When it broke
In 2022, 2023 and 2024, I could tell you almost exactly what it cost us to book one meeting: about 97 contacts emailed. Ninety-seven well-chosen people, a solid pitch, a decent offer. One conversation out the other end. For years that math worked, so we ran it.
Then, towards the end of 2024, it stopped.
I will not dress this up. In the months that followed, the number got worse and then it broke. Inboxes tightened, filters got smarter, and the same email that used to earn a reply now earned silence. For a few weeks, even months, it felt like the floor had dropped out of a channel we had built the whole business on. If you run outbound for a living, you know the specific kind of quiet I mean.
We had two options. Send more, or send better. The well-known international players in our space, the ones with a real reputation, were all reaching for "more": more volume, more domains, more automation. We went the other way.
From 97 to 33
Today, in mid-2026, it takes us around 33 contacts to book one meeting, not 97. And here is the part that convinced even me: that number barely moves. We see roughly the same 33, in a tight band, across different campaigns, different clients, different target industries, and completely different buyer personas. On one side, IT security leaders and CIOs at banks in Austria. On the other, heads of production and plant managers at manufacturers in Germany. Two worlds that share almost nothing land in the same place. When a result holds across that much variation, it is not luck. It is method.
One honest caveat, before this turns into a vanity metric. We watch 97-to-33 internally, because it tells us the method is working. But we rarely show it to a client, and we never build their reporting around it. It is not the point. The point is two things: do we generate real opportunities they can convert into business, and do we protect their brand while we do it. In a finite market, a reputation burned on careless outreach does not come back. The contact ratio is just how we know we are getting those two right, with less waste. If it ever came down to a better ratio or a client's good name, the ratio loses every time.
What actually changed
We research before we reach out, not after. The old model treated research as a tax on volume. The new one treats it as the work. Before we reach out to anyone, we know the account, the trigger, and the reason this person should care this quarter. This is where we use AI massively, and it is worth being precise about how. Not to save time, and not to write our messages, but to raise the quality of the research by an order of magnitude. More accounts understood properly, more signals caught, more real reasons to care than any human skimming a list could ever find. That sounds slow. It is the opposite. It means we stop spending contacts on people who were never going to answer.
We filter hard, and we are proud of how many we cut. Most outbound is built to keep lists big. Ours is built to make them smaller. On a typical list, we cut 8x more than we keep. We disqualify aggressively: wrong or unclear role, no or unclear ICP fit, negative signals, and so on. Every name we remove is a name we do not waste a touch on. In a finite market, the cut list matters as much as the contact list.
We judge the hook and the angle like it is the product. Because it is. We stopped asking "what can we say" and started asking "what is true for this account right now, and does it earn 30 seconds of attention." To be clear, we are not against automation. We use it wherever it genuinely helps: surfacing signals, enriching accounts, keeping four channels in sync. But we draw a hard line in one place. The AI finds the hook; it does not write the message. Every message is written by a human, not generated by AI. That is a deliberate choice, not a limitation. The judgment and the words stay human. Most messages do not survive that test. The ones that do, work.
We run four channels at once, not one channel harder. Email is now one instrument, not the orchestra. Around the same account we run LinkedIn Aircover so we are familiar before we ever land in an inbox, LinkedIn DMs for a human-to-human angle, and the phone when the moment is right. No single channel carries the whole load, so no single channel's bad week sinks the campaign. This is the core of how TrueReach works now, and it is the part I would not give back.
The side effects we did not plan for
There is a quiet side effect to sending less. Volume is what wrecks deliverability: the more you blast, the more spam filters and mailbox providers distrust you, and the more of your email quietly never reaches an inbox at all. By sending far less, to far better-qualified people, our sender reputation recovered. The emails that do go out are now far more likely to actually land. Restraint compounds.
And here is what the contact count alone does not show: the meetings themselves got better, not only easier to come by. In the old model, every so often a friendly independent consultant slipped through. Pleasant conversation, maybe a door opener, but never a buyer. That does not happen anymore. By the time we are in a room, the company is qualified, the person is qualified, and we already know what they are wrestling with before the meeting starts. We are not finding out whether there is a fit during the call. We are confirming one we had good reason to expect. And the team feels it: far less time lost to meetings that were never going to lead anywhere.
Why it fits our market
Here is the thing I did not expect: this rebuild did not just fix our numbers. It fit our market better than the old model ever did.
We sell into a finite TAM. The companies worth talking to are a known, countable set, not an endless lake to spray. The industry has a name for orienting everything around that set: account-based sales. For us it stopped being a phrase on a slide and became the actual operating model. The old "97 contacts per meeting" logic quietly assumed an almost infinite supply of names to burn. That assumption was always wrong for us. Burning through a finite market with mediocre volume is how you run out of market. Doing real work on each account, across four channels, with most names filtered out before they ever get touched, is what account-based sales looks like when you actually mean it.
And there is a second fit, one that is cultural. We sell into the DACH market, and DACH buyers have famously little patience for spray-and-pray. They notice immediately when an approach is generic, and they quietly write you off for it. A well-researched, relevant, low-volume approach is not just more efficient here. It is closer to how serious business gets done in this region in the first place: do your homework, show up with substance, respect the other person's time. The method the collapse forced on us turned out to be the one this market quietly respected all along.
So the collapse at the end of 2024, as bad as it felt in the moment, pushed us toward a better model at exactly the right time. I used to think we should have been running it all along. The honest answer is that we could not have. Research at this depth, on every single account, was not affordable before AI made it possible. The disruption and the thing that answers it arrived together. We did not survive the collapse by going back to what worked before. We survived it by building something better, and the numbers have held across very different clients and sectors.
If you are selling into DACH
If you are running B2B outbound into the German-speaking market and staring at the same quiet inbox we were, I will tell you what I wish someone had told me then: the answer is not more sends. It is fewer, better-aimed ones across more than one channel. Smaller list, sharper reasons, four ways in. Especially here, in a market that rewards the homework and quietly punishes the blast.
We came out of this prouder of the engine than we have ever been, and more optimistic about what comes next. The market did not get easier. We just got a lot better at it.
Martin Weiss, Co-Founder, BizXpand