bottom of this funnel moved for the first time. This week it moved twice, and both bookings came inside three days of first contact.
The broker frame works. Week 5's biggest lesson was that brokers were being pitched as if they were importers, and this week proved the correction: written to as businesses that sell classification, brokers replied at scale, and the two who moved fastest were a CFO with a technical compliance question and an operations manager who volunteered his mobile number. The registry list did something scraped lists have not: it put us in front of licensed brokers almost exclusively, which is exactly who the product serves best.
Week 5 made a promise: every booked call appears in the next report by name, booked, held, or missed. Here is the ledger, then the cards. Every quote is the prospect's own words copied from the thread, and every card names the channel, the inbox it landed in, and the date and time in UK hours.
| Call | Booked for | Status |
|---|---|---|
| Karan Virmani, CFO, Profreight | Tue 11 Aug, 8:00pm UK (3:00pm Eastern) | Booked. Karan asked on Sunday to move it to 9:30pm UK the same day or 8:00pm UK Wednesday. Waiting on Adrian's pick, needs confirming today |
| John Moran, Rock Solid Logistics | Thu 6 Aug, 7:00pm UK (11:00am Pacific) | Booked and due to have happened. We do not have the outcome on record, and we need it from Adrian: held, moved, or missed |
| Dylan Dickman, CEO, Known | Thu 30 Jul (last week) | Outcome still unconfirmed to us. Asked in the Week 5 report, asking again. He stays the highest-fit prospect in the pipeline until we know |
Adrian also mentioned on Sunday's call that he had pushed a calendar invite onto a strong prospect of his own for Wednesday 5 August. Whatever happened there, add it to this ledger and it gets tracked with the rest.
Last week's report promised that failures get named rather than buried. Two happened this week. Both were found mid-week, both were fixed within hours of being found, and both now have a rule that stops the repeat.
On launch day a technical failure at the sending platform stripped the email bodies out of the new campaign, and in the rebuild that followed, the sequence was cloned from the importer campaign. The result: the brokers registry list, 1,375 people in, was receiving copy written for importers, the precise audience mismatch the Week 5 report had just diagnosed. Some of the week's bluntest rejections trace to those sends.
Adrian caught it on Wednesday. The campaign was paused on the spot, the broker-frame copy was restored in two variants, one leading on capacity and margin, one asking whether they classify through CargoWise or Descartes today, and the campaign relaunched at 5:45pm UK the same day. The 1,375 who received importer copy get only the short follow-up, not a restarted sequence. Every reply and both booked calls above came in spite of this, which says something about the strength of the list itself.
The rule now: no copy variant goes live without written sign-off from Ripple. A social-proof line referencing the Jazz deal was also in the live rotation without that written confirmation, and it was pulled at relaunch for the same reason. It ships again the moment Adrian or Jonny confirms it in writing, and not before.
Karan Virmani, the CFO above, tried to book on Tuesday night and got "page not found." The link in that follow-up pointed at a calendar page that no longer existed. He told us directly, Adrian flagged it at 7:52am Wednesday, the corrected link was in Karan's inbox nine minutes later, and he booked within the hour. Every live sequence now carries the verified working link, and the booking chain has produced two clean bookings since, with the booking notifications firing correctly on our side both times. Whether each one landed in Adrian's diary is exactly what the ledger above exists to confirm.
The uncomfortable part is the unknown: a dead booking link fails silently, so we cannot know how many people before Karan clicked it, hit the dead page, and simply left. He is the one who said so. That is the strongest argument for the ledger above, where every booking, and every booking that should have happened, is tracked by name until it is closed out.
Week 5 diagnosed it, Week 6 proved it. The broker frame, where they run the classifier for their own clients and keep the billing, produced 27 replies, both booked calls, and technical questions from a CFO and two presidents. The registry list put us in front of licensed brokers almost exclusively, and the conversations are a level more serious than anything the scraped lists produced.
The copy mix-up did not happen while anyone was writing copy. It happened during a technical rebuild under time pressure, when the wrong sequence got cloned. The dead link did not happen in the calendar tool, it happened in a follow-up template pointing at a page that had been renamed. Both fixes are procedural: written sign-off before any variant ships, and the booking link verified as part of loading any campaign.
This week bounced 3.2% against last week's 1.5%, because official registry filings age faster than freshly scraped data and small brokerages often run unverifiable domains. The same data is why a CFO of a 37-year-old brokerage was on the list at all. The bounces were spread across 134 inboxes with no single inbox harmed, and the parked unverifiable addresses go through a recovery check before any future use.
John Moran gave us his number at 7:33pm and Jordan was on the phone with him before 9:30pm, with a call booked for the next day. Karan hit a dead link at night and had a working one nine minutes after it was flagged the next morning. Shunan asked for a company profile and had it in eight minutes. Every fast response this week moved a deal forward, and the one place we were slow, LinkedIn sitting empty for a second week, produced nothing at all.
Two booked calls and five live conversations outrank everything else, so the week starts with the calendar and ends with volume.
Adrian picks 9:30pm UK Tuesday or 8:00pm UK Wednesday, tells us, and Karan has his confirmation within the hour. The dossier with discovery questions is already live and linked from his card above.
Three bookings have unknown outcomes on our side: John Moran on Thursday, Dylan Dickman from 30 July, and the invite Adrian pushed himself for Wednesday. One line from Adrian on each, held, moved, or missed, and each gets its right next step, whether that is a follow-up, a rebook, or a clean close.
Ten days since a CEO said "let's have a call next week." The sequence is done, the relationship is Adrian's, and one short personal message settles it in either direction.
Both campaigns finished their runs over the weekend, so sending is currently at zero. Roughly two thousand more contacts close the gap to the volume target Sunday's call set: about a thousand parked addresses are already in re-verification, and the remainder comes from the same registry and database sources that built this week's list. Everything loads behind the broker copy that just proved itself, and nothing loads until the written sign-off from the leak section exists for every variant in the sequence.
Promised last week, displaced by the incident, still the right move. A fresh shortlist drawn from the same broker profile that produced every reply this week, filtered to people active on LinkedIn in the last thirty days, loaded on both accounts.
Where this leaves us. Six weeks in: three booked calls, two of them in the last seven days, all three with the exact buyer profile the engine was built to reach. The question is no longer whether the market replies, or even whether replies become calls. It is whether booked calls become held calls, and that is decided by the ledger discipline above and by Adrian's five minutes of input this week more than by anything we send.