How to Increase Contact Rate on Inbound and Outbound Leads

Contact rate is the share of unique leads you reach in a two-way conversation — not the share you dial. In our own client work, most teams stop at two attempts. On a 25% per-attempt reach model, two attempts reach 306 of 1,000 leads and six attempts reach 575: the same list, the same offer, 88% more conversations.

At a glance

• The metric. Contact rate = unique leads reached in a two-way conversation ÷ unique leads worked in the period. Dials, voicemails and delivered SMS are not contacts.
• The gap. In the campaigns we run, most cadences stop at 2 attempts. On the model below, attempts 3–6 add 269 contacts per 1,000 leads — more than attempt one produced on its own.
• The ceiling. More effort keeps working but with sharp diminishing returns. Pew Research Center lifted a phone survey response rate from 9% to 22% by going from a minimum of 7 call attempts over 5 days to a minimum of 25 over two and a half months — plus larger incentives and posted letters.
• The Australian constraint. Telemarketing calls to Australian numbers are only permitted weekdays 9:00am–8:00pm and Saturdays 9:00am–5:00pm, with no calls on Sundays or national public holidays. That is a 63-hour legal week.
• What to change first. Fix the denominator, then take the cadence from 2 attempts to 6 spread across different days and times. Everything else is smaller.

How is contact rate actually measured?

Contact rate is unique leads reached in a two-way conversation, divided by unique leads worked in the period. The denominator is people, not dials. If 1,000 leads entered your CRM in September and 380 had a real conversation that month, your contact rate is 38% — whether that took 1,200 dials or 6,000.

Three numbers get called “contact rate” and they are not the same thing:

• Connect rate — connected calls ÷ dial attempts. A per-dial efficiency number. Useful for judging list quality and caller ID reputation, useless for judging coverage.
• Contact rate — unique leads reached ÷ unique leads worked. The per-person number, and the one that feeds every stage downstream.
• Right-party contact (RPC) rate — unique leads where you reached the intended person, not their partner, receptionist or a reassigned number. On outbound lists this is materially lower than contact rate and it is the honest one to report.

A team that reports connect rate as contact rate can double its dialling, halve its connect rate, and show a flat number while its actual coverage improves — which is why the denominator has to be fixed before any lever is pulled.

Set the window explicitly too: contact rate over a rolling 14 days and over the life of the lead are different numbers, and long-tail lists look far worse on the first.

How it works

How to lift contact rate in four moves

01

Fix the denominator

Count unique leads reached in a two-way conversation, divided by unique leads worked. A dial, a voicemail or a delivered SMS is an attempt, not a contact.

02

Spread six attempts

Move the cadence from two attempts to six, staggered across at least four different time bands and days. Six dials in one morning behave like one.

03

Add a reply channel

Pair each unanswered call from attempt two onward with a consented SMS or email. That raises the reachable ceiling rather than climbing the same curve.

04

Re-derive the curve

Rebuild the attempts-to-contact curve from your own call logs each month. Replace the model’s reach probability and ceiling with your measured numbers.

How many call attempts does it take to reach a lead?

Here is the arithmetic end to end, so you can substitute your own inputs. This is a model, not a measurement of anyone’s campaign. Two assumptions, both stated:

• A reachable ceiling of 70% — on a 1,000-lead list, 700 people can be reached at all within a fortnight. The other 300 have dead or wrong numbers, screen everything, or never answer.
• A 25% per-attempt reach probability among the reachable people you have not yet reached, with attempts spread across different days and times.

Cumulative contacts after n attempts = 700 × (1 − 0.75ⁿ). That produces the attempts-to-contact curve:

Attempt	New contacts this attempt	Cumulative contacts	Cumulative contact rate
1	175	175	17.5%
2	131	306	30.6%
3	99	405	40.5%
4	74	479	47.9%
5	55	534	53.4%
6	41	575	57.5%
7	32	607	60.7%
8	23	630	63.0%
9	17	647	64.7%
10	14	661	66.1%

Three things fall out of it, and none of them are opinions:

1. Attempts 3 to 6 add 269 contacts — more than the 175 the first attempt produced. A cadence that stops at two throws away the largest single block on the curve.
2. More than half of every contact you will ever make arrives after attempt two. Of the 661 contacts made by attempt ten, 355 (54%) came at attempt three or later.
3. The curve flattens hard after six. Assuming you stop dialling a lead once you reach it, attempts 7 to 10 take 1,541 dials and return 86 contacts, where attempts 5 and 6 took 987 dials and returned 96.

