Playbook · Outreach
Sales Cadence: How Many Touches, Over How Many Days
By Marián Cabadaj · Published 13 September 2026 · 9 min read
A sales cadence is the fixed sequence of touches you run against a prospect: how many, on which channels, spaced how far apart. Most of the argument about cadences is about the first number. The one that decides the outcome is the third.
The failure I see most often is a cadence built as a calendar rather than as a list of reasons to make contact. Once the sequence has more slots than the rep has things to say, every touch after that point exists because the tool scheduled it, and the reader can tell.
This guide covers what a cadence actually is, why most stop working around step four, how many touches each kind of list deserves, which channels are good at which job, how to space the gaps, how to test a change without fooling yourself, and the four numbers worth watching. Every figure comes from The Bridge Group, Salesforce or TrustRadius directly.
What is a sales cadence?
A cadence is a repeatable sequence of contact attempts aimed at one prospect, defined by three things: the number of touches, the channels they run on, and the days between them. It is the operating schedule of outbound. Everything else, the copy, the offer, the list, sits inside it.
Every touch needs a reason of its own
The working definition worth holding is that each touch has to carry a reason that did not exist in the previous one. A new angle, a different pain, a piece of evidence, a question you have not asked. A touch whose only content is that the last one went unanswered is a reminder, and a reader who receives four reminders learns that you have nothing else to say.
Where the cadence ends and follow-up begins
A cadence runs before there is a conversation. Follow-up runs after one. They get confused because both are sequences of messages, but their content rules are opposite: a cadence has to invent its own reasons, while follow-up inherits reasons from what was said on the call. Our guide to cold email follow up covers the second half.
Why do most sales cadences stop working?
Because the rep runs out of things to say before the sequence runs out of steps, and the tool keeps sending anyway. The published data shows how little room there is to waste.
The Bridge Group's 2025 SDR Models, Motions and Metrics study of 351 B2B companies puts the monthly quota at 10 meetings held, down 40% since 2018, with 60% of SDRs hitting it, the lowest share in the study's history.
What this means for a B2B sales team: targets fell and attainment fell with them. Adding steps to a cadence is the cheapest response available and it is the one that has already been tried across the whole market for seven years. The number that moved was not touch count.
Salesforce's State of Sales 2026, a survey of 4,050 sales professionals across 22 countries fielded in August and September 2025, found 48% of reps say they lack the bandwidth for adequate cold outreach while already giving close to a full working day a week to prospecting.
What this means for a B2B sales team: a longer cadence is not free. It is paid for out of the same day that is already full, so every step you add to one sequence is a step removed from another contact. The question is never whether a fourteenth touch could work. It is what the same hour would have earned somewhere else.
How many touches should a cadence have?
Between 5 and 16, and the spread is that wide because the answer is a function of how the list was built. A best practice number copied off a vendor blog is a number fitted to somebody else's list.
| How the list was built | Touches | Over | Why that many | What goes wrong |
|---|---|---|---|---|
| Cold, bought or scraped | 12 to 16 | 21 to 28 days | Nothing happened. You are buying attention with repetition | Reads as pressure, because it is |
| Fit-matched, no signal | 9 to 12 | 18 to 21 days | They belong in the market but gave you no date | Every touch has to invent its own reason |
| Signal-sourced | 5 to 7 | 10 to 14 days | Something happened, and it is still recent | The signal expires faster than the sequence runs |
| Inbound or referred | 3 to 5 | 5 to 10 days | They came to you. Speed matters more than volume | Treating it like cold and losing the warmth |
Run your own numbers before you commit a quarter to them
Touch count only matters through what it produces. Put your own reply rate and meeting rate in and the sequence prices itself in working days.
Salesforce's same State of Sales 2026 research puts the average seller at 40% of their time actually selling, with the rest going to admin, research and internal work.
What this means for a B2B sales team: the cadence you design on paper assumes a full day of touches and gets four tenths of one. A 16-step sequence planned against an imaginary day becomes an 8-step sequence in practice, executed late, which is worse than an 8-step sequence executed on time.
