CTA architecture — heavy vs light, and the per-touch progression.

The premise

Every cold message has a body and an ask. The body earns attention through specificity, recipient-orientation, and a defensible value claim (Chapters 1 through 4). The ask converts attention into a measurable action. These two halves are not symmetrically weighted in their contribution to conversion: a mediocre body with a well-calibrated light ask routinely outperforms an excellent body with a heavy ask at the wrong touch.

The asymmetry is mechanical. The body is read in 8 to 15 seconds; the ask is read in under 2. The cost of the wrong ask is not that the recipient declines — it is that the recipient does not respond at all, because declining a heavy ask requires its own composition cost. A heavy ask at touch 1 produces, in our observation, a 3-5x lower reply rate than the same body with a single-question light ask, holding everything else constant.

The CTA-weight spectrum

Cold-message CTAs sit on a four-tier weight spectrum, from the highest commitment cost to the lowest:

• Heavy."Book a 30-minute meeting next week — here is my calendar." Full-commitment ask requiring calendar time, declared interest, and acceptance of the implicit framing that the pitch is worth a meeting before any value has been demonstrated.
• Medium."Would a 15-minute call next week make sense to walk through this?" A conditional meeting ask that acknowledges the recipient's right to decide whether it makes sense.
• Light."Is this a priority for you this quarter?" A single question requiring a one-sentence reply, no calendar commitment, no implicit framing about what comes next.
• Micro."Worth a 5-minute look?" or "Open to a quick yes or no on this?" A request for a one-word binary reply that costs the recipient seconds to send.

Each tier trades commitment for engagement, and the empirical curve is monotonic in the direction operators consistently underestimate.

The empirical conversion-rate differential

At first touch, holding the body constant and varying only the ask, the observed reply-rate ranges across well-targeted B2B sequences are:

The differential is not subtle. A first-touch heavy CTA at the median of its range produces roughly one tenth the reply volume of a first-touch micro CTA. The argument for heavy CTAs — higher quality per reply — does not survive the conversion math: the heavy-CTA reply is more likely to convert to a meeting, but cumulative meeting volume from a light or micro CTA run through a qualifying second touch is consistently 2 to 4x higher per thousand recipients sent.

The CTA-weight progression across the sequence

The correct CTA architecture is not a single weight applied to every touch — it is a progression that escalates as the recipient's implicit engagement accumulates. A recipient who has opened touch 3 without replying is a different audience from one seeing the sender for the first time at touch 1, and the ask should reflect it. The progression that, in our observation, produces the highest cumulative reply rate across a 7-touch sequence:

The progression respects two empirical facts. First: tolerance for heavier asks accumulates with implicit engagement — open, click, partial reply, dwell time. A heavy ask at touch 6 to a recipient who has opened five previous messages is not, mechanically, the same ask as a heavy ask at touch 1. Second: the implicit-rejection threshold sits around touch 8. Sending past touch 8 to a non-responder does not increase reply rate and does increase the probability of a recipient-driven spam-classification signal, which the sending domain absorbs as reputation damage.

The single-question CTA pattern

The dominant pattern at touches 1 through 3 is the single question requiring a one-sentence reply. Canonical shape: "Is improving [specific outcome] a priority for [recipient's team] this quarter?" Or: "Are you the team that owns [specific function] now, or has it moved elsewhere?" The question is anchored to a specific outcome the recipient cares about, and the answer is short enough that the cost of replying is lower than the cost of drafting a refusal.

The empirical reply rate on a well-targeted single-question CTA is 4 to 7%. The qualifying value is twofold. A "yes" identifies a recipient with declared priority, who moves to a medium CTA next. A "no" produces clean disqualification — a recipient removed from the sequence rather than continued through touches 2 through 7 to no economic effect.

The calendar-link-at-first-touch failure

The most common heavy-CTA failure: the operator drops a calendar link into touch 1 without first establishing that a meeting is what the recipient should want. The implicit framing — that the recipient and sender already have a meeting-shaped relationship requiring only a time slot — is, in a cold context, false. The empirical reply rate on a first-touch message containing a calendar link is roughly half of the same message without one, and roughly one quarter of the same message with a light CTA instead.

