Sales Statistics: Who Ran the Survey, and When
One page promises forty statistics and its own author says twenty. Which of the ranking publishers name their research, their population and their year.
A sales statistic carries four facts or none: who collected it, who was asked, when, and what was divided by what. Salesforce names its own seventh-edition research and dates the page, RAIN Group names a study of more than a thousand sellers without dating it, and The Bridge Group states the tightest population and gates the numbers.
Key takeaways
- Salesforce's page carries two different counts of its own statistics, the headline promising forty and the author's opening line saying twenty, which is the roundup format working as designed.
- The property missing most often is the study's year, and it is the one that decides whether two figures can be compared at all.
- The research with the tightest stated population sits behind a download form, so the figures circulating freely are filtered by accessibility rather than by quality.
- A borrowed figure is legitimate evidence in an argument and a category error as a benchmark, because you cannot inspect the population it was measured over.
Reviewed and updated September 2, 2026
Salesforce's sales statistics page is titled "40 Sales Statistics that Reveal How Teams Can Succeed in 2026". Four paragraphs into the same page, its author writes that he is "sharing 20 sales statistics that show how".
Both numbers are on one page, published by one company, under one byline. Nothing is being hidden and nobody is being dishonest. The headline count is a promise the format makes and the body does not have to keep, and once you notice that, the whole category becomes readable in a way it is not when you are collecting numbers off it.
This piece is about reading the sales statistics literature rather than adding to it. Four properties separate a figure worth quoting from a figure worth ignoring, and the pages currently ranking for this term score very differently against them.
The four things a figure has to carry
A percentage is a fraction that has been through a survey. Four facts have to survive that trip for the result to mean anything to you.
Who ran it. The publisher of the page and the organisation that collected the data are frequently different, and the page rarely makes the distinction visible. A figure lifted from a compilation that lifted it from another one has shed its qualifiers at every hop.
Who was asked. A percentage of sales professionals, a percentage of B2B buyers and a percentage of companies are three different populations, and the same headline number carries opposite implications across them.
When. Not the date on the article. The date of the survey. These diverge by years in this literature, and the divergence is invisible on the page.
What was divided by what. The denominator. This is the failure that runs deepest, and it is worked through for two adjacent territories in cold calling statistics and lead generation statistics, so it is not repeated at length here.
- Depends: The organisation that collected the data is named, not just the page reprinting it
- Depends: The population is described: sellers, buyers, companies or leaders
- Depends: A sample size or participant count appears somewhere a reader can reach
- Depends: The study's own year is stated, separately from the article's date
- Depends: The denominator is named where the figure is a rate rather than a share of respondents
- No: The publisher does not sell the remedy the figure implies
That last line is worth stating plainly rather than pretending otherwise. Every page ranking for this term is published by a company that sells to sales teams, and this one is on the site of an agency that sells outbound. That does not make the underlying research wrong. It does predict which figures get republished, because a figure describing a problem the publisher solves travels further than one that does not.
The figures that come with a population attached
Salesforce's page is the strongest of the currently ranking set on disclosure, and it is worth saying why before quoting anything from it. It names its own research, states the participant count, carries a byline and carries a date. Its figures are drawn from what the page calls the "Seventh Edition State of Sales Report", and it describes the underlying population as "4,000+ sales professionals". The page is dated 3 February 2026.
That is three of the four properties, from the organisation that ran the survey itself. Four of its figures bear directly on an outbound motion, and all four sit on the Salesforce page as fetched on 2 September 2026.
On the length of the cycle you are feeding, the Salesforce page gives "57% of sales professionals now say the sales cycle is getting longer." On what buyers do with untargeted outreach, the same page gives "73% of B2B buyers actively avoid sellers who send irrelevant outreach."
On where a seller's week goes, the Salesforce page gives "Sales reps spend 60% of their time on non-selling tasks." On the cost of a crowded stack the same page gives "42% of sales reps feel overwhelmed by too many tools", and, separately, that page gives "Overwhelmed sellers are 45% less likely to attain quota."
Now the part a roundup would leave out. Every one of those figures is a survey answer given by sales professionals about themselves. The first is a perception of a trend rather than a measurement of cycle length. The third is a self-estimate of time allocation, which is among the least reliable things a person can say about their own week. The last pair is an association between two survey answers, with no direction of causation attached to it.
None of that makes the figures useless. It makes them evidence about what sellers believe, which is a real and quotable thing, rather than evidence about what is happening to cycle lengths.