We call this the two-and-six gap: the distance between where cadences stop and where the curve’s economic knee actually sits. It is the whole answer to “but we already call them”. You do call them. You call them twice, and twice is the point at which you have collected roughly half of what the list will give you.

Substitute your own figures. At a 55% ceiling and 18% per-attempt reach, six attempts get you to 38.3%, not 57.5% — and on this page’s own yardstick the case for attempts 7 through 10 gets weaker, not stronger: a lower ceiling leaves less left to win, so those four attempts cost roughly 42 minutes of caller time per marginal conversation against about 32 minutes in the base case.

What does independent research say about attempt counts?

Almost all of the widely quoted “six attempts” sales statistics trace back to vendor studies whose datasets are no longer inspectable. One large, methodologically documented public comparison of call effort against reach does exist, and it is not from a sales vendor: Pew Research Center’s 2012 study of survey representativeness ran the same questionnaire twice, at standard and at high effort.

The standard survey made a minimum of 7 attempts per number over a five-day field period and achieved a 9% response rate. The high-effort survey made a minimum of 25 attempts per landline and 15 per mobile over two and a half months, raised the incentive from the $5 the standard survey paid to cell-phone respondents to $10 and then $20 for everyone, posted letters with a further $2 to landline households that had refused, left voicemails, and put its most persuasive interviewers on the phones. It achieved 22%.

Roughly three and a half times the call attempts, plus money and letters, produced about two and a half times the response rate — 9% to 22% — which is the shape of the attempts curve confirmed on someone else’s data, at a scale no sales team will ever fund.

Two caveats. Pew was recruiting survey respondents, not buyers, so the absolute rates do not transfer; and the lift came from incentives and letters as well as attempts, so it is not a clean attempts-only measurement. What does transfer is the method note: Pew staggered calls “over times of day and days of the week (including at least one daytime call) to maximize the chances of making contact”. Six attempts in one Tuesday morning is not six attempts. It is closer to one.

Which levers move contact rate, ranked by effect size?

Ranked by how much they move the number on a typical Australian list, largest first.

#	Lever	What it changes	Rough size on the model above
1	Attempt count, 2 → 6	Cumulative contacts per 1,000 leads	306 → 575 (+88%)
2	Spreading attempts across days and time bands	Per-attempt reach probability	Restores independence between attempts; six clustered dials behave like one or two
3	Adding a consented SMS or email reply path	Reachable ceiling	Converts “did not answer” into a reply channel; lifts the 70% ceiling rather than the curve
4	Time to first attempt on inbound	Reach on attempt 1 only	Largest single-attempt effect on inbound; no effect on cold outbound
5	Caller ID consistency and number reputation	Answer rate per dial	Withheld or rotating numbers get answered less; CLI is mandatory in Australia regardless
6	Data hygiene — mobile vs landline, format, deduplication	Denominator accuracy	Usually a reporting fix, not a contact fix, but it stops you optimising a fiction

Lever 4 has its own literature, set out in our guide to the speed-to-lead 5-minute rule and the research behind it; it matters here only because it overlaps with contact rate rather than adding to it. Lever 3 has its own constraints, covered in our guides to SMS outbound automation and consent and email and SMS deliverability at volume.

Why inbound and outbound contact rates are not the same number

Reporting one blended contact rate across both hides the only two facts that matter about them: they have different ceilings and different legal ground.

	Inbound (they enquired)	Outbound (you initiated)
Consent basis	The enquiry itself, if the form says what you will do with the number	No consent needed to dial, but Australian Privacy Principle 7 of the Privacy Act 1988 governs using the data for direct marketing, plus a Do Not Call Register wash for consumer numbers
Do Not Call Register	A call made with the account holder’s consent is not an unsolicited telemarketing call	Wash the list; the defence relies on a wash within the last 30 days
Calling hours	Standard hours apply unless the person consented to being called at that time	Standard hours apply
First attempt window	Seconds to minutes — intent decays fastest here	Days; there is no decaying event to chase
Sensible first channel	Call, with an SMS fallback in the same minute	Call, with the reply channel introduced after attempt 2
Where attempts should sit on the curve	Front-loaded: attempts 1–3 inside 48 hours	Spread: 6 attempts across 10–14 days and both time bands

An inbound lead has a clock on it and an outbound lead has a calendar on it, so the same six attempts have to be shaped completely differently even though the metric they feed is identical.

