Web Design
7 B2B Marketing Attribution Steps That Hold Up in 2026

Joshua Jackai
20 min

Published 18 November 2026 ยท Last updated and fact-checked November 2026
Short answer: B2B marketing attribution is how you give credit for pipeline and revenue to the channels and touches that produced them. Do it in seven steps: name the budget decision, capture the source on every lead, ask buyers, pick a model by sales cycle, attribute at account level, test before you move money, report monthly.
Picture a head of marketing at a 60-person US software firm (a composite, not a client). The CEO asks which channel to cut. HubSpot says organic search. Google Ads says paid search. LinkedIn says LinkedIn. The sales team says referrals. Every report is "right" by its own rules, and none of them can settle the argument.
I'm Joshua Jackai, and Demand Designer is my studio. I build B2B websites for US teams, and every build I ship has to answer one question for the client: where did this lead come from? These seven steps are the order I would set attribution up in if I joined that firm tomorrow. Each step names the research behind it, and where the research has limits, I say so.
Key takeaways
Capture the source on every lead before you buy software: Li and Kannan (2014) modelled individual touch data in the Journal of Marketing Research and found channel contributions "significantly different" from the metrics firms were using.
Treat any attribution report as observational: Gordon, Zettelmeyer, Bhargava and Chapsky (2019) compared methods against big Facebook field experiments, and observational methods "often fail" to recover the true effect.
Test branded search before you credit it: Blake, Nosko and Tadelis (2015) found in eBay experiments that brand keyword ads had no measurable short-term benefit.
Only test changes big enough to matter: Lewis and Rao (2015) ran 25 large field experiments covering $2.8 million in ad spend, and the median confidence interval on ROI was over 100 percentage points wide.
Price the tier before you plan the report: HubSpot's knowledge base (vendor, read 18 November 2026) limits deal create and revenue attribution reports to Marketing Hub Enterprise, which lists from $3,600 a month.
Step | What it fixes | Key number | Source type |
|---|---|---|---|
1. Name the decision | Reports nobody acts on | One decision and one date per report | My rule |
2. Capture the source on every lead | Leads with no origin | Individual touch records per customer | Peer reviewed |
3. Ask buyers how they found you | Dark touches no tool sees | Self-report checked against real exposure | Peer reviewed |
4. Pick a model by sales cycle | Arguing about the model | GA4 down to 3 models since November 2023 | Vendor |
5. Attribute at account level | Credit split across one buying group | $3,600 a month for HubSpot revenue attribution | Vendor |
6. Test before you move budget | Paying for sales you would get anyway | ROI intervals over 100 points wide | Peer reviewed |
7. Report monthly on pipeline | Lead counts that hide revenue | Last touch overincentivizes ad exposures | Peer reviewed |
What is B2B marketing attribution?
B2B marketing attribution is the practice of assigning credit for a lead, opportunity or closed deal to the marketing touches that came before it. In B2B it usually runs at account level, across a sales cycle of months and a buying group of several people, so it ties marketing to pipeline and revenue, not just form fills.
Consumer attribution asks which ad made one person buy. B2B attribution asks which mix of touches moved a whole company from stranger to signed contract. The buying group might include a marketing lead who read your blog, a finance lead who saw a LinkedIn ad and a CEO who heard your name on a podcast. Your CRM sees the first of those. Most tools see none of the last.
Lemon and Verhoef (2016) described the problem plainly in the Journal of Marketing:
"Customers now interact with firms through myriad touch points in multiple channels and media, and customer experiences are more social in nature," Lemon and Verhoef wrote in the Journal of Marketing.
That is why the seven steps below start with process, not software.
1. How to do B2B marketing attribution: name the budget decision first
The first step is writing down the decision attribution must support, such as which channel gets next quarter's budget, and the date it is due. In my experience, a report built without a decision measures everything and settles nothing. One decision and one date per report is the rule I start every setup with.
Good decision: "Do we keep LinkedIn Ads at $8,000 a month next quarter?"
