Buying Group Outbound: Why Prospecting One Person Per Account Stopped Working in 2026

Buying group outbound is the change most teams have not made yet

Every serious GTM report published this year says a version of the same thing, and almost nobody has rebuilt their outbound around it. B2B purchases are made by groups. Your prospecting is still aimed at individuals. That mismatch is the quiet reason a lot of well-executed outbound programs are producing meetings that go nowhere.

Salesforce's State of Sales report for 2026 landed this week and the headline everyone latched onto was the AI adoption number — 87% of sales teams now using AI in some form, with sellers reporting agents cutting prospect research time by roughly a third. That is a real finding. It is also the least interesting one, because it tells you how the work gets done rather than what work should be getting done.

The more consequential signal is buried in the operational research. LeanData's State of GTM Efficiency Report for 2026, based on more than 200 B2B leaders, found that more than half of companies have now shifted to opportunity-centric execution built around dynamic buying groups rather than individual lead records. Not account-based marketing in the 2019 sense. Buying groups: the actual set of humans who will collectively say yes or no, tracked as a unit, worked as a unit.

If your outbound still produces a lead, routes a lead, and books a meeting with a lead, you are running a 2019 motion into a 2026 buying process. Here is what the data says, why single-threading is now the single biggest silent killer in B2B pipeline, and a concrete playbook for buying group outbound you can start this quarter.

The data: nobody buys alone, and everyone prospects like they do

Three independent bodies of research converge here, and the convergence is what makes the case hard to argue with.

Committees got big and stayed big

Gartner's long-running figure puts the average B2B buying group at six to ten stakeholders. For anything above roughly $50,000 in annual contract value, Forrester and 6sense both put the median buying group at around 11 people, and 6sense's buyer research — drawn from more than 4,000 buyers across North America, EMEA and APAC — has the average sitting right around ten and holding steady year over year.

Ten people. Ten sets of priorities, budget anxieties, incumbent relationships and internal politics. And in most CRMs, the opportunity has one contact attached to it.

Buyers finish most of the decision before you arrive

This is where it gets uncomfortable. 6sense's research found that buyers delay contacting vendors until roughly two-thirds of the way through their journey, initiate that outreach themselves more than 80% of the time, and reach out first to the vendor they already intend to buy from. The shortlist forms early and hardens fast — 94% of buyers said they had already ordered their shortlist by preference before engaging any seller, and the vendor leading at that point wins 77% of the time.

Broader 2026 buyer research reinforces it: only about 17% of the B2B buying journey now involves a sales rep at all, with buyers completing well over half of their evaluation independently.

Put those two findings together and the implication for outbound is sharp. You are not trying to win a deal in the conversation. You are trying to be present, credible and known inside a group of ten people during a research window you cannot see, so that when the shortlist forms you are on it.

One contact cannot do that for you.

Multi-threaded deals win, and everyone knows it, and nobody does it

The win-rate data on multi-threading has been consistent for years and keeps getting rediscovered. Single-threaded opportunities close at a small fraction of the rate of opportunities with five or more engaged stakeholders — Foundry's research on multi-threading buyer journeys walks through the mechanics of why. On larger deals the lift from engaging multiple stakeholders is routinely measured in the double and triple digits.

And yet the majority of opportunities in the average CRM still have exactly one contact on them.

Here is the summary view:

Why single-threaded outbound fails in ways that look like other problems

The insidious thing about single-threading is that it never shows up as "single-threading" in your pipeline review. It shows up wearing four different costumes.

It looks like a messaging problem

You book a meeting with a manager who is genuinely interested. The meeting goes well. Then it dies. The obvious conclusion is that the pitch did not land, so the team rewrites the deck. But the manager was never the problem. They liked it, took it to a director who had a competing priority, and it stopped there. No amount of deck revision reaches the director.

