A cadence is the sequence of touches a rep sends to one prospect across channels and days to earn a conversation. The best cadences respect the buyer, pair every channel to the right moment, and give the rep a defendable playbook rather than a nagging checklist. This guide walks you through the full build: picking channels for your segment, sequencing touches by day and intent, writing the three message types that carry the load, and instrumenting the metrics that tell you which steps to keep and which to kill.
Time: 75 minutes
The right channel mix is determined by where your buyer already pays attention, not by what your engagement platform makes easy. Start by asking three recent closed-won customers how they first heard from you and which touch finally earned the meeting. You will almost always learn that one or two channels did the real work and the rest were noise. For a VP of Sales buying a CRM, phone and LinkedIn carry more weight than email alone. For a security engineer, email with a specific technical hook outperforms a cold call every time. Pick three channels for your first cadence, not five. Three gives you enough surface to be noticed without diluting your effort or looking automated. Email handles the asynchronous explain-and-prove work. Phone handles the live-objection and trust-building work. LinkedIn handles the social-proof and warm-touch work. SMS belongs only after a prospect has engaged somewhere else; cold SMS to a stranger reads as a scam, hurts your sender reputation, and runs into regulatory risk in many jurisdictions. Lock the channel set before you write a single word.
A cadence has a fixed length and a defined shape. The length is the number of business days from first touch to final touch. The shape is how touches are distributed across those days. For outbound to cold prospects, twelve to sixteen business days with eight to twelve total touches is the sweet spot for most B2B motions. Shorter than that and you are quitting before most buyers notice you exist. Longer than that and you are chasing people who will never buy, which costs you the opportunity to call someone who will. Load the first week heavier than the last. A reasonable shape puts three touches in days one through three, two touches in days four through seven, and the remaining four to six touches spaced out through days eight to sixteen. The front-loading is deliberate: the first touch rarely gets noticed, and the second and third touches are what convert a cold contact into a response. Avoid the mistake of spreading touches evenly across the window. Even spacing produces fewer meetings because no single week ever reaches the intensity needed to break through inbox noise.
Each touch should match the intent that fits the day of the sequence. Day one is a context-setting touch, usually email, because you are introducing a reason for the conversation. Day two is a validation touch, often LinkedIn, because you want the prospect to see that you are a real person tied to a real company before you call. Day three is the first live attempt, phone, because you have earned enough visibility that a dial will not feel like a cold-start. Day five is a reference or proof email, where you share a short specific result from a similar company. Day seven is a second phone attempt with a different ask than the first. Day ten is a value-add email, where you send a resource, a benchmark, or an insight tied to their role, with no meeting request attached. Day thirteen is a direct-ask email, short and plain, asking for a yes or a no. Day sixteen is the breakup, a one-line note that signals you will stop reaching out unless the prospect wants to engage. The sequence of channels matters because each one earns the right to use the next. A phone call on day one from a stranger is friction. A phone call on day three from someone whose name has appeared twice already is a conversation.
Most cadence copy fails because reps write every message the same way: a pitch about the product. The messages that actually book meetings fall into three categories, and every step in a healthy cadence maps to one. The first type is the context message, which names a specific trigger (role change, funding event, hiring pattern, technology adoption) and ties it to a tangible problem you solve. The second type is the proof message, which shares a short, specific result from a similar company in a sentence or two, no case-study PDF attached. The third type is the direct-ask message, which is three sentences long and asks for a reply, yes or no, with no soft framing. Avoid the fourth category that infects most sequences: the "just checking in" message, which is the cadence equivalent of saying nothing out loud. If you cannot write the message in a way that would make sense to the prospect even if they had never heard from you, cut it. Reps should rehearse all three message types aloud; the ones that sound stilted when read aloud will read as spam in the inbox.
True personalization at scale is a myth; relevance at scale is not. The job is to make every message read as if it was written for a cluster of ten similar prospects, not for a generic list of ten thousand. Build three gates into your sequence. The first gate is persona: group prospects by role, not by company size, because the pain language of a sales leader differs completely from the pain language of a CFO even at the same account. The second gate is trigger: segment by observable events like funding, headcount growth, product launches, or technology changes, because a specific trigger gives you a reason to reach out that the prospect will recognize. The third gate is a one-line variable the rep writes themselves per prospect: a sentence that references something only a human reading their LinkedIn or company blog would notice. That one line goes into the first and the fourth email, and nowhere else. Three gates deliver eighty percent of the lift that full handwritten personalization would, at ten times the throughput. Beware the trap of asking reps to personalize every line of every message; they will stop sending, or worse, they will send junk because the standard is impossible to maintain.