Which channels belong in the sequence?
All of them, used for the job each is good at. Multichannel fails when it means the same message pushed through four pipes, which the reader experiences as one message sent four times.
One of the four has a ceiling the others do not, and it is the one most plans over-allocate. Our guide to LinkedIn outreach covers what that channel can actually carry, from the limits LinkedIn refuses to publish down to the 200 characters you get in a note.
| Channel | The job it is good at | What it costs | Worth knowing |
|---|---|---|---|
| Carries detail and survives being read later | Low per touch, high per mistake | The only channel where a bad send damages the next 100 | |
| Proves you are a person and gives the name a face | Medium, and rate limited by the platform | A connection request is a touch, so count it as one | |
| Phone | Gets an answer today, including the answer no | High per touch, unlimited by any platform | The only channel that ends a cycle early |
| Voice note or video | Buys attention when text has stopped working | High, so it is a late-sequence move | Loses its effect the moment it becomes routine |
The phone is the channel most cadences under-use, and the reason is that it is the only one that can produce a no. A no ends the cycle, frees the slot and costs nothing further. Silence does none of those and looks better on a dashboard, which is exactly why teams drift towards it.
What this means for a B2B sales team: judge a channel by how fast it resolves a contact, not by its reply rate. A channel with a 4% reply rate that ends half your open cycles is worth more than one at 8% that leaves them all open.
How should the touches be spaced?
Tight at the start, wide at the end. One to two days between the first three touches, three to five days through the middle, then a week. Almost every tool defaults to an even gap, and an even gap is wrong at both ends of the sequence.
Why the early gaps are short
The first three touches are one attempt to be noticed, not three separate ones. Spread them over two weeks and each arrives with no memory of the last, so you pay the cost of a cold open three times and get the benefit of none.
Why the late gaps are long
By touch eight the reader has decided not to answer. The only thing that changes that decision is a change in their situation, which happens on their clock, not yours. Wide gaps let the sequence still be running when something moves.
Field note: the single change I have seen move a cadence most is moving the phone call from step nine to step two. Teams put the call late because it feels like the escalation. Late is where it has the least value, because by then the contact has already learned to ignore the name.
How do you test a cadence without guessing?
One variable, one list, enough contacts that the result is not noise. The reason this is hard is that a cadence has four variables at once and changing all four is what most teams call a test.
Split on the sequence, not on the subject line
Subject line tests are popular because they are easy and they answer a smaller question than they appear to. Testing a cadence means holding copy still and changing the structure: the number of steps, the order of channels, or the spacing. Our guide to cold email subject lines covers the copy half, which is a separate experiment.
TrustRadius's 2026 B2B Buying Disconnect, a January 2026 survey of 1,862 technology buyers and 444 vendors, found that 83% of buyers shortlisted three or fewer products.
What this means for a B2B sales team: the cadence is competing for one of three slots, and it is competing on relevance rather than persistence. A test that improves persistence and leaves relevance alone is measuring the wrong half.
Which cadence numbers actually matter?
Four, and each one points at a different fix. A metric that does not name a fix is a dashboard.
| Number | What it reads | The bad pattern | What it actually points at |
|---|---|---|---|
| Reply rate per step | Whether the message earns an answer at all | Step 1 is fine, step 4 is dead | The later steps repeat the first. Give each one its own reason |
| Meetings per 100 contacts | The whole cadence, end to end | High replies, few meetings | The ask is wrong, not the copy |
| Time to first reply | Whether the spacing matches the market | Most replies land after the sequence ends | The cadence is too short, or too tightly packed |
| Negative reply rate | Whether the list deserves the sequence | Rising while the reply rate holds | Targeting, not cadence. Stop tuning the sequence |
Field note: reply rate per step is the one worth instrumenting first, because it is the only number that tells you where the sequence dies rather than that it died. A cadence with a healthy average and a cliff at step four is a normal finding, and the average hides it completely.
The Bridge Group's same 2025 study puts pipeline per SDR at $3.78M, up sharply from $2.83M.