Bad decision: "Show us marketing's impact"
Owner: the person who moves the budget, not the person who builds the report
Cadence: the decision date sets the reporting date
Skip it if: you have one channel and no budget choice to make
Most attribution projects I see start with a tool demo. The vendor shows a dashboard with colorful paths, the team buys it, and six months later nobody can say which decision it changed. Start the other way round. Write one sentence at the top of the brief: "This report exists so that we can decide X by date Y."
The decision tells you what grain of data you need. A channel budget question needs channel-level credit for opportunities. A content question needs page-level credit for first touches. A sales handoff question needs to know which touches happened after a rep got involved. Those are three different setups, and trying to build all three at once is how attribution projects stall.
Steward and colleagues reviewed more than sixty years of B2B buying models and found that the field moved from single transactions to relationships, networks and, most recently, journeys. Their review is about how academics model buying, not about marketing reports, so I use it for the shape of the problem: B2B buying is a group process over time, and no single touch tells the story.
If your company runs on EOSยฎ, put the decision on the Scorecard as a weekly number with an owner. (Demand Designer is not affiliated with EOS Worldwide.) A number with an owner gets fixed. A dashboard with no owner gets ignored.
"Looking to the future, we recommend that scholars examine five areas of research: the impact of technology, modes of customer and supplier interaction, decision-making approaches, tensions between internal and external communities, and B2B marketing analytics," Steward, Narus, Roehm and Ritz wrote in Industrial Marketing Management.

2. Capture the source on every lead: best step for teams with messy CRM data
Source capture means every form fill and booked meeting stores where the person came from: UTM parameters, the referring page and the first page they saw. Without it, every model downstream is guessing. Li and Kannan (2014) built their model on individual touch records across online channels such as display, paid search, referral and email.
Store on the contact: original source, UTM source, medium and campaign, landing page, referrer
Store on the deal: the contact's source at the time the deal was created
Hidden fields: forms read UTMs from the URL and a first-party cookie, not just the current page
Calendar links: pass UTMs into the booking tool, or meetings arrive as "direct"
Check monthly: the share of new contacts with source "offline", "direct" or blank
This is the step most B2B websites fail, and it is a website problem before it is a marketing problem. I audit sites where the demo form drops UTMs because the visitor read three pages before converting, and the form only reads the current URL. Every one of those leads shows up as direct traffic. The fix is a small script that saves the first UTMs in a cookie and writes them into hidden fields on any form. I wrote up the full setup in website lead tracking.
Li and Kannan worked with a firm's record of every touch consumers made before converting on its website, across channels such as display, paid search, referral and email. They modelled how earlier touches carried over and spilled into later visits and purchases. The study was in a consumer setting, not a B2B pipeline. Its premise still applies to any business: firms that only have aggregate numbers judge channels on those, and Li and Kannan found the channel contributions from individual touch data differed significantly from the metrics firms were using.
Booking tools are the other common leak. A prospect clicks a LinkedIn ad, lands on your pricing page and books through an embedded calendar. Unless the UTMs pass into the booking, the meeting lands in the CRM with no source. My demo form vs calendar booking post covers how to keep the source when you switch.
"These channels' relative contributions are significantly different from those found by other currently used metrics," Li and Kannan reported in the Journal of Marketing Research.

3. Add self-reported attribution: ask B2B buyers how they found you
Self-reported attribution is a required open-text question on your demo form, such as "How did you hear about us?" It catches the podcasts, peer referrals and private Slack groups that no tracking script sees. Treat it as a second signal, not the truth: Jerit and colleagues (2016) designed a test that checks self-reported exposure against known exposure.
Question: "How did you hear about us?" as open text, not a dropdown
Required: yes, on demo and contact forms only
Code the answers: weekly, into 8 to 12 buckets (podcast, referral, LinkedIn, search, event)
Compare: self-reported bucket against tracked original source, side by side
Where they disagree: that gap is your untracked channel
I like open text because a dropdown teaches buyers your channel list and they pick the first plausible item. Open text gives you answers like "my CFO sent me your pricing page after a podcast". That one answer touches three channels, and your tracking saw only the last click.
