Best Time to Cold Call: What the Studies Counted

The study that most cold calling advice rests on was not a study of cold calls. It was an analysis of how quickly companies rang people who had already filled in a web form, run with a professor from MIT on data from six companies, and it is the source of the late afternoon and the Wednesday and Thursday answers that still circulate as facts about cold calling in 2026.

That is not a scandal. It is a good piece of research being asked a question it never asked itself. The useful version of this page is the same treatment cold calling statistics gives to success rates, applied to timing: take each published best time claim, find what produced it, and say what survives once you know what was counted.

The study underneath most of the advice

The Lead Response Management study is published on its own site and describes both its collaborators and its scope. Its stated aim was to work with "Professor Oldroyd from MIT" on research "crafted to ascertain the optimal timing for reaching out to web-generated leads, focusing on key metrics like the day of the week, time of day, and response intervals."

Its population is stated in the same paragraph: "We scrutinized three years of data from six diverse companies, analyzing over fifteen thousand leads and more than a hundred thousand call attempts." That is a real dataset with a real method, and the summary is where the familiar numbers come from. On days, the page reports that "Wednesday and Thursday emerged as the most effective days for initiating contact (49.7% more effective than the least productive day) and qualifying leads (24.9% more effective)". On hours, it reports that "Calling between 4 to 6 PM maximized contact chances (114% more effective than the least effective time)", with the eight to nine and four to five windows best for qualifying.

Now read the sentence that decides whether any of it applies to you. The study was "Focusing on leads obtained through web forms and contacted at least once". Every one of those fifteen thousand people had raised their hand. They were expecting a call, or at least would recognise why one was coming. A cold call reaches somebody who has done none of that, and the whole question of when they will pick up and how they will react is a different question. A nineteen year old measurement of inbound speed to lead is being quoted as the timetable for outbound dialling.

The datasets that are genuinely cold calls

Three publishers have put numbers behind cold dialling itself, and all three state a population, which puts them ahead of most of this literature. All three were read on 21 September 2026.

ZoomInfo publishes an analysis off its own platform and names the size in the same sentence as the finding: Tuesday and Wednesday "account for 44% of all demos booked in ZoomInfo's dataset of 1.4M+ outbound calls". It also splits the reason, reporting that Tuesday leads on connect rates while Wednesday leads on raw call volume, and it notes that in Europe, the Middle East and Africa the Wednesday figures improve.

CallHippo states its population in a data note: "The data referenced throughout this article was gathered by analyzing 72,000 call attempts over 90 weeks, collected from 3,000+ businesses across 20+ countries." Its day finding is that "Wednesday came out on top in our data as the strongest day for cold calling, with a 48% higher first-attempt connection rate compared to Monday", with Thursday second. Its hour finding names two windows, late morning and late afternoon, and reports the late afternoon window as sixty eight percent more effective than the late morning hour before noon, which it compares explicitly with the seventy one percent figure the industry repeats.

A third publisher reports on forty thousand outbound calls tracked through its own dialling and transcription stack, and lands on the same two windows, late morning and late afternoon, with the hours before ten and the lunch period performing worst.

So three cold calling datasets, and they do not agree on the day. ZoomInfo puts Tuesday first. CallHippo puts Wednesday first. The inbound study everyone borrows from puts Wednesday and Thursday first. The hours cluster more tightly than the days do, which is itself informative.

What was counted, which is where the disagreement lives

The three datasets are not measuring the same outcome, and once that is on the table the disagreement mostly dissolves.

A connect rate is dials that reach a human. A conversation rate is connects where the person stayed past the opening. A meeting or demo rate is the outcome anyone is paid for. Our cold calling tips page draws that split and argues that the first of the three is mostly a data and reputation question rather than a skill one. ZoomInfo's headline is about demos booked, so it is ranking the last of the three. CallHippo's day figure is a first attempt connection rate, which is the first. Those two can honestly disagree about the best day, because a day with the most people picking up is not necessarily the day with the most people willing to take a meeting once they have.

The hour findings survive better for the same reason in reverse: whatever you are counting, a person who is at their desk and between commitments is more reachable than one who is not, and late morning and late afternoon are when that is true in most office calendars.