It looks like a "no decision" problem

Somewhere between 40% and 60% of lost B2B pipeline goes to no-decision rather than to a competitor. That is what a buying group failing to reach internal consensus looks like from the outside. You did not lose to a rival. You lost to five people who never got aligned because only one of them ever heard your argument, secondhand.

It looks like a champion problem

Your contact changes roles, leaves, or gets reassigned, and the deal evaporates. With average tenure in GTM and ops roles where it is, this is not a rare event — it is a structural risk you are accepting every time you run a deal through a single relationship. A large share of stalled deals trace back to exactly this.

It looks like a timing problem

"They weren't ready." Sometimes true. Often it means the one person you reached was not the one experiencing the pain acutely enough to spend political capital. The person who was — the ops lead drowning in manual work, the CFO staring at a renewal — never entered your funnel because your list build returned one title per account.

Every one of those four failures is fixed at the top of the funnel, not in the deal review.

What buying group outbound actually looks like

Buying group outbound means the unit of prospecting is the account's decision-making group, not a person. Practically, that changes four things.

1. Your list build changes from titles to roles-in-a-decision

Stop building lists of "VP Sales at 50–200 person B2B SaaS." Start building groups. For a given ICP account, define the standing cast:

• Economic buyer — signs, cares about payback and risk
• Champion — feels the pain daily, will spend capital to fix it
• Technical or ops evaluator — will kill you on integration, security or data quality
• Blocker or incumbent owner — owns the tool you would replace
• End user — will decide whether it actually gets adopted

Five roles, not five titles. The titles that occupy those roles vary by company size and org design, which is exactly why "build me a list of VPs" produces such uneven results across a target list.

2. Your trigger changes from "matches ICP" to "someone in the group moved"

This is where signal-based outbound and buying group thinking fuse, and it is the highest-leverage change on this list. A buying group is not always in market. When it enters the market, someone in it does something observable:

• An ops lead posts on LinkedIn or Reddit about the exact problem you solve
• Two people from the same account view your profile or engage with your content inside a fortnight
• The company posts a job req that only makes sense if they are building the function you sell into
• Someone in the group comments on a competitor's post, or complains about the incumbent tool
• A funding round, a leadership hire, or a reorg changes who owns the budget

None of those are "the account fits our ICP." They are evidence that the research phase — the two-thirds of the journey you are otherwise blind to — has started. Catching a group during that window is the whole game, because after it, the shortlist is set.

The tooling problem here is real, and it is what most teams underestimate. Watching one signal type across a target list is manageable manually. Watching five signal types across five roles at three hundred accounts is not. This is the specific job Updately exists to do: capture warm intent signals across LinkedIn, Reddit and X, resolve them to the account, enrich and score the people around the signal against your ICP, and hand you a group rather than a lead. The point is not the automation. The point is that buying group outbound is arithmetically impossible to run manually at any useful scale.

3. Your sequencing changes from parallel spam to staged entry

The failure mode teams fall into when they first try multi-threading is blasting all five contacts with the same message in the same week. That is not multi-threading, it is a pattern that gets you flagged, internally and by the platform. Buying groups talk to each other. Five identical messages arriving on Tuesday is a story they will tell each other, and it is not a flattering one.

A staged approach works better:

• Week 1 — enter through the signal. Reach the person who actually did the observable thing. Reference it specifically and without flattery. This person has the highest reply probability because they have already declared the problem exists.
• Week 2 — go lateral, not up. Reach a peer in an adjacent function with a different angle on the same problem. Not a copy of message one. If the ops lead got the "manual work" angle, the demand gen lead gets the "wasted spend on unqualified pipeline" angle.
• Week 3 — go up, with context. Reach the economic buyer with the business case, and reference that you have been talking with their team — only if true. Executives forgive being second. They do not forgive being ambushed with a pitch they have no context for.
• Ongoing — stay visible to the whole group. Comments, useful content, relevant shares. This is the cheapest form of multi-threading and the one most teams skip because it does not show up in an activity dashboard.

