B2B buying group size has no single answer. It runs from two or three people on a small deal to more than twenty on a complex one, and the real number for any given deal depends far more on company size and purchase risk than on any industry average you will find quoted in a deck.
That distinction matters because most of the published numbers get misused the same way. A rep reads "the average buying group has 10 people," assumes their deal matches it, and either over invests in a three person purchase or under invests in a twenty person one. Marketing teams make the same mistake upstream, sizing a campaign's stakeholder outreach off a single borrowed statistic instead of the account list actually in front of them. The research below explains where these averages come from, why they disagree with each other, and what actually predicts the size of the group sitting across from you.
What the Research Says About Group Size
The most cited starting point is Harvard Business Review's 2015 finding that an average of 5.4 people formally signed off on each B2B purchase. A year later, HubSpot cited CEB research putting the number at 6.8 stakeholders, a 25 percent jump from 5.4 in just eighteen months, according to CEB's Brent Adamson.
By 2026 the count has grown again, but by how much depends on who is counting. Forrester's State of Business Buying, 2026 report puts a typical purchase decision at 13 internal stakeholders plus 9 external participants, 22 people in total, with the number climbing further on complex or AI related purchases. That is roughly four times the 2015 figure.
Not every source agrees that groups have grown that far. Influ2's 2026 buyer survey found half of buying groups had only 2 to 4 people, and 42 percent had 5 to 9, with groups of 10 or more showing up only at the largest companies. Four separate counts, four different answers, and all of them are defensible once you see what each one is actually measuring.
Lay the four studies side by side and a decade long pattern shows up even where the exact figures disagree: 5.4 in 2015, 6.8 a year later, somewhere between 2 and 9 for most deals in 2026, and 22 once a research firm counts the external voices a buyer consults before a vendor is even contacted. The direction is consistent even when the headline number is not, which is itself a useful finding. Groups have not shrunk in a decade, and nothing in the research suggests they are about to.
Why the Published Numbers Disagree
The gap is not noise. It is methodology. HBR's 5.4 and HubSpot's 6.8 count only the people who formally sign off, the inner ring of approvers. Forrester's 22 widens the lens to include external participants, analysts, consultants, and peer references a buyer consults before a vendor is even contacted. Influ2 surveyed working sales and marketing teams directly about their most recent live deals, which skews toward the mid market accounts that make up most real pipelines rather than the flagship enterprise purchases that dominate analyst commentary.
Influ2's own writeup makes this explicit, noting that Forrester's average "reflects the enterprise high end rather than a typical deal." Both numbers are real. They are just answers to slightly different questions, one about the biggest deals a research firm tracks, the other about the deal in front of an average rep this week.
None of this makes the averages useless, but it does mean quoting a single number without its definition is close to meaningless. "10 people" could mean approvers only, or it could mean approvers plus every analyst call a buyer made before reaching out. Ask which one before you plan a sales motion around it.
The Variable That Actually Predicts Size
Industry and product category explain surprisingly little of the variation between deals. Two variables do almost all the work: company size and how much the purchase costs.
Put those two together and the published averages stop being contradictory and start being a map. A 40 person startup buying a point tool sits near the HBR end, 3 to 5 people. A 3,000 person enterprise evaluating a cross functional platform sits near the Forrester end, well into the teens once external influencers are counted. Most B2B sellers spend their day somewhere in the HubSpot to Influ2 range, 5 to 9 people, which is also the band most reps consistently underestimate because only 2 or 3 of those people ever reply to an email directly.
Two concrete accounts make the pattern easier to use. A 60 person marketing agency buying a 200 euro a month prospecting seat will usually have one approver, the person whose budget covers it, and maybe a second who checks that it plugs into the existing stack. A 1,200 person manufacturer evaluating a platform that touches sales, finance, and IT will have a reviewer from each of those functions before procurement is even looped in, plus whoever signs the annual contract once the case is made. Neither account is unusual. They are just different points on the same two variable scale, and treating them with the same outreach plan wastes effort on one and leaves the other half mapped.
Why an Average Is the Wrong Tool for Your Deal
None of the four numbers above tell you anything about the account sitting in your pipeline right now. An average is built to describe a population, not to predict one member of it, and a single deal is one member. Treating "10 people" as a target to hit on every opportunity leads to the same mistake in both directions: chasing five extra contacts on a three person purchase that never needed them, or closing a twenty two person decision having spoken to two.
The more useful question is not "what is the average," it is "who, specifically, is in this group." The full guide to mapping a buying group covers the five roles that keep appearing across the research, from the economic buyer down to the person who can only say no, and how to find each one on a live account in about thirty minutes. For a narrower look at the three roles that decide most outcomes, specifically the economic buyer, the champion, and the blocker, this breakdown of buying group roles walks through how to tell them apart when none of them volunteer the information.
The same dynamic appears at the top end of the range Forrester describes. On the largest, most scrutinized accounts, the group does not just get bigger, it also gets slower and more procedural, with a different playbook for enterprise buying committees once legal, security, and multiple business units are all seated at the table.
How to Size Your Own Buying Group This Week
Skip the industry average and build a number from the two variables the research actually supports.
- Start with headcount. A company under 200 employees rarely produces a group above 5. Above 1,000, assume 10 or more unless the purchase is trivially small.
- Weigh the blast radius. Count the departments that would notice if the tool disappeared tomorrow. Each one is a likely stakeholder, whether or not they have spoken to you yet.
- Name names, not a count. A number without identities does not help a forecast. Automatic role classification with a confidence score turns a guess about group size into a labeled list of actual people, so the team is correcting a draft instead of starting from zero.
The published research is useful for calibrating expectations before a deal starts. It is not a substitute for finding out who is actually in the room on the account you are working today. If your pipeline review still asks "how many stakeholders" and accepts a single number as the answer, that number is worth less than a short list of names, because a count tells you nothing about who still needs to hear from you this week.
