Why You Keep Losing the Ones That Mattered
The move in short
Custom Build — A pipeline triggered when a deal is moved to Closed-Lost in the CRM that pulls the original proposal PDF, the full email thread with the prospect, any recorded call notes, and the stated rejection reason, runs an LLM analysis to extract the real objection pattern, price sensitivity signal, and the point in the conversation where momentum shifted, and accumulates these into a rolling loss-pattern register — updated each time a deal is lost — with a quarterly digest delivered to the managing director showing recurring themes, the competitor named most often, and one suggested change to the pitch process.
The Company
Hantwerk Digital Studios is an independent brand and digital design consultancy based in Copenhagen. Around 28 people, mix of brand strategists, UX designers, and a small production team. They work mostly with mid-market companies — the kind of clients who are too big for a freelancer but not big enough to go straight to a global agency. Their work is genuinely good, and they've got a real reputation in the Danish market, with some recognition beyond it.
The Pain
Sofie, the managing director, knows they win the small stuff and the big stuff. It's the middle tier that's the problem. A 200K–400K branding project goes to pitch, the team puts serious time into it, and then they hear nothing, or they get a polite email saying they've gone in a different direction. Sofie's gut says it's sometimes pricing, sometimes something about how they present the strategic side, maybe occasionally a competitor she keeps seeing pop up. But she doesn't actually know. There's no record. Each loss gets absorbed into the next week, and the team moves on. She's been running on instinct for three years, and it's a reasonable instinct, but she can't do anything concrete with it.
The Move
When a deal is moved to Closed-Lost in the CRM, a pipeline kicks off automatically. It pulls the original proposal PDF, the email thread with the prospect, any call notes logged, and whatever rejection reason was entered. An LLM works through all of it and looks for a few specific things: what the real objection seems to be (not just what was written in the CRM field), whether anything in the conversation suggests price was a problem, and roughly where things started to go quiet or turn.
All of that feeds into a running register that builds up over time. Every lost deal adds to it. Once a quarter, Sofie gets a short digest: what objection patterns are coming up repeatedly, which competitor is being named, and one concrete suggestion about the pitch process based on what the data is actually showing.
It's built in Make or n8n, connected to whatever CRM they're using, with the LLM step handled through the API. It takes a few days to set up properly.
The blind spot
The rejection emails and the CRM notes already exist. All of this information is sitting there. The thing that doesn't happen is anyone reading across all of it together, because in the middle of running the studio, nobody has the time to go back through six months of lost deals and look for a pattern. So they don't, and the losses stay separate rather than adding up to something useful.
The pattern
The same basic move works in a few other situations:
- A law firm that loses tenders and wants to know whether it's fee structure, team composition, or something in how they write the submissions
- A SaaS company reviewing churned accounts to find out what the early warning signs actually were, not what the team assumed they were
- A recruitment agency tracking why placed candidates leave within 12 months, and whether there's a pattern by client or role type