Disclosure: This article is published by Datamagnet. Vendor and tool claims are self-reported or drawn from public documentation, retrieved 2026-09-17, unless otherwise noted.
Signal Attribution: How to Prove Which Trigger Actually Closed the Deal
Ask your best rep why a deal closed and you'll usually get a story about the final demo. Ask what actually happened and the real answer is messier: a job-change alert three months earlier, a funding announcement, two LinkedIn comments, and then the demo. In 2025, 6sense found buyers log an average of 16 interactions with the vendor they eventually choose, over a 10.1-month buying cycle (6sense, 2025 Buyer Experience Report). Crediting the last one and ignoring the other 15 isn't attribution — it's a guess with a CRM field attached. This guide walks through how to log every signal that touches a deal, weight it honestly, and build a system that can actually answer "what closed this?"
TL;DR
B2B buyers rack up 16+ interactions with their chosen vendor and often pick a winner before sales ever gets a call, per 6sense's 2025 Buyer Experience Report. Last-touch tracking misses a 36% "dark social" gap (HockeyStack, 2025). Signal attribution fixes this by logging every trigger as a timestamped event instead of crediting the final touch.
Key Takeaways
- In 2025, 6sense found B2B buyers interact with their eventual vendor 16 times over a 10.1-month cycle, and 94% of buying groups rank a preferred vendor before ever talking to sales (6sense, 2025).
- Last-touch attribution misses a 36% "dark social" gap between what buyers say influenced them and what standard tracking tools actually record (HockeyStack, State of Revenue, 2025).
- Deals sourced through a known signal, like a past champion's job change, see 114% higher win rates and 54% bigger deal sizes (UserGems).
- 76% of CRM admins say less than half their organization's data is accurate, which is why signal attribution needs timestamped events, not CRM notes (Validity, State of CRM Data Management 2025).
- Build attribution in four layers: log every signal with a timestamp, tag it to a trigger type, correlate it to close date, then weight credit. Never award it all to one touch.
Why Can't You Just Credit the Last Touch?
You can't credit the last touch because most of the buying decision already happened by the time it fires. Forrester's 2026 State of Business Buying report puts the typical buying decision at 13 internal stakeholders and 9 external influencers, with that number climbing further for complex or strategic purchases (Forrester, State of Business Buying 2026). A single "last touch" can't represent a decision that twenty-plus people influenced at different moments. HockeyStack's analysis of over 8,500 B2B buyers found a 36% gap between what buyers said influenced their purchase and what standard first-touch and last-touch tracking actually recorded (HockeyStack Labs, State of Revenue, 2025). This dark social gap shows why last-touch attribution alone misses most of a deal's real momentum, which typically builds during research and engagement phases before formal sales involvement.
Isn't it strange that teams will argue for hours over which channel gets marketing credit, then let sales attribution default to "whatever the rep remembers typing into Salesforce"? The table below shows why the two models produce such different answers to "what actually closed this deal."
| Dimension | Last-Touch Attribution | Signal Attribution |
|---|---|---|
| Credit assignment | 100% to the final touch before close | Weighted across every logged signal in the lookback window |
| Data source | CRM activity log or rep memory | Timestamped events: job changes, funding, engagement, hiring signals |
| Dark social visibility | Blind to it — a 36% gap between reported and tracked influence (HockeyStack, 2025) | Captures pre-sales research and engagement as scored signals |
| Buying committee coverage | Represents one person's final action | Reflects input from the 13+ stakeholders typically involved (Forrester, 2026) |
| Auditability | Not re-testable — no record of earlier touches | Weights can be re-tested against win-rate data over time |
Most teams don't have a bad attribution model — they have no attribution model, just a CRM field someone fills in from memory. "Last touch" isn't really a methodology. It's what's left over when nobody logged anything earlier in the deal.
Citation capsule: HockeyStack's analysis of over 8,500 B2B buyers revealed a 36% gap between what buyers said influenced their purchase and what standard first-touch and last-touch tracking actually recorded (HockeyStack Labs, State of Revenue, 2025). This dark social gap demonstrates why last-touch attribution alone fails to capture most of a deal's real momentum, which typically happens during research and engagement phases before formal sales involvement.
If you want a real answer to "what closed this deal," you need signal-based selling infrastructure that logs events as they happen, not weeks later from memory.
How Many Signals Actually Touch a Deal Before It Closes?
More than you're currently tracking. B2B buyers now average 27 total vendor interactions before purchase, up from just 17 in 2019, according to Forrester's longitudinal buyer research (Forrester, cited via Sword and the Script, 2021). That's a mix of digital research and human conversations, tallied across every vendor a buyer considers, not just the one they eventually choose. Meanwhile, 6sense found that 94% of buying groups rank a preferred vendor before a salesperson ever gets a call, and those buyers still average 16 direct interactions with that vendor over the following months before signing (6sense, 2025 Buyer Experience Report). Put those two data points together and the pattern is clear: most of a deal's real momentum builds in research and engagement activity that your CRM never logs as a formal "touch."