The buyer-side figures, and the missing date
RAIN Group publishes a large sales statistics page and is unusually clear about provenance: the figures carry its own name throughout, and the page states that its data is drawn from "our expansive global sales skills study of over 1,000 sellers and sales managers". Population named, publisher and researcher the same organisation.
Two RAIN Group figures matter more to an outbound team than anything on the seller-behaviour pages, because they describe the buyer's side of a cold approach. Its page gives "82% of buyers accept meetings with sellers who reach out to them" and, separately, "82% of buyers look up providers on LinkedIn before replying to their outreach."
Read the first one carefully before carrying it anywhere. It is a statement about buyers who accepted at some point, not a reply rate, and it is not a prediction about your campaign. What it supports is the narrower and still useful claim that meeting acceptance from cold outreach is ordinary buyer behaviour rather than an exception. The second one has a direct operational consequence: a profile a prospect finds after reading your message is part of the message, which is why sender identity is not a cosmetic decision.
The gap on that page is the third property. RAIN Group describes the basis of the figures as "our expansive global sales skills study of over 1,000 sellers and sales managers" and gives no year for that study anywhere in the visible copy, while the embedded page metadata puts the last modification in July 2025. A reader arriving in 2026 has no way to tell from the page itself how old the research is.
That is the most common failure in this literature and it is not a small one. A figure whose year you cannot establish cannot be compared with a figure from a different year, which means it cannot be used to argue that anything has changed.
The best-specified population is the one you cannot read
The Bridge Group publishes research into sales development roles, and its page states the population precisely: "Metrics & motions across 350+ B2B companies", with the body confirming that "351 B2B companies participated" in the current edition, which it describes as the tenth round of the research.
That is a better population statement than anything in the roundups. The metrics themselves sit behind a download form.
The pattern is worth naming because it shapes what circulates. Research with a clearly defined population is expensive to run, which is why the organisations running it gate the results. Roundups are free. So the figures that travel furthest across the open web are the ones a compiler could reach without asking permission, and the ones with the tightest methodology are the ones a reader has to trade an email address for.
The practical response is not to disparage the gated reports. It is to notice that a roundup's coverage of a topic is a filtered sample of the research, filtered by accessibility as much as by relevance.
- Salesforce: Seventh Edition State of Sales Report
- RAIN Group: its own global sales skills study
- The Bridge Group: the tenth round of its own research
- Each is quoting itself rather than a compilation
- The citation chain is one link long
- Salesforce: 4,000+ sales professionals
- RAIN Group: over 1,000 sellers and sales managers
- The Bridge Group: 351 B2B companies
- All three name a count on the page
- None of the three states a margin of error
- Salesforce: the article is dated February 2026
- RAIN Group: no date in the visible copy
- The Bridge Group: an edition number rather than a year
- An edition number is not a date a reader can use
- A study year and an article date are different facts
Two live pages that show how far a date can drift
The date problem is easier to see on a specific page than to argue in the abstract, and two examples are sitting in this SERP.
Gong's cold calling statistics page carries both of its dates openly, which is more than most do. It reads "Published on: August 16, 2021" and "Last modified on: March 4, 2026". Same URL, same figures, one date five years newer than the other. A roundup citing that page in 2026 will inherit the modification date, and the analysis behind the figures is from the earlier one unless the page says otherwise.
The second example is starker. LinkedIn's URL for its State of Sales report still resolves, and the page it serves is titled "The State of Sales 2017 US Report". Anyone citing "LinkedIn's State of Sales" from that address today is citing a survey run nine years ago, and nothing about the citation would look wrong to a reader.
Neither publisher has done anything improper. Both are examples of the same structural fact: a URL is stable and a survey is not, and the citation carries the URL.
- 2017LinkedIn State of Sales, US
The report title still served at the URL people cite as current
- August 2021Gong cold call analysis published
The publication date printed on the page
- July 2025RAIN Group page last modified
Recorded in the page metadata, absent from the visible copy
- February 2026Salesforce State of Sales, seventh edition
Dated on the page, with the participant count stated
What to do with any of this
The reason the four properties matter is that a borrowed figure is usually being used to answer a question your own records could answer better.
Consider what a published statistic is for. You want to know whether your own number is normal. Your own number is measured on your actual population, over a period you chose, using a definition you control. A published one was measured on an undescribed population, in an unstated year, using a definition you cannot inspect. The comparison only means something if the two populations match, and you have no way to check.
Two internal comparisons beat every external figure and both are free. The same period last year, measured the same way, holds the offer and the market constant. One segment against another inside the same campaign holds everything constant except the variable you are testing. The conditions under which a cross-company comparison is defensible at all are set out in sales benchmarking, and the short answer is that the denominators have to match.