Inbound response studySix companies, three years

• Called people who filled in a web form
• Counted contact and qualification
• Published with a named academic collaborator
• Best days given as midweek
• Not a study of cold calling

Platform call logsVendor datasets

• Cold dials placed through one product
• One counts demos booked, another counts connections
• Each states how many calls it looked at
• Their best days do not agree
• Published by companies selling to callers

Your own call logThe only local answer

• Your list, your reps, your market
• You choose which outcome to count
• Needs several reps pooled to read
• Accumulates across a quarter
• The only version that detects your own changes

The age problem, and the vendor problem

Two further discounts apply before any of this reaches a dialling plan.

The first is age. The inbound study is from the era when a desk phone was where a business call landed. Since then the direct dial has largely become a mobile, spam labelling has arrived on the receiving handset, and the working day has stopped being the same shape for everybody. A finding about when people are at their desks is not obviously portable to a world where the phone is in a pocket. None of the pages repeating those numbers says when they were produced.

The second is that every publisher of a cold calling timing figure sells something to people who make cold calls. That does not make the measurements wrong; it decides which ones get published and repeated. A dialler vendor has a commercial reason to find that timing matters a lot, because the product that reports timing is the one being sold. Treat a vendor finding as a named vendor's finding, which is how it is quoted above, and be more careful with the ones that arrive with a recommendation attached.

There is a third constraint that overrides all of it and is not about performance. Calling hours are regulated, the federal floor is not the whole picture, and several states run narrower windows than the federal one. What that means in practice, and which layer binds, is in when cold calling is against the law. A window that tests well and sits outside a permitted calling hour is not available to you.

Reading a best time claim

• ✓Yes: The dataset is named and its size is stated
• ✓Yes: The calls were cold rather than follow ups to an enquiry
• ✓Yes: The outcome counted is named: connects, conversations or meetings
• ✓Yes: The year the data was collected is given, not just the article year
• ✓Yes: Times are the prospect's local time rather than the caller's
• ✕No: A nineteen year old inbound study is being quoted as cold calling data
• ✕No: The window sits outside a permitted calling hour in a state on your list

Measuring it yourself, which takes a fortnight

Your own call log answers this better than any of the sources above, because it is measured on your list, your reps and your market, and it is the only version that can detect your own changes.

Record four counts and keep them separate, as cold calling statistics argues: dials attempted, connects, conversations and meetings booked. Stamp each with the prospect's local hour and weekday. Then read connects by hour band rather than by exact hour, because the exact hour is noise at any volume a small team produces.

Here is the shape of the reading on invented figures, offered only to show how thin the base usually is. Suppose a rep places six hundred dials in a fortnight, spread across four hour bands, and connects with thirty six people in total: eleven in the late morning band, seven at lunch, eight in the early afternoon and ten in the late afternoon. Those figures are invented and describe no team. Eleven against ten is not a finding, and a fortnight of one rep's dialling will almost never produce one. Two things help. Pool several reps before reading anything, and accumulate across a quarter rather than a week, because the same four counts over three months say something the same four counts over a fortnight cannot.

The other reason to measure rather than inherit is that your best window is a property of who you call. Owner operated businesses, shift based operations and corporate offices are reachable at different hours, and no published average can tell you which one your list resembles. That is the argument our cold calling tips page makes about inheriting a window from a listicle.

The short version

The most repeated best time to cold call figures come from a study of inbound web form leads at six companies, published with its population and its scope stated, and quoted for nineteen years as though it were about cold calls. Three genuine cold calling datasets exist, each from a vendor, each stating its size, and they disagree about the best day because they count different outcomes.

The hours agree more than the days do: late morning and late afternoon, in the prospect's local time, inside permitted calling hours. Treat that as a default, measure your own connects by hour band across a quarter with several reps pooled, and keep the four counts separate so the number that moves points at the thing that caused it. What cold calling is and where it still works is the wider frame. If the honest answer is that you need more conversations rather than better timed ones, see what a first campaign produces on the written channels.

Figures above were read on the publishers' own pages on 21 September 2026, with dated snapshots kept as evidence, and each is attributed to the page it appears on. RevenueFlow runs email and LinkedIn rather than phone and holds no calling data of its own. Verify current figures at source before relying on them.