The drop is mechanical: the calendar link signals that the sender's goal is the meeting, not the recipient's problem. The recipient's mental model collapses to "another salesperson optimizing for the booked call." A calendar link at touch 6 or 7, after a light-CTA progression has produced no reply, is a different message — its appearance there, after deliberate restraint, reads as escalation rather than presumption.

The "are you the right person" opener problem

A widely-deployed light-CTA pattern: "Are you the right person to speak with about [function] at [company]?" Structurally this is a light CTA — a single question requiring a one-word reply — and operators argue its routing function makes it efficient at scale.

In our observation, the "are you the right person" opener underperforms a value-anchored single-question CTA by a factor of two to three on reply rate. The failure mode is mechanical: the message forces the recipient to do the sender's routing work without delivering any reason to engage. The implicit response is "why is this my problem to solve." The pattern reads as a thinly-disguised lead-qualification step, and recipients have learned to ignore it. The value-anchored variant — "is [specific outcome] something your team is working on now" — solves the same routing problem (a "no" tells the sender to look elsewhere) while delivering an actual reason to respond.

The yes-no CTA

The highest-converting single-CTA pattern is the explicit yes-or-no ask: "Worth a quick yes or no on whether this is a fit?" The pattern works because it explicitly grants permission to reject, and the cost of a one-word "no" is lower than the cost of ignoring the message and managing the lingering uncertainty. The qualifying value is exceptionally clean: a "yes" is a meeting candidate that moves directly to a medium CTA; a "no" is definitive disqualification. The pattern produces the highest-quality disqualification signal on the spectrum.

The "soft pass" CTA

Adding "no worries if not the right time" — or a near variant — to a light or medium CTA produces a measurable lift on reply rate. The mechanism is psychological: recipients respond more frequently when they feel they have an explicit out. The empirical lift is 15 to 35% over the same CTA without the soft pass, concentrated in the "maybe later" segment — recipients not actively engaged but not closed-off, who would otherwise have produced no signal. The signal they produce when given an out is often a soft yes with a follow-up suggestion ("not this quarter, but check back in Q4") — a higher-value signal than silence.

The "explicit out" CTA

A more aggressive variant: offering a single-click removal option explicitly within the message body. "If this is not relevant, a one-word reply with 'remove' takes you off the list — no friction." The pattern feels counterintuitive — operators worry it cannibalizes engaged recipients — but the empirical effect on net response rate is positive. The mechanism: recipients who exercise the explicit out are recipients who would otherwise have flagged the message as spam. A "remove" reply removes the contact and produces no reputation signal; a spam flag propagates to the inbox provider's reputation system and degrades deliverability for every other recipient. The explicit out converts a damaging spam signal into a benign list-management action.

Reply-to-thread vs calendar-link CTA

Among recipients who do convert to a meeting, the operational question is the conversion path: reply-to-thread (the recipient replies, the sender proposes times, the meeting is scheduled in the thread) versus calendar-link (the recipient clicks, the meeting is self-served). The empirical differential favors reply-to-thread on first-meeting conversion by 30 to 60%. The recipient who has just engaged with a cold message is not yet in a state of "I will commit calendar time to this sender" — they are in a state of "I am willing to continue the conversation." The reply-to-thread path matches that state; the calendar link assumes a state not yet reached. The operational cost of reply-to-thread is non-zero, but the lift is consistent enough that calendar-link-only CTAs should be a deliberate choice, not a default.

CTA placement

Placement within the message produces a measurable effect on response rate, independent of weight:

• End-of-message. The default. The body builds the value claim, then asks. Effective when the body is short enough that the recipient reaches the end.
• Mid-message.The CTA is embedded after a one-sentence value claim. Underperforms end-of-message, because the recipient's implicit checklist ("why am I being asked, and is the answer compelling") is not yet complete when the ask arrives.
• PS-placed. The CTA appears after the signature as a postscript. Observably effective — PS placement produces a reply-rate lift of 10 to 25% over end-of-message on otherwise identical messages, driven by scanning behavior. The PS is the second-most-read element after the subject line, and a PS-placed CTA gets seen even by recipients who skimmed the body.