The research is about method, not a verdict. Jerit and colleagues built an experimental benchmark to compare what people said they saw with what an experiment had actually shown them. Their study was about news exposure, not buying journeys, and it does not tell you how accurate a demo-form answer is. What I take from it: a self-report is worth most when you check it against a benchmark, and in attribution your tracked source is the nearest benchmark you have. My rule is to use both. Tracked source tells you the path the browser took. Self-reported source tells you what the buyer remembers as the reason. When they disagree on a pattern, not one lead, you have found a channel worth testing.
Do not ask on every form. A newsletter signup does not need it. Ask where intent is high and the extra field is worth the friction. My form fields post covers that trade-off.
"Media exposure is one of the most important concepts in the social sciences, and yet scholars have struggled with how to operationalize it for decades," Jerit and colleagues wrote in Communication Methods and Measures.

4. Which B2B attribution model should you use? Pick by sales cycle
For most US B2B firms with a 30 to 120 day sales cycle, use a W-shaped model for the budget report and first touch for content decisions. An attribution model is the rule that splits credit across touches. Google Analytics 4 has offered only three models since November 2023, per Google's help center, so W-shaped means using your CRM.
First touch: credits the touch that created awareness; suits demand creation questions
Last touch: credits the touch before conversion; suits short cycles and direct response
Linear: equal credit to every touch; suits teams starting out
U-shaped: most credit to first touch and lead creation; suits inbound B2B
W-shaped: adds the opportunity-creation touch; suits sales-led cycles over 60 days
Data-driven: a statistical model fitted to your paths; needs volume most B2B firms lack
Here is what changed for US teams. Google's help center (vendor, read 18 November 2026) says first click, linear, time decay and position-based models "are no longer available as of November 2023" in GA4. That leaves data-driven, paid and organic last click, and Google paid channels last click. HubSpot's knowledge base lists nine models, including U-shaped, W-shaped and full path, but W-shaped only appears in deal create and revenue reports.
Shao and Li (2011) set out the case for data-driven multi-touch attribution at the ACM SIGKDD conference: give credit to more than the last ad, based on each touch's contribution. The catch for B2B is volume. In my judgment, a data-driven model fitted on 40 closed deals a quarter is mostly fitting noise. Rule-based models are blunt, but they are honest about being blunt.
My default for a US B2B firm with a 30 to 120 day cycle is W-shaped for the budget report and first touch for content decisions, with self-reported data beside both. Pick one model for the budget meeting and keep it for at least two quarters. Switching models every quarter makes every channel look like it moved. The matrix further down this page compares all eight options side by side.
"Rather than giving all the credit to the last ad a user sees, multi-touch attribution allows more than one ads to get the credit," Shao and Li wrote in the KDD proceedings.

5. Attribute at account level: best step for buying groups
Account-level attribution rolls every contact's touches up to the company and credits the deal, not each person. It stops one deal being counted as three wins because three people filled in forms. Purmonen, Jaakkola and Terho (2023) define the B2B journey around buying and usage center members.
Associate contacts to companies: automatic by email domain, checked weekly
Associate contacts to deals: every person on the buying side, not only the signer
Credit the deal: one opportunity, one credit pool
HubSpot tier: deal create and revenue attribution need Marketing Hub Enterprise (vendor)
Price check: Enterprise lists from $3,600 a month plus a $7,000 onboarding fee (vendor, read 18 November 2026)
The most common error I find in B2B CRMs is deals with one contact attached. The rep adds the person who booked the call and nobody else. The CFO who read your pricing page and the operations lead who watched the webinar never get linked, so their touches never count. The attribution report then over-credits whatever brought in the first contact.
