Ask ten sales leaders how to build a sales cadence and you will get ten touch counts, all quoted with total confidence and none with any reasoning attached. Eight touches, twelve touches, sixteen over six weeks. The number is presented as the answer when it is actually an output of three decisions nobody discusses: how many prospects you have, how much time your team has, and how quickly your market decides.
This guide treats sales cadence design as a resource allocation problem rather than a template choice. You will get the reasoning behind touch count, the spacing patterns that outperform, three worked cadences for different deal shapes, the arithmetic that sizes a cadence to your headcount, and clear rules for when to stop.
What a sales cadence is really allocating
A cadence distributes a fixed quantity of rep attention across a list. That is the whole mechanism. Every design decision is a trade between depth per prospect and breadth across the list, and the two move in opposite directions.
A rep with 25 hours of outbound time a week can run a 14-touch cadence across 40 prospects or a 6-touch cadence across 200. Both are defensible. What is not defensible is choosing a touch count from a blog post and discovering three weeks later that the team is silently skipping the phone steps because the day does not contain enough hours.
So start from capacity, not from a template. The best sales cadence is the most touches you can genuinely complete, applied to the smallest list that still hits your number.
How many touches a sales cadence needs
The commonly cited figure is that it takes around eight touches to reach a prospect, while most reps stop after two. Call-recording research such as Gong Labs’ analysis of cold calls supports the persistence half of that claim. The gap between those numbers is where most pipeline is lost, and it is genuine. But eight is a median across wildly different situations, and applying it uniformly is how teams over-invest in bad-fit accounts.
| Prospect type | Touches | Duration | Reasoning |
|---|---|---|---|
| Inbound, high intent | 5–7 | 7–10 days | Already interested; speed matters more than persistence |
| Cold, well-targeted ICP | 8–12 | 17–21 days | Needs recognition built from nothing |
| Trigger-led (funding, hiring, job change) | 6–9 | 10–14 days | Window is time-bound; relevance is already high |
| Enterprise, multi-threaded | 12–16 | 30–45 days | Several stakeholders, longer internal cycles |
| Re-engagement, closed-lost | 4–6 | 14 days | Prior context means fewer touches to a decision |
Notice that trigger-led sequences need fewer touches than cold ones. Relevance substitutes for repetition, which is the strongest practical argument for building cadences around sales trigger events rather than around a static list.
Spacing: the part that matters more than the count
Two cadences with identical touch counts can perform very differently depending on how those touches are distributed. The pattern that consistently works is tight at the start, loose at the end.
Front-loading works because recognition decays. Touch one and touch two need to be close enough that the second one lands while the first is still faintly familiar. Once recognition exists, later touches can be spaced further apart without losing it, and wider gaps reduce the sense of being pursued.
| Phase | Days | Touches | Gap | Goal |
|---|---|---|---|---|
| Establish | 1–8 | 4–5 | 1–3 days | Build recognition fast |
| Develop | 9–16 | 3–4 | 3–4 days | Advance the argument |
| Resolve | 17–21 | 2–3 | 4–5 days | Force a yes or a no |
Three spacing errors account for most underperformance. Even spacing across the whole window wastes the recognition window entirely. Weekend touches on B2B lists get buried by Monday. And clustering two channels on the same day makes the sequence feel automated, which is the one impression that cannot be recovered.
Three sales cadence templates
These are complete and runnable. Adjust the channel mix to what your data supports, and check the capacity arithmetic in the next section before committing.
The standard cold cadence: 10 touches, 21 days
| Day | Channel | Content |
|---|---|---|
| 1 | Trigger-led opener, one idea, under 90 words | |
| 2 | Connection request | |
| 4 | Phone | First dial, no voicemail |
| 5 | New angle with one proof point | |
| 8 | Phone | Second dial plus voicemail |
| 9 | References the voicemail directly | |
| 12 | Message or InMail for tier one | |
| 15 | Phone | Third dial, different time of day |
| 18 | One short, specific question | |
| 21 | Honest close-the-loop |
The trigger cadence: 7 touches, 12 days
Compressed because the trigger has a shelf life. Days 1 and 2 both carry touches, and the whole thing resolves before the event goes stale.
- Day 1: Email naming the trigger explicitly, plus a profile view
- Day 2: Connection request or InMail for high-value accounts
- Day 3: Dial, with voicemail
- Day 5: Email with a peer example from the same situation
- Day 8: Dial at a different hour
- Day 10: LinkedIn message with a direct question
- Day 12: Close the loop and return to nurture
The inbound cadence: 6 touches, 8 days
Here speed dominates everything else. The first touch should happen in minutes, not hours, and the case for that is covered in our piece on speed to lead.
- Minute 5: Dial immediately
- Minute 10: Email referencing exactly what they downloaded or viewed
- Day 2: Second dial at a different hour
- Day 3: LinkedIn connection request
- Day 5: Email with the specific resource their behaviour implies they need
- Day 8: Final dial and close the loop
Sizing a sales cadence to your team
Here is the calculation that should precede any cadence decision, and it takes about five minutes.
Take a rep with 25 hours of genuine outbound time per week. Assume three minutes per dial including notes, ninety seconds per personalised email, and one minute per LinkedIn action. The standard ten-touch cadence above contains four dials, five emails and two LinkedIn actions, costing roughly 21 minutes of rep time per prospect across the full 21 days.