Citation capsule: B2B buyers now average 27 total vendor interactions before purchase, up from 17 in 2019 according to Forrester's longitudinal research. Meanwhile, 6sense found that 94% of buying groups rank a preferred vendor before sales gets involved, with those buyers averaging 16 direct interactions with that vendor over the following months. This gap means most buying momentum occurs outside traditional sales touchpoints.
For a deeper look at catching these early moments, see how job-change signal APIs surface buyer activity long before a demo request ever lands in your pipeline.
Why Can't Your CRM Answer "Which Signal Closed This Deal"?
Your CRM can't answer that question because most of what's actually stored in it isn't reliable enough to build a real attribution model on top of. Validity surveyed 602 CRM administrators in 2025 and found 76% believe less than half of their organization's data is accurate and complete, and 37% said they've lost revenue directly because of it (Validity, State of CRM Data Management 2025). The average organization loses 16 sales opportunities per quarter to CRM inaccuracy alone. That makes notes-based attribution inherently unreliable for real decision-making: if a rep's memory of "what worked" is the input, the output is a guess dressed up as a data point, not a system you can audit, re-weight, or trust when a bigger deal is on the line.
Watching teams try to reconstruct attribution after the fact is a familiar pattern. A deal closes, leadership asks what drove it, and someone opens the account and starts guessing from stale activity logs and half-written notes. By then the job change that started it all happened five months ago, and nobody tagged it as anything.
The fix isn't a cleaner spreadsheet. It's logging signals as structured, timestamped events the moment they fire — a job change, a funding round, a LinkedIn engagement spike. That beats relying on a rep to remember and manually note them weeks later, and it's the difference between a CRM field and an attribution system.
Citation capsule: A Validity survey of 602 CRM administrators found that 76% believe less than half their organization's CRM data is accurate and complete, with 37% reporting direct revenue loss from bad data. The average organization loses 16 sales opportunities per quarter to CRM inaccuracy, making notes-based attribution inherently unreliable for decision-making.
How Do You Log and Tag Every Buying Signal?
You start by logging every signal as a timestamped event the moment it fires, then tagging it to a specific trigger type — skip either step and you're back to guessing which one mattered. Signal attribution runs on four layers total; this covers the first two.
1. Log every signal as a timestamped event, not a note. Every job change, funding round, hiring spike, or engagement moment needs a hard timestamp the moment it happens. Datamagnet's Create Signal endpoint lets you set up monitors for job changes, new posts, and engagement events. Each fire gets recorded automatically instead of depending on someone noticing it.
2. Tag each signal to a specific trigger type. A job-change signal, a company engagement signal, and a keyword engagement signal all mean different things about buyer intent. Store the trigger type alongside the timestamp so you can later ask "which trigger type correlates most with closed-won," not just "did something happen."
How Do You Correlate and Weight Signal Credit?
You correlate every signal to the close date inside a defined lookback window, then split credit across all of them instead of awarding it to one touch. These are layers three and four of the same framework.
3. Correlate signals to close date inside a time window. Decide on a lookback window first — 90 days is a reasonable starting point for most B2B cycles. Then pull every signal that fired on an account or contact inside that window before close. This is where timestamp-free CRM notes fail and structured signal logs win: you can actually query "what fired in the 90 days before this deal closed" instead of relying on memory.
4. Weight credit across signals instead of picking one winner. A simple, defensible starting model: split credit evenly across every signal inside your lookback window. Then adjust weight upward for trigger types your win-rate data already favors. If champion job-change signals beat generic content downloads for your team, weight them higher — but keep every signal in the record so the model can be re-tested later.
Route the whole pipeline through webhooks so signals land in your CRM the moment they fire. Use the Champion Tracker cookbook as a working template for the job-change piece specifically.
<!-- [ORIGINAL DATA] -->Teams that move from note-based tracking to timestamped signal logging typically find their "unknown source" bucket in closed-won analysis shrinks fast. Deals that used to get filed under "inbound" or "referral" start showing a specific job-change or engagement signal sitting 60 to 90 days upstream of the close date.
Citation capsule: Signal attribution requires four layers: logging every signal as a timestamped event at the moment it fires, tagging each to a specific trigger type, correlating signals to close date within a defined lookback window, and weighting credit across all signals instead of picking one winner. Starting with a 90-day lookback window and even credit distribution allows teams to refine weights once win-rate data shows which trigger types perform best.
What Does Signal Attribution Prove About Champion-Sourced Deals?
It proves that a specific, trackable trigger — not vague "relationship strength" — is doing measurable work. UserGems' research on champion tracking looked at deals involving a past champion, someone a rep already knew who moved to a new company. Those deals saw 114% higher win rates, 54% bigger deal sizes, and closed 12% faster than deals without a known champion in play (UserGems, Hidden Gems study). That's not a soft "relationships matter" claim. It's a specific, attributable trigger — a job-change signal firing on a known contact — that correlates with three measurable outcomes at once: win rate, deal size, and cycle time. That's exactly the kind of claim signal attribution is built to prove with a timestamped record, instead of assuming it from a rep's gut feel.