What this means for a B2B sales team: more pipeline is being carried by the same number of people, so the cost of a cadence that wastes touches has gone up rather than down. Efficiency per contact is now the constraint, which is an argument for a shorter sequence against a better list, not a longer one against the same list.
The cadence at a glance
One table, for the version of this page you will actually come back to.
| Decision | The short answer |
|---|---|
| Touch count | 5 to 7 on a signal-sourced list, 12 to 16 on a cold one |
| Spacing | 1 to 2 days early, 3 to 5 in the middle, a week late |
| Channels | Email for detail, phone to resolve, LinkedIn for a face |
| Where the call goes | Step two, not step nine |
| When to stop | When the reasons run out, not when the steps do |
| What to test | Structure, with the copy held still |
| First number to fix | Reply rate per step, never the average |
How Leadalf runs cadence
Everything above assumes the hard part is the sequence. It is not. The hard part is that a cadence has to invent a reason for every touch, and inventing reasons is exactly what a rep runs out of around step four.
The differentiated point is this: a cadence is only as short as its list lets it be. The reason a signal-sourced list needs 5 touches and a cold one needs 16 is that the first arrives with a reason already attached and the second does not. Most cadence advice tries to solve a list problem with sequence design, which is why it converges on more steps every year.
What the workflow looks like
Leadalf supplies the reason, and leaves the sequence to the tool you already run it in.
- Signals in: it watches LinkedIn, Reddit, X, the Meta Ads Library and job postings, so a contact enters a sequence because something happened rather than because they matched a filter.
- Scored against your ICP, not a keyword: each signal is scored against the ICP boundary, which is what keeps a keyword match from becoming a step one.
- Delivered with the reason attached: the people behind passing signals are enriched and pushed into Instantly or HeyReach carrying the reason they qualified, which is the first line the sequence would otherwise have to invent.
See how the signal sourcing works end to end, or what a filtered lead costs in credits. If you want the role that runs the cadence rather than the cadence itself, our guide to the SDR role covers the daily shape of the work.
It does not write your sequence and it does not send for you. Those belong in Instantly or HeyReach, and a vendor that claims both ends is usually weaker at one of them.
FAQ
How many touches should a sales cadence have?
Between 5 and 16, and the range is that wide because the number depends on how the list was built rather than on a best practice. A signal-sourced list where something just happened needs 5 to 7 touches over 10 to 14 days, because the reason for contact is already there and it expires. A cold bought list needs 12 to 16 over three to four weeks, because nothing has happened and repetition is the only thing buying you attention. Copying the cold number onto a warm list is the most common way to make a good list perform like a bad one.
How long should the gaps between touches be?
Short at the start and longer at the end. One to two days between the first three touches, then three to five days, then a week. The reason is that the early touches are still part of one attempt to be noticed, so spreading them out breaks the thread, while the later ones are separate attempts and need enough distance that they do not read as pestering. Fixed spacing across the whole sequence is the default in most tools and is wrong at both ends.
Should a cadence use more than one channel?
Yes, but for the jobs each channel is good at rather than for the sake of coverage. Email carries detail and survives being read three days later. The phone is the only channel that can end a cycle today, including ending it with a no, which is worth more than a month of silence. LinkedIn proves you are a person. A multichannel cadence that sends the same message through four pipes is still one message, and the reader experiences it as four.
When should a cadence stop?
When the reasons run out, not when the step count does. If you can write the next touch without inventing a pretext, keep going. If the next touch would exist only because the sequence has a slot for it, stop and move the contact to a quarterly list. The teams that run 20 touches are usually running 6 reasons and 14 reminders, and the reader can tell which is which.
Does a cadence need different copy for each step?
It needs a different reason for each step, and the copy follows from that. A new angle, a different pain, a piece of evidence, a question you have not asked yet. What does not work is the same message rewritten, which is what a follow-up that opens with a reference to the previous email usually is. Our guide to cold email follow-up covers how many of those a sequence can carry before the reply rate falls away.