First Touches That Earn Replies

A structure that works

Most effective first touches have four parts, and each fits in a sentence or two.

The reason. Why you are writing to this person now: the trigger you found in research (Lesson 3.1), a mutual connection, or something they said publicly. If you cannot state the reason in one line, you are not ready to write.

The relevant point. One observation or insight about their situation, ideally the kind of commercial insight Lesson 2.2 described: something true about their world that they may not have connected to their own position. Not a list of your features.

The credibility. One line of proof: a similar company you helped and the result, stated plainly and only if you can evidence it.

The ask. One question that is easy to answer. 'Is this on your list for this half?' is easier to answer than 'Do you have 30 minutes on Tuesday?', and an answer of any kind starts a conversation.

What to leave out is as important: no paragraph about your company, no three calls to action, no attachment, no fake 'Re:' in the subject, no pretending you have met.

What the data says about length, language and questions

The most detailed public dataset on email replies is old, but it is primary and specific. In 2016 Boomerang, which makes an email add-on, analyzed over 40 million emails sent by its users where a reminder had been requested if no reply came. Messages of 50 to 125 words had the best response rates, each above 50% in that sample, with rates declining slowly for longer emails and more steeply for very short ones. Messages written at a third-grade reading level performed best. Emails asking one to three questions were about 50% more likely to get a response than emails asking none. Subject lines of three to four words had the highest response rates.

Two cautions. This is 2016 data from people emailing contacts they chose to follow up with, most of them presumably not cold prospects, so the absolute response rates are far higher than any cold campaign will see; Instantly's 2026 benchmark, by contrast, puts the platform average for cold email at 3.43%. And it is correlational. The direction of the findings, shorter, plainer, with a real question, is consistent with common sense and with later practice, which is why they are still worth knowing. The exact numbers are not targets.

The first message also carries extra weight in a sequence. Belkins' 2026 study of 7.5 million emails sent in 2025 found the first email had the highest reply rate of any single step, even though follow-ups together produced more replies. Your opener is the message most likely to be read and answered; spend your effort there.

What marks a message as mass-produced

Buyers have become very good at spotting automation, and AI drafting has produced its own tells. Generic flattery ('I was impressed by your company's growth'). Personalization that is true of everyone ('As a leader in your industry'). Overlong, perfectly balanced paragraphs. Three bullet points of benefits. A question that cannot be answered in a sentence.

The cure is not to hide the automation; it is to make the content specific. One verified fact about the person, one insight relevant to their situation, one plain question. Gartner's 2025 survey adds a reason to keep it accurate as well as specific: 69% of buyers reported inconsistencies between suppliers' websites and what sellers told them. A first touch that promises something your website does not is a poor start.

The Challenger research summarized in Lesson 2.2 framed the seller's job as teaching, tailoring and taking control. In a first touch, tailoring is the whole game: the reader should be able to tell that this message could not have been sent, unchanged, to anyone else.

Learning from your own campaigns in Kavanah

Outside averages are a starting point; your own buyers are the real test. Kavanah keeps the evidence from your past cadences and lets you query it.

get_winning_openers returns personalized opening sentences from past cadence sends that earned a reply, meetings first, excluding contacts who later opted out, together with the reply rate of personalized versus generic sends. It is a quick way to see what kind of opening your buyers actually respond to, and whether personalization is paying for itself.

suggest_cadence_variants gathers evidence for writing subject lines and choosing sender names for a new cadence: past automatic winner decisions, first-touch subject arms ranked by reply rate with their contact counts, and which subject traits went with higher reply rates. Treat it as evidence to reason from, not a generator of copy.

When a step is set to personalize, the opening line is generated per contact from what the workspace knows, and it fails closed: if there is not enough real context to open honestly, it does not fall back to a generic body under a personalized pretext. For a one-off message, propose_follow_up drafts an email to a deal contact and can save it as a Gmail draft for you to edit and send yourself.