Purmonen and colleagues pulled together research on customer journeys, B2B purchase processes and buying centers into one framework. It is a conceptual paper, not an experiment, so it tells you what to model rather than what the numbers are. Its point for attribution: the journey belongs to a group with goals, it loops back on itself, and it continues after the purchase.
Price check before you buy software for this. HubSpot's pricing page (vendor, read 18 November 2026) lists Marketing Hub Professional from $800 a month with three core seats and a $3,000 onboarding fee, and Enterprise from $3,600 a month with five core seats and a $7,000 onboarding fee. Contact create attribution works on Professional. If your budget question only needs first touch by channel, Professional plus clean contact-to-deal associations may be enough.
"Placing equal importance on the purchase and usage stages of the journey and acknowledging its iterative nature," Purmonen, Jaakkola and Terho wrote in Industrial Marketing Management.

6. Test before you move budget: best step before any big cut
An incrementality test switches a channel off, or holds out part of the audience, to measure what it adds beyond sales you would have got anyway. Attribution reports cannot do this alone. In Blake, Nosko and Tadelis's (2015) eBay experiments, brand keyword ads showed no measurable short-term benefit.
Best candidates: branded search, retargeting, any channel that "always wins" last touch
Method: geographic split, time-based on/off, or the ad platform's lift study
Length: long enough to cover your typical time from first touch to opportunity
Read: opportunities created, not clicks or form fills
Expect: wide uncertainty, so test big changes, not small ones
This is the step that changes budgets. Attribution reports are observational: they show which touches came before a sale, not which caused it. Ads are targeted, and Lewis and Rao call the selection bias that follows a crippling concern for observational methods. Gordon, Zettelmeyer, Bhargava and Chapsky (2019) compared observational methods with large randomized advertising experiments on Facebook and found they often failed to recover the effect the experiments measured.
Lewis and Rao (2015) show why tests are hard: across 25 large field experiments with US retailers and brokerages, the median confidence interval on ROI was over 100 percentage points wide. Those were consumer firms, most reaching millions of customers. I expect a B2B firm with a few hundred opportunities a year to get even less precise answers. So I only test changes big enough to matter, such as pausing branded search in half your US regions for a quarter. eBay is a consumer marketplace, not a B2B pipeline, so treat Blake and colleagues' result as a reason to test your own brand terms, not as proof they do nothing for you. The same study found that on non-brand keywords, new and infrequent users were positively influenced by ads, while frequent users who were not influenced took most of the spend.
Li and Kannan's model, a different method from a holdout test, also flagged cases in which retargeting may actually lower the chance of conversion. I covered a branded and non-brand Google Ads version of this test in should you cut non-brand Google Ads.
"Selection bias, due to the targeted nature of advertising, is a crippling concern for widely employed observational methods," Lewis and Rao reported in the Quarterly Journal of Economics.
"As an extreme case, we show that brand keyword ads have no measurable short-term benefits," Blake, Nosko and Tadelis reported in Econometrica.

7. Report monthly on pipeline, not leads: best step for the CEO meeting
The final step is one monthly report that shows pipeline and closed revenue by channel, using the model you chose in step 4, with self-reported source and test results beside it. Leads alone hide quality. In a game-theory model in Marketing Science, Berman (2018) showed last-touch attribution overincentivizes ad exposures.
Rows: channels, using the same buckets as your self-reported coding
Columns: opportunities created, pipeline value, closed won, cost, cost per opportunity
Beside it: self-reported share by channel and the latest incrementality test
Lag: report closed revenue on deals created at least one sales cycle ago
One line at the top: the decision from step 1 and the recommendation
The report I build for clients fits on one screen. It does not show every touch path. The CEO wants to know three things: which channels create opportunities, what each opportunity costs, and what we are changing next month. Everything else goes in an appendix.
Berman modelled advertisers buying across several publishers and found that the popular last-touch method pushes advertisers toward too many ad exposures, often lowering their profits, while a Shapley value method raised profits compared with last touch. It is a game-theory model of online ad markets, not a study of B2B pipeline reports. I read it as a caution: whatever rule you use to give credit also shapes where your team spends next.
