The thesis is that finding customers becomes predictable if you split it into roles — whoever opens conversations, whoever closes, whoever looks after the account afterwards — and if you count backwards from the target to the number of conversations required. Around that sit email templates and an enthusiastic defence of the method.
What I took from it
Splitting roles, translated to one person, becomes splitting days. Prospecting and delivery do not mix, because they compete: delivery always looks more urgent and eats the prospecting time, and three months later the pipeline is empty and it is too late to fix. Since reading it, finding customers has fixed blocks in the week and gets treated as production, not as something done when a gap appears.
The second idea worth keeping is the backwards arithmetic. If you know how many conversations produce a proposal and how many proposals produce a job, the target stops being a wish and becomes a number of conversations per week, which is something actually under your control.
The third is subtler: the first message should not sell anything, it should ask who the right person is. That still works because it asks for almost nothing, and nobody minds being asked for a redirection.
What I left out
Almost the entire method. It was written to sell subscription software in 2011, before every manager's inbox became a landfill; the templates this book popularised are literally the format that now gets recognised and deleted in two seconds. And the role specialisation assumes a team and a volume most of its readers do not have. It is a historical book: it explains where the thing that annoys all of us came from.