A cadence is only as good as what happens when the prospect replies, and most teams lose their best responses because reply handling is undefined. Build three reply paths before you turn the sequence on. The positive reply (interested, wants to talk) must be answered within one business hour during work days; faster if the segment is small and expensive to generate. Lag over an hour and conversion drops by half. The neutral reply (ask a question, request information, not now) earns a human response the same day with a specific next step, never a template. The negative reply (not interested, unsubscribe, wrong person) exits the cadence immediately and updates the record. Never argue with a no; never continue the sequence after one. On top of reply handling, define disqualification criteria that remove a prospect regardless of reply. A prospect who leaves the company, moves out of your ICP, or is already a customer should drop out of the cadence automatically. Treat the cadence as a conversation that respects the prospect's signals. Teams that get this right earn their next cadence attempt at the same account; teams that ignore replies get blocked at the domain.
Never launch a new cadence to your full list. Run it against sixty to one hundred prospects first, split across two or three reps, and measure for one full cycle before scaling. The pilot has three jobs. First, surface the operational breaks: reps who cannot find the LinkedIn variable, phone numbers that bounce, email deliverability issues, templates that render oddly on mobile. Second, measure early signal: open rate on the first email, meaningful reply rate, meetings booked, and cost per booked meeting. Third, hear the rep feedback: which steps feel pointless, which messages are hard to send, which touches generate confusion rather than curiosity. Change nothing during the pilot except operational fixes. Resist the urge to rewrite messages mid-run because an early send performed badly; you need the full cycle of data to know what is signal and what is noise. After the pilot, cut the two lowest-performing steps, rewrite the two highest-opening-but-lowest-replying messages, and only then expand. Teams that skip the pilot and launch wide burn their best prospect lists on a sequence they would have fixed in week one of a pilot.
Most cadence dashboards track vanity: emails sent, open rate, click rate. Those numbers are easy to game and tell you almost nothing about whether the cadence is working. The four metrics that matter are meaningful reply rate (positive and neutral replies divided by prospects touched), meeting-booked rate, meeting-held rate, and cost per sales-accepted opportunity. Meaningful reply rate tells you if the message resonates. Meeting-booked rate tells you if the message converts attention into commitment. Meeting-held rate tells you if the prospects you are booking are real. Cost per sales-accepted opportunity tells you if the whole machine makes economic sense. Track these at the sequence level, at the step level, and by persona. A sequence that performs well in aggregate often has one or two steps producing all the lift and three or four steps producing noise; without step-level metrics you cannot find them. Review the numbers every two weeks for the first quarter, then monthly. Kill any step that produces less than ten percent of its share of meaningful replies for two cycles in a row. Keep the system lean. A cadence that keeps shrinking is a cadence that keeps improving.
For most B2B motions, twelve to sixteen business days with eight to twelve touches is the sweet spot. Shorter than that and you quit before buyers notice you exist. Longer than that and you are chasing people who will never buy, at the cost of calling someone who will. Transactional segments can run shorter; enterprise multi-threaded sequences sometimes run longer, but rarely more than four weeks of active touches.
Email, phone, and LinkedIn carry the load for the vast majority of B2B segments. Email handles the asynchronous explain-and-prove work, phone handles live objection handling and trust, and LinkedIn warms the connection and provides social proof. SMS belongs only after a prospect has engaged somewhere else; cold SMS to strangers reads as a scam, hurts sender reputation, and invites regulatory risk in several jurisdictions. Pick three channels for the opening sequence, never five.
True handwritten personalization at scale is a myth. Aim for relevance at scale instead. Group prospects by persona, segment by trigger event, and require one rep-written variable per prospect that appears in two specific messages. This produces eighty percent of the lift of full personalization at ten times the throughput. Reps who are told to handwrite every line stop sending or send junk; the goal is a defendable, repeatable standard.
Define three reply paths before launch. Positive replies get a human answer within one business hour during work days. Neutral replies get a same-day human response with a specific next step, never a template. Negative replies exit the cadence immediately and update the record. Never argue with a no, never continue after an unsubscribe, and never let an out-of-office auto-response fire the next step. Reply handling is where cadences earn their next attempt at the account or get blocked at the domain.
For a well-targeted cadence in a defined persona and trigger segment, meaningful reply rates of three to seven percent are realistic. Below three percent, the problem is almost always message-market fit or list quality, not volume. Above eight percent, you likely have warm-ish data or a segment with an unusually acute pain, and the next test is to see how far you can scale before the rate drops. Chasing meaningful reply rate matters more than chasing open rate, which has become unreliable because of image pre-fetching.
Review the metrics every two weeks for the first quarter after launch, then monthly. Kill any step that under-produces its share of meaningful replies for two cycles in a row. Rewrite the two highest-opening-but-lowest-replying messages after each full cycle. Avoid rewriting mid-cycle because an early send performed badly; you need the full data cycle to separate signal from noise. Cadences improve by subtraction more often than by addition.
Strkr gives you the channel sequencing, personalization gates, reply handling, and step-level metrics to run this playbook in one system. Build once, pilot in a week, scale what works.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