If you're logging job-change signals as timestamped events per the system above, you can pull this exact comparison for your own pipeline instead of taking someone else's benchmark on faith. Datamagnet's Person Engagement Signal endpoint extends the same logic to people who aren't yet a "champion" but are actively engaging with your content or team on LinkedIn.
Citation capsule: UserGems' research on champion tracking found that deals involving a past champion who moved to a new company saw 114% higher win rates, 54% bigger deal sizes, and closed 12% faster than deals without a known champion. This specific, attributable trigger—a job-change signal on a known contact—correlates with three measurable outcomes simultaneously, proving that signal attribution reveals impact missed by generic relationship claims.
What Comes Next for Signal Attribution?
What comes next is attribution that runs continuously in the background instead of getting reconstructed after a deal closes. More of the buying journey now happens through trackable digital signals — job changes, engagement, funding events. Against that backdrop, the manual "ask the rep what worked" model looks increasingly like guesswork dressed up as data.
The shift already underway is from "explain this deal after it closes" to "score every account in real time against the signals already proven to correlate with wins." That's less about buying a bigger MTA platform and more about making sure every signal gets logged, tagged, and time-stamped in the first place — the raw material any attribution model, simple or sophisticated, actually depends on.
Why Should You Start Logging Signals Before You Try to Attribute Them?
You can't attribute what you never logged. Buyers touch a vendor 16 times before choosing them, and buying decisions now involve 13 internal stakeholders plus 9 external influencers (Forrester, State of Business Buying 2026). Add in that 76% of CRM admins don't trust their own data, and the fix becomes obvious: structured, timestamped signal capture, not a smarter spreadsheet formula. Layer in trigger tagging, a close-date correlation window, and weighted credit, and "which signal closed this deal" stops being a guess. See how a live LinkedIn Signal API tracks every trigger as it fires — set up your first signal monitor this week.
Frequently Asked Questions
What is signal attribution in B2B sales?
Signal attribution tracks which specific buying signals — job changes, funding events, engagement spikes, hiring surges — correlate with a deal closing, instead of crediting only the final touch. It requires timestamped signal logs, not CRM notes written from memory after the fact.
How is signal attribution different from marketing attribution?
Marketing attribution typically tracks channel-level touches like ad clicks and email opens. Signal attribution tracks buyer-side events happening independent of your marketing, like a target contact changing jobs or a company raising funding. Both feed the same goal: proving what actually influenced a deal, based on 6sense's finding that buyers average 16 interactions with their eventual vendor (6sense, 2025).
Which attribution model works best for signal-based selling?
A weighted multi-signal model outperforms single-touch models for signal-based selling, since Forrester's 2026 research shows the typical buying decision involves 13 internal stakeholders and 9 external influencers (Forrester, State of Business Buying 2026). Start by splitting credit evenly across every signal in a defined lookback window, then adjust weights once your own win-rate data shows which trigger types perform best.
Can you automate signal attribution without a full MTA platform?
Yes. Most of the work is capturing signals as timestamped events at the source. Register job-change and engagement signal monitors with webhook delivery, and every trigger lands in your CRM automatically. That gives you the raw data a full multi-touch attribution platform would otherwise need imported manually.
How far back should you look for signals before a deal closes?
Ninety days is a reasonable starting window for most B2B cycles, though it should scale with your sales cycle length. Buyers now log up to 27 total vendor interactions before purchase (Forrester, via Sword and the Script, 2021), so a 30-day window risks cutting off signals that mattered earlier in the journey.
Sources
- 6sense, 2025 Buyer Experience Report, retrieved 2026-09-17, https://6sense.com/science-of-b2b/buyer-experience-report-2025/
- Forrester, State of Business Buying 2026, retrieved 2026-09-17, https://www.forrester.com/press-newsroom/forrester-2026-the-state-of-business-buying/
- HockeyStack Labs, State of Revenue, retrieved 2026-09-17, https://www.hockeystack.com/lab-blog-posts/state-of-revenue
- Validity, State of CRM Data Management in 2025 (via PR Newswire), retrieved 2026-09-17, https://www.prnewswire.com/news-releases/validity-releases-state-of-crm-data-management-in-2025-report-revealing-disconnect-between-data-quality-and-ai-implementation-302499899.html
- UserGems, Hidden Gems Study, retrieved 2026-09-17, https://www.usergems.com/blog/our-champions
- Forrester B2B buyer interaction data, cited via Sword and the Script, retrieved 2026-09-17, https://www.swordandthescript.com/2022/05/b2b-sales-interactions/
- Datamagnet, Create Signal endpoint, retrieved 2026-09-17, https://docs.datamagnet.co/api-reference/endpoints/signal-create
- Datamagnet, Webhooks, retrieved 2026-09-17, https://docs.datamagnet.co/api-reference/webhooks