Lag matters in B2B. If your average cycle is 90 days, this month's closed revenue came from touches that started last quarter. Show pipeline created this month and revenue closed on older deals, and label both. If you want a quick view of where your site is losing pipeline before any of this, my free Pipeline Leak Scorecard checks the forms, tracking and handoff in about fifteen minutes.
"The popular last-touch method is shown to overincentivize ad exposures, often resulting in lower advertiser profits," Berman reported in Marketing Science.

B2B attribution models compared: matrix by sales cycle
Use this matrix to pick a model for the budget report before you open any software. Each row shows what the model credits, which sales cycle it suits, where it breaks and where US teams can run it in 2026. The live page lets you filter the rows by cycle length and by tool.
Model | Credits | Best for | Where it breaks | Where to run it (read 18 November 2026) |
|---|---|---|---|---|
First touch | The first known touch | Content and awareness decisions | Ignores everything that closed the deal | HubSpot Professional and Enterprise |
Last touch | The touch before conversion | Short cycles, direct response | Over-credits branded search and retargeting | GA4, HubSpot, ad platforms |
Linear | Every touch equally | Teams starting out | Treats a newsletter open like a demo | HubSpot (GA4 removed it in November 2023) |
U-shaped | First touch and lead creation | Inbound B2B, cycles under 60 days | Ignores what happens after the lead | HubSpot |
W-shaped | First touch, lead creation, opportunity creation | Sales-led cycles over 60 days | Needs clean contact-to-deal links | HubSpot Marketing Hub Enterprise |
Data-driven | Fitted to your conversion paths | High volume, many conversions | Unstable on small B2B deal counts | GA4 |
Self-reported | What the buyer remembers | Dark touches: podcasts, referrals, communities | Memory, not a log: check it against tracked source | Any form tool |
Holdout test | The extra outcome a channel adds | Big budget cuts or increases | Slow, wide uncertainty | Geographic or time split |
Calculator: how a model splits credit on your own journey
The calculator in the live version of this page takes the touches on one real deal, in order, and the model you choose. It returns the share of credit each channel gets. Defaults are placeholders, not benchmarks, so type in a journey from your own CRM.
With an example journey of four touches (LinkedIn ad, blog post, webinar, branded search demo request), first touch gives LinkedIn 100% of the credit and last touch gives branded search 100%. Linear gives each 25%. Under the usual U-shaped convention (vendors vary), the first touch and the lead-creation touch get 40% each and the middle touches share 20%, so 10% each. Under W-shaped, with the lead and the opportunity both created at the branded search touch, branded search gets 60%, LinkedIn 30% and the blog post and webinar 5% each. The same deal, four answers. That is the reason step 4 tells you to pick one model and keep it.
What is the difference between B2B and B2C attribution?
B2B attribution credits accounts and deals over months, across a buying group; B2C attribution usually credits one person's purchase over days. B2B also has fewer conversions, so data-driven models have less to learn from, and offline touches such as calls and events matter more.
The practical difference is where the data lives. B2C teams can often run attribution inside the ad platform or analytics tool because the purchase happens on the site. B2B teams need the CRM, because the conversion that matters, a closed deal, happens weeks later in a sales conversation. That is why steps 2 and 5 are CRM hygiene tasks, not analytics tasks, and why I build source capture into the website itself.
Frequently asked questions
What is the best attribution model for B2B?
For most US B2B firms with a sales cycle over 60 days, a W-shaped model fits the budget report because it credits first touch, lead creation and opportunity creation. Pair it with self-reported source and at least one holdout test before cutting any channel. No single model is right everywhere.
What is self-reported attribution in B2B?
Self-reported attribution asks buyers how they heard about you, as a required open-text field on demo forms. It captures podcasts, referrals and communities that tracking misses. Code the answers into channel buckets and compare them with tracked source. Answers are memory, not a log, so treat them as one signal.
