Cold outreach cadence playbook D1 D3 D5 D7 D9 D11.

Six touches across 11 days. Two-day gaps. Each touch carries a different angle, never a copy paste of the prior message. The cadence runs across email primary, LinkedIn secondary, Twitter tertiary on warm prospects.

The open-rate floor is 45 percent by touch four. The reply-rate floor is 3 percent by touch six. Drop below either floor and the sequence pauses while the list and deliverability layer get rebuilt. The cadence is paired with the founder funnel service.

1. Pick one ICP slice and one trigger signal. A cadence built for three personas at once averages out to a message none of them answer.
2. Validate the list before writing a word. Only the rows that survive firmographics, signal recency and deliverability checks enter the sequence.
3. Write six angles, not six versions of one email. Signal, pivot, proof, bump, cross-channel, breakup.
4. Fix the spacing at two days and the length at 11 days, then leave both alone for the first 200 sends so the numbers mean something.
5. Set the two floors up front: 45 percent open by touch four, 3 percent reply by touch six. A miss sends you back to the list and the inbox setup, never to more volume.

The D1 / D3 / D5 / D7 / D9 / D11 spacing exists because the prospect inbox compresses recall on a 48-hour window. A 24-hour cadence reads as automation; a 5-day gap loses the thread. Two-day gaps land inside the recall window without tripping the spam pattern detectors that fire on aggressive frequency.

Six touches captures the long tail. On a validated list, 38 percent of replies arrive between touch one and touch three. The remaining 62 percent arrive between touch four and touch six. Cutting the cadence at three touches throws away most of the qualified pipeline.

Each touch carries a distinct angle. Touch one opens on a signal observed in the last 14 days. Touch two pivots to a different angle on the same problem. Touch three drops a proof. Touch four bumps for engagement. Touch five crosses channels. Touch six closes the loop. No touch repeats the prior angle; if it does, the sequence underperforms by 30 to 50 percent on reply rate.

One signal observed on the prospect in the last 14 days. A recent funding round, a new model release, a conference talk, a blog post, a hiring signal. One sentence on the signal, one sentence on the angle, one sentence on the ask. Subject line under 6 words.

Open-rate target on touch 1 alone is 35 to 42 percent on a validated list. Reply-rate target is 1.5 to 2.5 percent. Below those numbers the signal layer is wrong; the list goes back through the validate-prospects gate.

Same problem, different angle. If touch 1 led with the signal, touch 2 leads with the operator-pattern lift. If touch 1 led with a question, touch 2 leads with a proof. The body length is shorter than touch 1, never longer. Subject line is fresh, never RE: the touch 1 thread.

Touch 2 reply rate stacks with touch 1. Combined reply rate target by D3 is 2.5 to 4 percent on the engaged subset (opens of touch 1).

One operator-pattern lift, anonymized. The format is fixed: one-line setup, one-line outcome, one-line ask. The proof chosen matches the prospect cohort observed in the validate gate, not a generic case study. Touch 3 carries the highest single-touch reply rate in the sequence (target 2 to 3 percent standalone).

One sentence reply on the touch 1 thread. Not a forward, not a new thread. The body is shorter than 25 words. The phrase structure is a question, never a statement. Touch 4 is the only touch in the sequence that uses the original thread.

Open-rate target by D7 is 45 percent (the cadence floor). Below that floor the deliverability layer is broken; the cadence pauses for an inbox health check.

LinkedIn or Twitter, never email. A connection request with a note that references the touch 3 proof. Note length under 200 characters. The cross-channel ping signals the cadence is not blast email and lifts the touch 6 reply rate by 1.5 to 2x.

One sentence. Closing the loop. No ask. The structure is a single line that gives the prospect permission to opt out, with an offer to ping again in 90 days. Touch 6 reply rate target is 1.5 to 2.5 percent, weighted heavily toward yes-defer-to-Q-next replies that re-enter the cadence on a 90-day cycle.

The emails that get answered read like one person writing to one person. Open on the signal you actually observed, say in one line why it made you write, and ask for something small with a date on it. A reply is the goal, not a booked call; the call comes from the reply.

