Method
Pipeline stops being a guess when you can name the inputs
Forecasting isn't a spreadsheet skill. It's what becomes possible once the thing feeding the spreadsheet is a system with known inputs and measured conversion at every step.
1. Triggers beat firmographics
Most targeting stops at the firmographic layer: industry, headcount, region, funding. That describes a company that could buy. It says nothing about whether they're going to buy this quarter.
What actually moves someone from indifferent to interested is a change in their world. A new VP of Operations in month two, wanting a visible win. A second site or depot that broke the reporting that used to work on one. An ERP migration that orphaned a workflow. A compliance deadline with a date attached.
So the first thing we do is go through your closed-won deals looking for what was happening at those accounts right before they bought. That pattern, not the industry label, becomes the search.
What this changes in practice: a list of 1,000 companies that fit a trigger will outperform a list of 10,000 that merely fit a profile, on every metric that matters — reply rate, meeting rate, and the quality of the conversation when it happens.
2. Lists are built, not filtered
Pulling 10,000 contacts from a database takes four minutes and produces a list that looks identical to the one your competitor pulled. Worse, database records go stale fast, and a bounce rate above a few percent is the quickest way to destroy a sending domain.
We build account by account. Every contact is verified, every account carries a research note explaining why it's on the list at all. If we can't write a sentence about why this company should care this month, it doesn't belong there.
A thousand contacts built this way takes real time. It's most of what the setup fee pays for.
3. Deliverability is the hidden variable
This is the part that kills outbound programmes silently. Your emails go to spam. Nothing errors. Your dashboard shows sends going out and nobody replying, and the conclusion you draw is that outbound doesn't work for your market.
- Never send cold from your primary domain. If reputation goes, it takes your invoices and your support replies with it.
- Secondary domains, properly authenticated. SPF, DKIM and DMARC configured before a single message goes out.
- Warm for weeks, not days. Volume ramps slowly. There is no shortcut, and anyone offering one is selling you a burnt domain.
- Test continuously. Placement is checked on an ongoing basis, because a domain that landed last week can stop landing this week.
4. Research is the message
The gap between a 1% reply rate and a 5% reply rate is almost never the subject line. It's whether the first sentence proves a human looked at this specific company before writing.
"I saw you're in field service" is a merge field wearing a disguise, and every operations manager has learned to delete it on sight. "You posted for a second dispatch coordinator in Leeds three weeks ago" is a different thing entirely — it's evidence, and evidence earns the next fifteen seconds.
The research isn't preparation for the message. The research is the message.
This is also the honest reason we cap client count. Research at this depth doesn't scale by adding accounts; it scales by getting better at finding the signal, which is slow.
5. Five numbers make it arithmetic
Once these are stable, forecasting stops being an argument:
| Number | What it tells you | Typical range |
|---|---|---|
| Contacts reached | The only true input dial | ~1,000/mo |
| Reply rate | Whether targeting and copy are right | 3–6% |
| Positive rate | Whether you're talking to the right person | 1–2% |
| Meetings booked | Whether the pitch converts interest | 6–11 |
| Meetings attended | The number worth paying for | 5–9 |
Each number has a distinct failure mode, which is what makes them useful. Low reply rate is a targeting or research problem. Good reply rate but few positives means you're reaching the wrong seniority. Plenty of positives but few bookings is a pitch problem. Bookings that don't attend is a qualification problem. You can't diagnose any of that from one aggregate number.
What we deliberately don't do
- Paid media. Different craft, different feedback loop. We don't have a media-buying track record and won't spend your budget pretending otherwise.
- Sales coaching. Neither of us has closed enterprise deals in a Western market. We'll book the meeting; your team should run it.
- Volume blasting. Cheap to do, and it damages the domain, the brand and eventually the market.
- Scraped or purchased lists. Bad data, legal exposure, and it defeats the entire method.
The honest summary: this is not a clever growth hack. It's a well-run version of something most companies attempt badly, applied to a buyer who still reads their email. The advantage is in the execution, not the idea.