Where an external figure does earn its place is in an argument rather than in a measurement. Quoting Salesforce's finding that most surveyed buyers avoid sellers sending irrelevant outreach is a legitimate way to support a decision to narrow a list, because it is evidence about buyer attitude and it is being used as such. Quoting the same figure as a benchmark for your own reply rate is a category error.
The metrics an outbound team should be producing instead, with the denominator written beside each one, are in lead generation KPIs, and the arithmetic underneath the cost side is in cost per lead.
The short version
A sales statistic carries four facts or it carries none: who collected it, who was asked, when, and what was divided by what. The pages ranking for this term score well on the first two and poorly on the third.
Salesforce's page names its own seventh-edition research and a population of more than four thousand sales professionals, and dates itself, which makes its figures quotable as evidence about what sellers report. RAIN Group names its own study of more than a thousand sellers and sales managers and puts no date in the visible copy. The Bridge Group states the tightest population of the three, 351 participating companies, and puts the metrics behind a form, which is the pattern that decides what circulates freely.
Check the study's year rather than the article's. One page prints two dates five years apart from a single URL, and a nine-year-old survey still answers at an address people cite as current.
Then stop shopping for a yardstick. Your own last period and your own segment splits are measured on the population you actually contact, and they survive the objection that ends every argument about a borrowed benchmark, which is that you do not know what it counted. If you would rather see what that population looks like for your market before spending anything, we will build a campaign and show the reasoning.
Third-party figures above were quoted from the publishers' own pages as fetched on 2 September 2026. Verify current figures at the source before relying on them.
Sources: Salesforce sales statistics, whose own summary of the research behind its figures reads "The full report offers a comprehensive look at how 4,000+ sales professionals are navigating this shift."; RAIN Group sales statistics; The Bridge Group SDR research; Gong cold call statistics; LinkedIn Sales Solutions
Frequently asked questions.
Frequently asked questions- What makes a sales statistic worth quoting?
- Four facts have to be reachable from the page: the organisation that collected the data, the population it asked, the year the research was run, and what was divided by what where the figure is a rate. A page supplying the first two and omitting the third is common, and the omission matters because a figure whose year is unknown cannot be used to argue that anything has changed.
- Which sales research actually publishes its own data?
- Among the pages currently ranking, three quote themselves rather than a compilation. Salesforce draws on what it calls its Seventh Edition State of Sales Report and describes the population as more than four thousand sales professionals. RAIN Group cites its own global sales skills study of over a thousand sellers and sales managers. The Bridge Group states that 351 B2B companies participated in the current round.
- Why do statistics roundups repeat the same figures?
- A page whose headline is a count has to reach that count, and the cheapest route is to restate what is already on the page or lift what is already elsewhere. Compilation citing compilation also drops qualifiers at each hop, so a finding about a specific sample in a specific year arrives as a bare sentence with a percent sign attached to the citing page's date.
- Should I use published benchmarks to set targets?
- Not for target setting. A published figure was measured on a population you cannot inspect, in a year the page may not state, using a definition you cannot check, so the comparison tells you about the two definitions rather than the two programmes. Your own previous period and your own segment splits hold the offer, the market and the definition constant, which no external figure can.
About the author.
B2B cold email experts helping companies generate qualified leads through done-for-you outreach campaigns.
RevenueFlow Team
Explore more.
Ready to scale your outreach?
We build GTM engines that book real meetings. See the receipts.
Related articles.
Dark Funnel: A Measurement Gap, Not a Channel
The dark funnel names buyer conversations that carry no identifier to join on. Why better tooling cannot close it, and what the three sold remedies buy.
SCOTSMAN: Eight Criteria and a Three Way Score
SCOTSMAN adds Competition and Originality to the usual six, which turns qualification into a test of whether a real deal is one you can actually win.
Highspot Competitors: The Field Just Consolidated
Six names on a 2026 Highspot shortlist come from four owners. Two consolidations, both dated on the vendors' own pages, and no vendor publishes a rate.
AI-Guided Selling: What the Recommendation Rests On
A next best action is a prediction fitted to your own closed deal history, which is small, biased toward what was already tried, and typed in by the people it describes.
Descript Pricing: Every Editor Seat Adds to One Pool
Descript publishes its whole rate card. The number it does not publish is the denominator: two of three allowances are metered per Editor seat and pooled.
Revenue Acceleration: Three Problems, One Label
Revenue acceleration names three different problems at once. Only one of them, the time lost in the seams between teams, is what the cross-functional framing adds.