Use this sequence when your agency sells one nameable service (paid search, lifecycle email, brand identity, technical SEO) and you can say something specific about how each target company does it today. It is a research-first sequence. The list is short, the observation is real, and the ask grows smaller with each step.
Who to put in it
Heads of marketing, marketing directors, and in companies under a hundred people the founder who runs marketing on the side. One person per company, or two if the service sits between marketing and another function such as product or sales. Pick companies in one sector per campaign so the sentence about who you fix this for is true. Twenty to forty companies per wave is realistic, because each observation takes time to write.
What to change before sending
observation is a custom field and every row needs one, written in a single sentence a stranger could verify: a landing page that loads slowly on a phone, a newsletter that has not changed format in a year, ad copy that contradicts the website. The personalisation article covers how custom fields flow into the sequence and how to catch a blank in the To check queue before it sends. If you want to test two openers, use A/B variants on the first step only and read the comparison table after the wave; the method is in A/B testing cold email subject lines and variants.
Name a first_step you genuinely sell at a fixed scope. If everything you do is a retainer, rewrite step two honestly rather than inventing an audit.
How to read the replies and when to stop
“Send me the proposal” after step three is the yes this sequence is built for. Send it as a tracked offer link, following the golden rules of sending offers and proposals, and move the company into the proposal follow-up sequence. “We have an agency” is a Maybe later; ask when the contract renews and tag the company. “How did you find that?” is a genuine question about your observation and often turns into a conversation; answer plainly.
Do not measure this campaign by reply rate alone; a handful of conversations with the right companies is the goal, and reply rate is the wrong north star explains why. Stop after step four, and revisit silent companies only when you have a new observation to make.