Cut every sentence the prospect could have written about themselves. Replace a claim with a number the prospect can check, and keep the proof to one anonymized operator result that matches their cohort. Read the email out loud before it ships: if it sounds like a template, it will be read as one.

ReachInbox or Smartlead for the sequencer (warmup gates wired in). One inbox per persona, capped at 35 sends per inbox per day. Domain warmup ages new sender domains for 21 days before the first touch ships.

Apollo + Clay + Apify for list build. Validate-prospects gate kills any row that fails firmographics, signal-recency, or deliverability checks. List size after the validate gate typically lands at 40 to 55 percent of the input list.

HubSpot or Attio for the CRM. PostHog for the activity dashboard. The audit ledger reads from PostHog and ships every Monday with open rate, reply rate, qualified-meeting count, and the abandon-rule status.

Cold email never ships from the company's main domain. We register two to three lookalike sender domains per campaign, point each one at the main site, and set SPF, DKIM and DMARC before the first warmup email goes out. Each domain carries two or three inboxes, and every inbox is capped at 35 sends per day.

New sender domains age for 21 days on warmup before touch one. If a domain's placement drops, it comes out of rotation and the others absorb its share at the same per-inbox cap, so one burned domain never drags the main domain's reputation with it.

Every row is re-verified within 7 days of its first touch. Catch-all domains, role inboxes (info@, sales@) and anything the verifier marks risky are dropped, not sent at a lower volume. After the validate gate the list usually lands at 40 to 55 percent of what went in, and that smaller list is the one that clears the 45 percent open floor.

Hard bounces above 2 percent on any send day stop that inbox for the day. Replies that say "not me" route the next touch to the named colleague rather than retrying the same contact, and every unsubscribe or opt-out lands on a shared suppression list that all future campaigns read first.

Eleven days, six touches, then stop. That window is long enough to reach the touches where most replies land, since 62 percent of replies on a validated list arrive between touch four and touch six, and short enough that the account is out of the sequence before it starts to read as pressure. Running past D11 rarely adds replies; it mostly adds unsubscribes and spam complaints.

Shorten it only when the list is thin on signal. If the D5 open rate is already under 35 percent, the problem is deliverability or the list itself, and more touches will not fix it. Lengthen it by spacing, never by adding touches: a prospect who answers the breakup with a not-now re-enters on a 90-day cycle with a fresh signal, rather than getting touch seven on D13.

Across channels the clock is shared. The D9 LinkedIn or Twitter ping counts as one of the six touches, not a parallel sequence, so the prospect sees one conversation across channels instead of two campaigns running at once.

Every Monday we compare the week's open rate, reply rate and qualified meetings against the floors. We change one variable per cycle, the subject line, the touch three proof, or the signal source, and hold it for at least 200 sends before reading the result. Changing three things at once tells you nothing about which one moved the number.

Winning variants roll into the default template for the next campaign; losing ones are logged with their numbers so nobody re-tests them. Replies are tagged by reason (timing, budget, wrong person, interested) and the most common objection becomes next cycle's touch two angle.

Three abandon conditions. The cadence pauses on any one of them and the list goes back through the validate-prospects gate before the next attempt.

1. Open rate under 35 percent at D5: deliverability or list problem.
2. Spam complaint rate over 0.3 percent at any touch: pause sender domain.
3. Reply rate under 1 percent at D11: ICP fit problem, rebuild signal layer.

Pair the cadence with the founder funnel engagement when the abandon rule fires more than once per quarter; the failure is upstream of the cadence and lives in positioning, ICP, or product.

FORKOFF runs the cadence for you. The sandbox covers the first 14 days, with the 35 percent open-rate floor checked at D5. After the sandbox, the engagement moves to a 90-day retainer at outcome-priced rates with the weekly report shipping every Monday. Sandbox sized by application.

Sister playbook: founder-led growth covers the inbound side of the same buyer journey. Pair them for the full pipeline. Internal tool: CPQV calculator for the measurement layer.