B2B attribution often gives all the credit to the final touch, like praising the scorer while forgetting the passes that set up the shot. The person who fills out the last form gets the points, while the rest of the buying group disappears from view. That incomplete picture becomes costly when a company is trying to win larger, more senior deals from two audiences at once, like builders who adopt the product and executives who sign for it. The signals that show who really contributed are usually already in the stack. They just aren't in the same place.

The buying group is bigger than the person who filled in the form

When Palo Alto Networks changed what its BDRs pass to sales, from individual leads to buying groups, they saw a 17x increase in pipeline progression, deal sizes that more than doubled, and a 17% greater closed-won rate from opportunities with a buying group attached¹.

The buying group is larger for a reason. Forrester cites typically 13 internal stakeholders and 9 external influencers², all getting involved at varying points of the buyer journey. A lead-level report only follows one of those individuals.

For companies that grow through technical adoption, the gap is wider still. The person with the densest data trail, the developer who found and adopted the product, is rarely the person who signs the contract. The executives who do sign, such as the CIO, the VP of Engineering, procurement and the CISO reviewing security posture, leave far fewer signals. The most visible user has the least purchase authority, and the most powerful buyer is close to invisible.

The fix is to report at an account level, and roll every contact, anonymous visit and ad impression up to the account it belongs to.

Most of the decision is made before your first reported touch

6sense's 2025 Buyer Experience Report found that buyers end up choosing from their Day One shortlist 95% of the time. They contact sellers about 61% of the way through their journey, and the average cycle runs to 10.1 months³. That fits the wider pattern that most potential buyers aren't in-market at any given moment, but will be at some point⁴.

So the shortlist is largely set before your first reported touch. Form-fill reporting only picks up the journey once the buyer is well advanced, and credits whoever is there at the end.

For companies with bottom-up adoption, usage itself can help build the shortlist. Months of product activity, documentation visits and community engagement may come before anyone fills in a form, and last-touch reporting cannot credit any of them. The account then looks as though it appeared from nowhere. Account-level reach and engagement reveal whether you were already influencing the buying group before anyone raised a hand, making the otherwise dark part of the funnel visible.

The signals exist, but they live in different places

Gartner's 2025 Tech Marketing Benchmarks Survey found that 48% of technology marketers ranked measuring overall ABM success among their top three challenges, and 34% named measuring attribution⁵. Measurement is the most-cited problem in the survey.

The problem compounds when signals live in separate places. One tool reports intent, another reports touchpoints, and the CRM reports revenue. For product-led companies there's often a fourth, product analytics, which tends to sit apart from the marketing automation and CRM systems where attribution has to run. Each is accurate about its own slice, but none can tell you what happened in the account. Most stacks are also built up in pieces, by different teams at different times, so some tools are only partly configured, or not used at all, and their signals never reach the place where attribution happens. Multi-touch attribution at contact level doesn't fix this. It spreads credit across one person's touches and still ignores the rest of the group.

Six steps from last touch to account-level attribution

The move from last-touch to account-level attribution is operational work. Here is how each step looks when signals are split between technical users and executive buyers.

  1. Take stock of the signals. List which tools are switched on, which are feeding data into the account view, and which are licensed but dormant. Treat this as an inventory. Stitching can't join signals that were never captured.

  2. Define the account. Agree the target account list and ICP, then map the buying group roles inside each account. Attribution can't be account-level if nobody has agreed what the account is. For companies with a large technical user base, this starts with a practical question: which individual users belong to which enterprise accounts? Several people from the same company domain adopting the product independently is an expansion signal, but only if someone is tracking it.

  3. Stitch the identity. Tie product usage, community activity, site visits, ad impressions, events and CRM contacts to one account record. Modern reporting platforms can match anonymous visits to companies and link a person's history once they identify themselves. Company-level matching is generally easier than person-level, and match rates vary by setup, so ask any provider for evidence from a comparable data set.

  4. Measure reach and engagement before pipeline. Track the percentage of target accounts reached, the percentage engaged, how many buying group contacts have engaged, and which channels deliver reach you wouldn't otherwise have had. Product-qualified signals belong here too, meaning usage that points to account-level expansion intent. Senior buyers who know you for something else will look like quiet accounts long before they look like opportunities, so pipeline is the wrong first measurement. Track how many senior contacts have engaged inside accounts your technical users already reach, and how many target accounts show engagement from more than one role. ForgeX lists buying group contacts engaged as a core account-based engagement metric, and warns that setting pipeline expectations too early puts cross-functional buy-in at risk⁶. If pipeline is the first success metric, the program may be judged before its foundations have produced reliable numbers.

  5. Credit influence to opportunities and revenue. Run multi-touch at account level, broken out by channel and campaign, across opportunities created, won and lost. Technical activity, community engagement and executive-side content consumption all roll up to the same account view. Treat the model as a guide, because some influence never leaves a trace. Self-reported "how did you hear about us" data and incrementality tests help fill that gap.

  6. Close the loop with sales. Build two lists: accounts sales is working that marketing hasn't touched, and accounts marketing has warmed that sales hasn't called. Put them in a prioritization view sales will open without being asked to. For long, multi-stakeholder cycles, add a third lens: accounts that combine strong technical adoption with emerging executive engagement.

Measurement that changes while the program is still running

If the answer is already in the stack, the job is getting the systems to answer together. In our piece on how ABM teams build an AI advantage that survives platform transition, we argued that MCP, the Model Context Protocol, is becoming the standard way AI agents connect to the platforms teams already run. Applied to reporting, that connection turns measurement from a monthly pack into a live feed that can change the program as it runs. It also gives the tools a stack already has a second life. A platform nobody logs into can still answer questions through an agent, provided the data in it is sound and the definitions are agreed. Some examples include:

  • Accounts surface themselves. An agent spots accounts outside the target list that are surging on relevant content, or where several technical users have started using the product, and proposes them for the program with their engagement history attached. A person approves.

  • Sales gets the whole buying group. When an account crosses an engagement threshold, the account owner receives a message listing every engaged contact, their roles, what they read, and which roles are still silent.

  • Coverage gaps close automatically. Accounts sales is working but marketing isn't reaching are added to campaign audiences the same day.

  • Budget follows reach. Channels delivering incremental reach get a recommended increase each week, and the ones duplicating it lose budget.

None of this needs new strategy, but it needs clean definitions behind the data. An agent connected to numbers without a shared definition of an engaged account will produce confident answers that nobody can trust. Whoever looks after the stack is best placed to set those definitions.

Account-level attribution is operational work, and it needs continual monitoring. The strategy decision has usually been made already, whether that's moving upmarket, selling to a new audience or both. What remains is the measurement to support it, so your team can see the whole play, not just the scorer, and refine the tactics that will drive the next win. The game has changed. The scoreboard is next.

Sources: 1. Forrester, How Palo Alto Networks Drives Revenue With Buying Groups, 2026. 2. Forrester, The State of Business Buying, 2026, press release, January 2026. 3. 6sense, The B2B Buyer Experience Report 2025. 4. LinkedIn B2B Institute and Ehrenberg-Bass Institute, How B2B Brands Grow, the 95-5 rule. 5. Gartner, 2025 Tech Marketing Benchmarks Survey: Account-Based Marketing Insights, May 2025. 6. ForgeX and DemandBase, Account-Based Measurement & Reporting Certification.