It's the Buying Group, Stupid
Written by Robin Caller on May 20, 2026
Thirty years on, James Carville’s line still earns its keep. He hung it on the wall in the Clinton war room to stop bright people getting distracted from the only thing that mattered. B2B marketing needs the same sign on its own wall, and the line needs to be: it’s the buying group, stupid.
We are still grading individuals. Still scoring leads. Still passing single names to sales as if one person inside a buyer ever made the decision on their own. Forrester’s State Of Business Buying 2024 puts the average B2B buying decision at thirteen people, with eighty-nine per cent of purchases involving two or more departments.
Eighty-six per cent of those purchases stall along the way. Eighty-one per cent of buyers say they are unhappy with the vendor they finally choose. That is not a lead problem. That is a buying-group problem, and the lead-by-lead orthodoxy that demand gen was built on is no longer fit for what it is supposed to do.
Pretty much every B2B marketing team I speak to knows this. They will nod through it on a panel. They will quote Forrester at you over coffee. And then they will go back to the office and ask the same MQL question they have asked for fifteen years: which individual just crossed the threshold, and have we passed them to sales fast enough? The reason is not stupidity. It is that the operational model behind the MQL is still in place.
The platforms grade individuals. The handoffs are individual. The bonus structures are individual. And the buying group, the thing that actually signs the order, sits invisibly behind all of it.
That is what has to change. And it is changeable, because the signals are there. We just have to stop wasting them.
Identify the company, stitch the signals
The work splits into four jobs: identify the company, stitch the signals, resolve the people, rank the roles. The fifth, and the one most teams keep failing at, is to fix the operating model that stops the first four from sticking.
Start with the obvious one. When an unknown visitor lands on your pricing page, you do not know who they are, but you usually know roughly which building they came from. Reverse IP lookup is no longer experimental. It has been a vendor category for the best part of two decades and it is table stakes.
Match the IP, dedupe against your CRM, and you have moved from “anonymous visitor” to “someone at Lloyds is looking at our pricing page for the third time this week.” That is a different conversation. It is not yet enough to act on, but it is the first link in a chain that, properly built, identifies the buying group rather than the individual.
The next link is signal stitching. Most companies have all of the signals already and treat each one as its own little kingdom.
The systems do not lack data: web visits, form fills, email opens, content downloads, webinar attendance, third-party intent in 6sense or Bombora, first-party intent buried in the product, CRM activity in Salesforce or HubSpot, sales-engagement opens and clicks in Outreach or Salesloft. The buying group leaves a trail across all of them, and none of them talks to the others unless someone has done the work to make it happen.
That work is unglamorous. It does not earn anyone a stage at a conference. It is what separates the teams that can see a buying group from the teams that cannot.
Resolve the people, rank the roles
Once the signals are stitched, the next job is to resolve the humans. This is where most B2B operations fall over. Knowing that thirteen people at Lloyds are circling you is interesting. Knowing which thirteen is what you need.
The good third-party data sets, used properly, get you most of the way: org charts, named contacts, role and seniority, recent role changes, public professional activity, even budget signals. Stitched against your first-party engagement data, you start to see real shape: the senior practitioner who has read four of your articles, the procurement person who has just been added to the email thread, the CFO who has not engaged at all and is therefore the silent risk.
None of this is exotic technology any more. The pieces are off the shelf. The problem is that most teams have never asked the platform to do this job, because they were still optimising for the single MQL.
The hardest step, and the one most often skipped, is ranking within the group. Identifying thirteen people is necessary. Knowing which of them is the economic buyer, which is the champion, which is the technical evaluator, which is the procurement gatekeeper, and which is the silent blocker, is what makes the difference between a passable account-based programme and a useful one.
The signal that matters most for the economic buyer is usually absence. They have not visited the site. They have not opened the email. The deal will not close until someone in your sales team has earned five minutes of their time. Most lead-scoring models cannot represent absence as a signal, which is why they keep missing the buyer.
You will not get this perfectly right from a distance. You will get it directionally right, and you will refine it the moment a real conversation begins. That is fine. Directionally right is a hundred times more useful than individually scored. The whole point of the buying-group model is that you stop pretending you know exactly who is making the decision and start engaging the people who actually are.
The fifth job is the operating model
Now the part nobody likes hearing. Almost none of this is a technology problem. The tools have existed for years. Reverse IP is mature enough. Identity resolution is mature enough. Intent data is mature enough.
The CDP, the data warehouse, the activation layer: all in place, in some configuration, for any company that wants to do this properly. What is not in place is the operating model. Most B2B revenue teams are still measured on lead volume, lead conversion, and SQL throughput. The people writing those targets have not caught up with the people writing the buying-committee research. So the ops team builds the buying-group view, the sales team ignores it because they get paid on individuals, and the whole thing falls back to the MQL within a quarter.
If you want a one-line test for whether your team has crossed over, it is this. When a target account shows real engagement, can your team name the buying group, rank the members, point to the signals each one has left, and decide as a group whether to act? If yes, you have moved on from lead-by-lead. If no, you are still in the world Carville’s sign is shouting at.
A few things follow from this if you take it seriously. You stop running MQL as your primary metric and start running account-level engagement as the primary. You stop sending single leads to sales and start sending account briefs that name the group. You give your SDRs and AEs a view of every signal the account has produced, not just the form fill that triggered the alert. You build content for the group rather than for the persona, which usually means more practitioner-grade material and fewer buyer-stage downloads. You accept that the deal cycle is going to look messier on the dashboard, because the dashboard was built for the wrong unit of measurement in the first place.
The buying group meets generative search
There is one more uncomfortable wrinkle that the Forrester data forces. Almost ninety-five per cent of B2B buyers in that same 2024 report said they expected to use generative AI to support their purchase decision in the following twelve months. The buying group is not just thirteen people any more. It is thirteen people, each with a chatbot helping them prep, summarise vendor sites, and rank options before any human conversation takes place.
If your content is not visible to those models, you have lost a meaningful share of the group before they even read your homepage. The buying-group problem and the GEO problem are the same problem. The unit of audience is a committee. The unit of discovery is now partly machine. Both shifts point in the same direction, and both kill the single-lead MQL.
At LeadScale, this is the shift we work on every day. The pattern we see across enterprise programmes is uncomfortably simple: accounts that close are usually accounts where sales can name the group, and accounts that stall are usually accounts where they can only name one person.
Here is what that pattern looks like in shape. The named lead is a marketing director who downloaded the gated piece. The unmapped signals around the account tell a different story: a CFO who has never visited the site, a procurement team forwarding the pricing page internally, a technical lead attending two webinars without registering. The deal moves when sales stops working the named lead and starts working the silent CFO. Not because the marketing director is unimportant, but because the marketing director was never the buyer.
The engine we have built exists because the lead-by-lead model is breaking, and because the buying-group model is the only one that holds up against what Forrester’s data and most senior B2B marketers’ lived experience are both pointing towards. We validate leads, but the more interesting work is the layer above: who else at the company is in motion, what signals tie them together, how that picture is handed across to a sales team in a form they can actually use, and how it stays current as the group changes shape across a long deal.
None of this is hard to argue. It is hard to operationalise. The argument has been won on the slide deck and lost on the org chart. The teams that fix the org chart first are the ones who will catch the next decade of B2B growth. The rest will still be scoring individuals while their competitors are already in conversation with the group.
It’s the buying group, stupid. Put it on the wall.