Twenty-five hours is 1,500 minutes, which supports about 71 prospects in flight at any time, or roughly 24 new prospects entering per week once the pipeline is full.
| Cadence | Minutes per prospect | Prospects in flight | New per week |
|---|---|---|---|
| 6 touches, email and LinkedIn only | ~8 | ~185 | ~60 |
| 10 touches, three channels | ~21 | ~71 | ~24 |
| 14 touches, phone-heavy | ~34 | ~44 | ~15 |
Now work backwards from your target. At a 10% sequence reply rate and 40% of replies converting to meetings, 24 prospects a week produces roughly one meeting a week per rep. If the target is four, you need either four times the capacity, a materially better list, or a lighter cadence across more prospects. That is a real strategic choice, and it is invisible until you do the arithmetic.
Rules that raise reply rate without adding touches
Before extending a sales cadence, exhaust the changes that cost nothing.
- One idea per touch. Messages carrying three arguments get answered on none of them.
- Every touch advances. If a message could be deleted without the sequence losing meaning, delete it.
- Reference the previous touch. This is what separates a sequence from a set of unrelated messages.
- Never send “just following up”. It carries no information and signals you have nothing to add.
- Vary the ask size. Early touches should ask for an opinion; only later ones ask for time.
- Protect deliverability. More email touches means more volume per domain. Read cold email deliverability before increasing send counts.
Message quality inside the cadence matters more than the cadence structure itself. Our guide to cold email sequence structure covers what each individual email should carry.
When to stop, and what happens next
Every sales cadence needs an exit, and most teams handle this badly by either giving up at touch three or continuing indefinitely.
- Explicit no: stop immediately across all channels and record the reason. It is the most valuable data you will collect, and under the ICO’s direct marketing guidance an objection must be honoured without friction.
- Not now: exit the cadence, set a dated re-entry, and stop treating it as a live opportunity.
- Total silence: exit at the planned end, then move to a low-frequency nurture of roughly one useful touch a month.
- Bounce or wrong number: stop and correct the record, because carrying bad data forward compounds. See CRM data hygiene.
Silence is not rejection. It usually means wrong timing, and the account should re-enter when a trigger fires rather than on an arbitrary date. That is the practical value of monitoring buying intent signals against your closed and dormant lists.
Testing a sales cadence without fooling yourself
Most cadence “testing” is storytelling. A team changes four things at once, sees replies rise, credits whichever change they liked best, and rolls it out. Two months later performance is back where it started and nobody knows why.
Sample size is the first constraint people ignore. To detect a move from an 8% reply rate to 11% with any confidence, you need several hundred prospects per variant. Below roughly 200 per side you are reading noise, and the smaller the effect you are chasing, the larger the sample required.
| What you are testing | Prospects per variant | Realistic read time |
|---|---|---|
| Structural change (touch count, channel added) | 300+ | 4–6 weeks |
| Spacing pattern | 300+ | 4–6 weeks |
| Opening message angle | 200+ | 2–4 weeks |
| Single subject line | 500+ | 2–3 weeks |
Four rules keep the results honest. Change one variable at a time, even when that feels painfully slow. Split the list randomly rather than by rep or by territory, since rep skill will otherwise explain your result. Let every cohort finish the full sequence before reading it, because late replies are disproportionately the good ones. And write the hypothesis down first, so you cannot retrofit an explanation to whatever the data happens to show.
Where volumes are too low for clean testing, which is common for teams selling into small markets, stop testing copy and test targeting instead. List quality produces larger effects than cadence structure, so it is detectable at smaller sample sizes.
Adapting the cadence to deal size
The right cadence for a £5,000 annual contract is not a shortened version of the right cadence for a £150,000 one. The economics differ, and so should the shape.
| Deal size | Cadence shape | Personalisation | Contacts per account |
|---|---|---|---|
| Under £10k | 6–8 touches, email-led | Segment level | 1 |
| £10k–£50k | 10–12 touches, three channels | Trigger plus one researched detail | 1–2 |
| Above £50k | 12–16 touches over 6 weeks | Individually written | 3–5, multi-threaded |
Multi-threading is the change that matters most at the top end. Above about £50,000, sequencing three or four stakeholders in the same account concurrently outperforms pursuing one contact harder, because enterprise decisions are rarely made by the person who answers first.
Frequently asked questions about sales cadence design
How many touches is too many?
Beyond about 14 touches on a cold prospect, marginal returns collapse and irritation rises. Watch where replies actually land in your own data. If nothing arrives after touch nine, the touches after it are costing you goodwill for nothing.
Should every prospect get the same cadence?
No. Tier the list. High-value accounts justify phone touches and manual research; the long tail gets email and LinkedIn. Uniform cadences either over-serve the tail or under-serve the accounts that matter.
What is a good reply rate?
For a multichannel cold cadence on a well-targeted list, 7% to 12% at sequence level. Cold email alone typically runs 3% to 5%. Anything above 20% usually indicates a warm list rather than an exceptional cadence.
Does adding phone touches actually help?
Yes, when you have verified mobile numbers. Phone is the most expensive touch per unit of rep time, so it belongs on tier-one accounts. See cold calling scripts for what to say when someone answers.
How do I stop the cadence when someone replies anywhere?
Cross-channel reply detection has to be automatic. Manual monitoring fails at volume, and the failure is visible to the prospect. A unified inbox with shared sequence state handles it properly.
Design from capacity, not from a template
A sales cadence is a plan for spending finite attention. Work out how much you have, decide how deep you want to go per prospect, and let the list size fall out of that. Teams that pick a touch count first end up with a cadence nobody completes, which performs worse than the shorter one they could have run properly.
ZenBee builds cadences across LinkedIn, email and phone over verified data, with signal-based entry and automatic cross-channel exit. Request a demo or see pricing.