What the law lets you do in Australia

General information only, not legal advice. Every rule below is linked to the regulator that made it.

Voice calls. The Telecommunications (Telemarketing and Research Calls) Industry Standard 2017 applies to all voice calls to Australian numbers that offer, advertise or promote goods or services — including calls to numbers not on the Do Not Call Register, and including calls back to your own enquirers. Permitted telemarketing calling times, unless the consumer has consented to being called at that time:

Call type	Weekdays	Saturday	Sunday	National public holidays
Telemarketing calls	9:00am – 8:00pm	9:00am – 5:00pm	No calls allowed	No calls allowed
Research calls	9:00am – 8:30pm	9:00am – 5:00pm	9:00am – 5:00pm	No calls allowed

The same standard requires calling line identification on every attempt, requires the return-contact number to stay live for at least 30 days, and requires the call to be terminated immediately if the person asks. ACMA’s enforcement options run “from formal warnings to penalties of up to $250,000”.

The Do Not Call Register. Consumer numbers must be washed. Two details that change outbound list strategy and are routinely got wrong: the Register states that “Business telephone numbers are not eligible for registration”, though a dual-purpose number “may be eligible for registration if it is used mostly for private/domestic purposes” — which covers a great many sole traders’ mobiles. And a marketer is not in breach if they “had washed their list in the last 30 days and the number was not on the register”, so the wash is a 30-day cycle, not a one-off.

SMS and email. These sit under the Spam Act 2003, not the calling standard, and the Spam Act sets no permitted-hours window at all. ACMA’s three rules are consent (express or inferred, with the sender carrying the burden of proving it), accurate sender identification that stays correct “for at least 30 days after you send the message”, and an unsubscribe honoured “within 5 working days”. Note the asymmetry: you may legally SMS at 8pm on a Sunday when you may not call, which is why a reply channel is the cheapest way to lift the ceiling on the curve. Whether you should is a brand question, not a legal one. Sector-specific obligations are in our guide to AI outbound in regulated Australian industries.

The 63-hour window: what a 6-attempt cadence actually costs

Add up the permitted calling times and Australian outbound has a hard ceiling of 63 legal dialling hours a week — 55 across the weekdays plus 8 on Saturday. A conventional Monday-to-Friday, 9-to-5 roster covers 40 of them. The remaining 23 hours — 37% of the legal window, and the two blocks where consumers are most likely to be at home, being 5:00pm to 8:00pm on weekdays and Saturday morning — are uncovered.

Now the labour. For 1,000 leads at six attempts, dropping each lead from the list once you reach it, budgeting 1.5 minutes for an unanswered dial including ring time, voicemail and disposition, and 6 minutes for a connected conversation including notes:

• 2 attempts: 1,825 dials, 306 contacts, about 69 caller-hours.
• 6 attempts: 4,101 dials, 575 contacts, about 146 caller-hours.
• The difference: 77 extra hours for 269 extra contacts — roughly 17 minutes of caller time per additional conversation.

New leads per month	Dials at 6 attempts	Caller-hours/month	What that requires
Under 150	Under 615	Under 22	One person, by hand, with a shared calendar. A cadence tool is optional.
150 – 600	615 – 2,461	22 – 88	One to two callers, a dialler, and a cadence that survives someone taking leave.
600 – 1,500	2,461 – 6,152	88 – 219	More roster than a 63-hour window absorbs at the 5–8pm and Saturday peaks. Attempt 1 has to stop being a person.
Over 1,500	Over 6,152	Over 219	Two or more full-time callers doing nothing but dialling, before anyone sells anything.

That is the honest cost, and it is why cadences collapse to two attempts without anyone deciding to. The roster runs out, the peak hours go unstaffed, and attempts 3 through 6 quietly never happen. The arithmetic is the same whether a person, a dialler or an AI outbound calling and SMS system makes the attempts — only the hours change, and the hours are the whole constraint.