4. Your metrics change from replies to group coverage

If you only measure reply rate, you will optimise toward single-threading forever, because one well-targeted message to one person has the best reply rate per send. It also has the worst deal outcome.

Track instead:

• Contacts engaged per opportunity — the single best leading indicator of win rate available to you
• Group coverage — what percentage of the defined roles you have touched at each active account
• Time from first signal to second contact engaged — how fast you expand once you are in
• Single-threaded opportunity rate — the share of your pipeline sitting on one relationship, which should be a number your VP looks at weekly

That last one is the metric to introduce first if you introduce only one. Most teams discover it is well over half their pipeline, and the discovery alone changes behaviour.

Where AI actually helps, and where it makes things worse

Given the Salesforce finding that 87% of teams are now using AI, it is worth being precise about this, because the naive application of AI to outbound makes buying group problems worse rather than better.

Where it helps: research at group scale. Understanding five people's context — their role, their tenure, what they have posted, what the company just announced, what tool they currently use — is exactly the work that used to make multi-threading economically infeasible. Sellers in the Salesforce data report agents cutting research time by around a third, and research is the binding constraint on multi-threading. That is a genuine unlock.

Where it hurts: generating five variations of the same message. If AI just lets you send five times as many templated messages, you have industrialised the exact behaviour that makes buying groups distrust vendors. Growth Unhinged's 2026 State of AI in GTM research found the pattern clearly — near-universal AI adoption, but impact concentrated narrowly in intent-driven outbound and market intelligence rather than in raw message generation. The teams getting results are using AI to know more, not to send more.

The test is simple. If your AI spend increased send volume, you probably made things worse. If it increased the number of distinct, well-researched humans per account you can credibly reach, you made things better.

A four-week implementation you can actually run

You do not need a replatform. You need a different definition of a target.

Week 1 — Audit. Pull every open opportunity and count contacts engaged in the last 30 days. Calculate your single-threaded rate. Present it. This is the political work and it is the part that makes the rest possible.

Week 2 — Define the group. For your top ICP segment, write down the five roles. Get sales and marketing to agree on them. Map the three most common title variants for each role at your typical account size.

Week 3 — Wire the signals. Pick three signal types you can genuinely monitor — profile views, post engagement on your and your competitors' content, and hiring signals is a strong starting trio. Route them to the account, not the person, so that a signal from anyone in the group lights up the whole group.

Week 4 — Rebuild one sequence. Take your best-performing sequence and rewrite it as a staged group entry: signal contact, lateral peer, economic buyer, with genuinely different angles for each. Run it against 50 accounts. Compare contacts-engaged-per-opportunity against your control.

Four weeks, no new platform required to start, and the measurement is clean.

Takeaways

• Buying group outbound is the motion the 2026 data supports. Committees of six to ten decide; more than half of companies have already moved to opportunity-centric execution around dynamic buying groups, per LeanData's research.
• Single-threading is the disguised cause of your no-decision losses. It masquerades as a messaging problem, a timing problem and a champion problem. Measure your single-threaded opportunity rate and it stops hiding.
• The research window is the only window. Buyers contact vendors two-thirds of the way through, with 94% having already ranked their shortlist. Signals are the only visibility you get into the phase that actually decides the deal.
• Enter through the signal, expand laterally, then go up. Five identical messages in one week is not multi-threading. Staged entry with genuinely different angles per role is.
• Point AI at research, not at send volume. The teams seeing real impact used AI to make multi-threading affordable, not to make templating faster.

The uncomfortable summary is that most outbound programs are optimised for a metric — reply rate on individual sends — that is actively hostile to how B2B purchases get made. Fixing that does not require more activity. It requires accepting that the account, not the person, was always the unit that mattered, and that the only way to work an account properly is to know when its buying group woke up.

If you want to see what that looks like operationally, Updately turns signals across LinkedIn, Reddit and X into scored buying groups rather than isolated leads — which is the part of this playbook that does not scale by hand.