The "next steps if interested" framing

A specific augmentation of the medium and heavy CTA: explicitly articulating what happens after a positive reply. "If a 15-minute call makes sense, I'll send three time options and a one-page brief ahead of the call." The pattern lifts response rate on medium and heavy CTAs by reducing the recipient's uncertainty about what they are agreeing to. It does not lift — and in fact slightly suppresses — response rate on light or micro CTAs. The light CTA succeeds because its commitment cost is low; articulating downstream steps raises the implied commitment and reintroduces the friction the light CTA was structured to avoid. Correct for touches 4 through 7, wrong for touches 1 through 3.

The CTA at touch 7 — the last touch

The final touch operates under different conversion math. By touch 7, the recipient has either engaged enough to have replied or implicitly rejected by sustained non-response. The two dominant last-touch architectures, both producing observed reply rates of 0.5 to 1.5%:

• The explicit close."I'll assume timing isn't right and stop reaching out — happy to circle back next quarter if that's better." The pattern leverages mild loss aversion: the sender is closing the loop, and a non-zero fraction of non-responders surface a "actually, let's talk" reply at exactly this moment.
• The pivot ask.A reframe to a different question — "is there someone else on the team I should be reaching out to instead" — which produces a routing reply from recipients who would not have replied to the original ask. Structurally a light CTA on a previously-non-responsive contact, and it recovers a meaningful tail of the cohort.

Operators sometimes attempt a heavy CTA at touch 7 (the "last chance to book" pattern). Reply rate is observably worse than either architecture above, because the heavy ask amplifies the recipient's existing rejection signal rather than reframing it.

Common operator failures

• Heavy CTA at first touch. Calendar link in touch 1. Reply-rate drop of 50-75% versus a light first-touch CTA. The most expensive single mistake on the spectrum.
• No CTA at all. The message trails off after the value claim. The recipient defaults to no action. Reply rate is roughly half of the same message with a light CTA.
• Multi-step CTA in one message."Let me know if this resonates, and if so here's my calendar — also happy to send a deck first if useful." Three parallel commitments to evaluate; cognitive load suppresses response. Single CTA per message, always.
• Generic "let me know" close. A non-CTA masquerading as a CTA. Produces the no-CTA floor while consuming the structural space a real CTA would occupy.
• Identical CTA across all 7 touches. The recipient reads the sequence as a single rejection event repeated. Collapsing the progression removes the only mechanism by which a sequence outperforms a single touch.

Pre-write CTA checklist

• Touch number identified; CTA weight matches the position in the progression
• Single ask, not multiple — one action, one decision point
• Light or micro at touches 1-3; medium at 4-5; heavy or last-touch at 6-7
• Value anchor present — the ask is tied to a recipient outcome, not generic routing
• Soft pass present on medium and heavy CTAs — explicit out granted
• Reply-to-thread default for first-meeting conversion; calendar link only at touches 6+
• PS placement considered for end-of-message CTAs — measurable lift at zero cost
• Last-touch architecture decided in advance — explicit close or pivot, not heavy-CTA recycle
• Generic "let me know" closes removed

Where this fits

CTA architecture is the inflection point of the message components in Chapters 1 through 5. The principles, the subject line, the opening line, and the value proposition exist to earn the attention the CTA converts. A disciplined CTA — calibrated to the touch position, anchored to a recipient outcome, single-ask, with the right placement — extracts the maximum reply rate from a body the earlier chapters have already optimized.

Personalization at scale (Chapter 6) addresses how the CTA and body get tuned without collapsing operator economics. Multi-touch sequencing (Chapter 7) addresses the per-touch frame progression that pairs with the CTA progression here. A sequence with the right CTA progression but the wrong body progression produces the same dilution as the inverse — the two are designed together.