What to change first

1. Fix the denominator. Recompute last quarter as unique leads reached ÷ unique leads worked, and separately as right-party contacts. If the two are more than 10 points apart, you have a data problem, not a persistence problem.
2. Count your real attempt distribution. Not the cadence you designed — the histogram of attempts actually logged per lead. In our own client work we typically find a median of 2 against a designed 6 — that is our observation across the campaigns we run, not a published statistic.
3. Take the cadence to six, spread across at least four distinct time bands inside the 63-hour window, including at least one 5:00pm–8:00pm attempt and one Saturday morning attempt.
4. Add a consented reply channel from attempt 2. It raises the ceiling rather than climbing the same curve.
5. Re-derive the curve monthly from your own logs and replace the model’s two assumptions with your own numbers.

On the economics: in our own client work we typically see that doubling contact rate roughly doubles the deals that come out of the same list. That is our operating experience across the campaigns we run, not a published study — there is no sample size or window behind it, and we would rather say so than dress it up. How we define and measure lift is set out on our methodology page. One honest wrinkle: contact rate and speed to lead overlap, because answering faster is part of how contact rate improves, so their effects do not multiply cleanly — a team that fixes both should expect less than the product of the two. This page sits inside a wider set on the pipeline stages between a click and a closed deal, which is where to see contact rate relative to set rate and close rate.

Frequently asked questions

What is a good contact rate for sales calls?

There is no credible single benchmark, because almost nobody publishing one states their denominator. On the six-attempt model above, a fresh, well-formatted Australian consumer list lands near 57% within a fortnight; a two-attempt cadence on the same list lands near 31%. Compare yourself against your own prior quarter using a fixed denominator of unique leads worked. Any figure quoted without a denominator and a window is not a benchmark.

How many times should you call a lead before giving up?

Six spread attempts is where the curve’s economic knee sits: attempts 3–6 add 269 contacts per 1,000 leads, while attempts 7–10 take half as many dials again and return 86. Keep going past six only when a closed deal is worth enough to justify roughly 17 minutes of caller time per marginal conversation. Pew Research Center’s high-effort survey went to a minimum of 25 attempts per landline and still only reached a 22% response rate against 9% at seven attempts — persistence keeps paying, but not proportionally.

What are the legal calling hours for sales calls in Australia?

Under the Telecommunications (Telemarketing and Research Calls) Industry Standard 2017, telemarketing calls to Australian numbers may only be made on weekdays between 9:00am and 8:00pm and Saturdays between 9:00am and 5:00pm, with no calls on Sundays or national public holidays, unless the consumer has consented to being called at that time. Research calls get a slightly wider window. The full requirements, including calling line identification and call termination, are published by the Do Not Call Register. General information only, not legal advice.

Does a voicemail or a delivered SMS count as a contact?

No. A contact is a two-way exchange with the intended person — a right-party contact. Voicemails, ringouts, delivered SMS and opened emails are attempts, and counting them as contacts is the most common way a contact rate gets inflated into uselessness. A reply to an SMS is a contact; the delivery receipt is not. If you use SMS or email, ACMA’s spam rules require consent, accurate sender identification and an unsubscribe honoured within 5 working days.

Do I have to check business numbers against the Do Not Call Register?

The Register itself states that “Business telephone numbers are not eligible for registration”, with a limited exception for business fax numbers. But a dual-purpose number “may be eligible for registration if it is used mostly for private/domestic purposes”, which describes a large share of sole trader and micro-business mobiles — so a B2B list built from mobile numbers should still be washed. The safe-harbour position is a wash within the last 30 days. General information only, not legal advice.

Why does my contact rate look fine while my pipeline does not?

Usually because the reported number is a connect rate, or because right-party contacts are being counted alongside gatekeeper conversations. Split it three ways — connect rate per dial, contact rate per lead, right-party contact rate per lead — before concluding anything. Dormant lists need their own baseline; the mechanics are in our guide to running a database reactivation campaign.

Contact rate feeds set rate, which feeds show rate, and each is measured differently; the sales pipeline stages hub maps the whole sequence.