Does GA4 have multi-touch attribution?
GA4 has data-driven attribution, which spreads credit across touches using your own conversion data, plus two last-click models. Google removed first click, linear, time decay and position-based models in November 2023. For B2B deal credit, most teams still need their CRM, because GA4 does not see closed revenue.
How much does B2B attribution software cost?
It depends on the tool. HubSpot's revenue attribution needs Marketing Hub Enterprise, listed from $3,600 a month plus a $7,000 onboarding fee (vendor, read 18 November 2026). Contact create attribution comes with Professional, from $800 a month. Clean source capture on your forms costs less than either.
How long does it take to set up B2B marketing attribution?
Source capture and a self-reported field can go live in a week if the website is in good shape. Clean contact-to-deal associations take a month of rep habit. Useful reports need one full sales cycle of clean data, so plan for a quarter before trusting the numbers.
How I set up attribution at Demand Designer
Every B2B site I build for US teams ships with source capture already wired: first-touch UTMs saved in a cookie, hidden fields on every form and calendar, and original source written to the CRM contact and deal. That is steps 2 and 3 done on launch day, before any attribution software.
On a Figma to HubSpot rebuild for Snappy Kraken, shipped in 21 days, the new site contributed to a 30% rise in MQLs. My fixed-price builds start at $3,250 (USD). Whoever builds your site, run the free Pipeline Leak Scorecard first: it shows whether your forms are dropping the source before you spend a dollar on attribution tools.
Sources
Peer-reviewed research
Berman, R. (2018). Beyond the last touch: Attribution in online advertising. Marketing Science, 37(5), 771 to 792.
Blake, T., Nosko, C., and Tadelis, S. (2015). Consumer heterogeneity and paid search effectiveness: A large-scale field experiment. Econometrica, 83(1), 155 to 174.
Gordon, B. R., Zettelmeyer, F., Bhargava, N., and Chapsky, D. (2019). A comparison of approaches to advertising measurement: Evidence from big field experiments at Facebook. Marketing Science, 38(2), 193 to 225.
Jerit, J., Barabas, J., Pollock, W., Banducci, S., Stevens, D., and Schoonvelde, M. (2016). Manipulated vs. measured: Using an experimental benchmark to investigate the performance of self-reported media exposure. Communication Methods and Measures, 10(2 to 3), 99 to 114.
Lemon, K. N., and Verhoef, P. C. (2016). Understanding customer experience throughout the customer journey. Journal of Marketing, 80(6), 69 to 96.
Lewis, R. A., and Rao, J. M. (2015). The unfavorable economics of measuring the returns to advertising. The Quarterly Journal of Economics, 130(4), 1941 to 1973.
Li, H., and Kannan, P. K. (2014). Attributing conversions in a multichannel online marketing environment: An empirical model and a field experiment. Journal of Marketing Research, 51(1), 40 to 56.
Purmonen, A., Jaakkola, E., and Terho, H. (2023). B2B customer journeys: Conceptualization and an integrative framework. Industrial Marketing Management, 113, 74 to 87.
Shao, X., and Li, L. (2011). Data-driven multi-touch attribution models. Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 258 to 264.
Steward, M. D., Narus, J. A., Roehm, M. L., and Ritz, W. (2019). From transactions to journeys and beyond: The evolution of B2B buying process modeling. Industrial Marketing Management, 83, 288 to 300.
Industry and vendor sources (not peer reviewed)
Google Analytics Help. Get started with attribution. Read 18 November 2026. Vendor source.
HubSpot Knowledge Base. Understand attribution reporting. Read 18 November 2026. Vendor source.
HubSpot. Marketing Hub pricing. Read 18 November 2026. Vendor source.
Free workbook
Pipeline Leak Scorecard
Find the leaks between your traffic and your pipeline. Score your site in about fifteen minutes, with it open in front of you.
12 questions, 8 pages, 15 minutes. Free.







