You spend on five channels and one deal closes. Which one gets the credit? Most founders guess, and the guess gets more expensive every quarter you keep making it. Here is what B2B marketing attribution actually is, and the groundwork before you touch a single model.
THE SHORT VERSION → Attribution assigns credit for a closed deal to the touchpoints that produced it. Skip it, and every platform in your stack will happily claim that credit for itself, which is how budget gets moved on vibes instead of evidence. → Only 52% of senior marketing leaders can prove marketing’s value internally (Gartner, 2024). Attribution is the mechanism that gets you into that 52%. → The reason it’s gotten harder, not easier: Forrester’s own buying study found the average number of buyer touchpoints jumped from 17 to 27 in the two years spanning the pandemic. That’s 27 things to allocate credit across, not three. → Three definitions have to be solid before any model matters: what counts as a touchpoint, what the full conversion path looked like, and what attribution window you’re measuring inside. Get these wrong and every report downstream inherits the error. → The most common failure at this stage is defaulting to last-click because it’s what the tools show out of the box. It rewards whatever closed the deal and erases whatever built the pipeline, which is the opposite of useful for a budget decision. → This post covers the basics only. The next post breaks down all seven attribution models and the exact math behind each one. |
Here’s a conversation that happens in almost every B2B company between $1M and $20M ARR. The founder asks which channel produced the last five closed deals. Marketing points at a dashboard showing LinkedIn and organic search. Sales says half those buyers mentioned a referral or a podcast episode on the first call, and none of that shows up anywhere. Everyone is technically right, and nobody can answer the actual question. That gap, between "we spent money and deals closed" and "we know which spending caused which deals," is what B2B marketing attribution exists to close.
It looks like a reporting problem. It’s a decision problem wearing a reporting problem’s clothes. Every dollar in next quarter’s budget is a bet on a channel, and without attribution you’re placing that bet on whichever channel argues loudest in the room. Gartner’s 2024 survey found only 52% of senior marketing leaders can prove marketing’s value and get credit for its contribution to business outcomes. Half the room cannot defend its own budget when asked. Attribution is how you get to a defensible answer instead of a confident one, and B2B companies feel this more acutely than most, because the sales cycle is long enough and the buying group large enough that a confident guess and a correct answer rarely look the same.
What it actually is
Strip away the tooling, and attribution is one plain idea: assigning credit for a closed deal to the marketing and sales touchpoints that led to it. A buyer read a blog post in March, attended a webinar in April, took a demo in May, and signed in June. Attribution decides how much of that June signature belongs to March, how much to April, how much to May.
It needs a formal exercise because every platform in your stack is already answering that question, badly, and in its own interest. The ad platform says the ad gets credit. The CRM says the last form fill gets credit. Sales says the demo gets credit, because that’s the meeting they ran. None of them are lying. Each is reporting on the one touchpoint it can see and presenting a partial view as the whole picture.
Every platform in your stack will tell you it deserves the credit. Attribution is what you build when you stop believing any one of them. |
Why it’s getting harder, not easier
Attribution becomes unavoidable at a specific point: when you outgrow the ability to remember how each client found you. Below $1M ARR, most founders genuinely know, because there are a dozen clients and they spoke to all of them personally. Past that point, and especially once the referral network that got you here saturates, the number of simultaneous channels and the number of people on each buying committee both grow past what memory can hold.
The buyer’s side of the journey is making this worse at the same time. Forrester’s own B2B buying research, tracked across surveys in 2017, 2019, and 2021, found the average number of buyer touchpoints held nearly flat through 2019 - 16 to 17 - then jumped to 27 in the two years spanning the pandemic. It wasn’t a steady climb. It was a plateau, then a spike, and the spike is the number that matters now. Twenty-seven touchpoints is not a list you eyeball in a CRM and intuit an answer from. It’s a dataset that needs a defined method, which is the entire discipline of B2B marketing attribution: turning a sprawling, multi-person buying process into a credit assignment you can actually defend.
The three definitions that have to be solid first
Before any model is worth discussing, three concepts need fixed, unambiguous definitions. Get these wrong and every attribution report downstream inherits the error, regardless of how sophisticated the model on top of it looks.
What counts as a touchpoint?
Any buyer interaction you can actually capture: a blog visit, an ad click, a webinar registration, a demo, a case study download. The word "capture" matters most here. A conference conversation nobody logs is a real-world touchpoint that doesn’t exist in your data. Attribution only works with what gets recorded, which is why tagging discipline outranks model sophistication.
What is a conversion path?
The ordered sequence of touchpoints one buyer moved through before converting. Order matters as much as contents; the same three touchpoints in a different sequence tell a different buying story. A checklist of "channels that were involved" is a much weaker artifact than a full, ordered path.
What is an attribution window?
The length of time before a conversion during which a touchpoint still earns credit. Thirty days is the default on most ad platforms, built for short consumer purchase cycles. It’s close to useless for a B2B deal that took five months and eleven touchpoints, because everything from month one silently falls outside the window and gets treated as if it never happened.
A worked example, with real numbers
A $4M ARR services firm closes three deals worth $180K combined in one month. Deal one, $40K: the buyer clicked a LinkedIn ad two weeks before signing. Deal two, $90K: the buyer read the newsletter for four months, then requested a demo. Deal three, $50K: a referral partner intro, preceded by a case study download from the site a week earlier. Credit last-touch only, and the CRM says LinkedIn drove $40K, the newsletter drove nothing despite four months of engagement on the $90K deal, and the referral partner gets full credit for $50K the case study likely assisted. A founder acting on that report cuts newsletter spend and doubles LinkedIn, exactly backwards from what the $90K deal, half the total revenue in the example, actually needed to close. That is the cost of skipping attribution: not a vague inefficiency, a specific wrong reallocation of real budget toward the smallest deal in the set. |
The two mistakes that cost the most
The default failure isn’t picking the wrong model, it’s not picking one at all and inheriting whatever the CRM shows out of the box, almost always last-click. Last-click hands 100% of credit to whatever the buyer touched immediately before converting and erases what built the demand in the first place. It looks precise because it’s a single clean number, and that precision is exactly what makes it dangerous: a founder reviewing a report showing "LinkedIn: 8 deals, Organic: 1 deal" reads that as fact, not as an artifact of which model happened to be switched on.
The second mistake compounds the first: trusting each platform’s self-reported numbers and adding them up. Every platform claims credit for buyers who touched more than one channel, so four or five tools summed together will regularly report well over 100% of actual closed revenue. That isn’t a bug in any one tool. It’s what happens when nobody assembles one system of record across all of them, which is the actual job attribution does. Neither mistake is a failure of effort. Both are the natural result of trusting whatever number is easiest to pull instead of building the one number that’s actually correct.
// where mrktrs fits Most firms are one model away from a defensible number. We build the tagging, the CRM structure, and the model selection that turns "we spent money and deals closed" into a number you can act on. mrktrs owns growth for B2B firms between $1M and $20M ARR. |
Frequently asked questions
Do I need attribution software to start, or can I do this manually?
Manually, if deal volume is low enough to review by hand: export closed-won deals, pull each contact’s touchpoint history, tag the source yourself in a spreadsheet. This gets tedious past roughly 15 to 20 closed deals a month, usually the point a dedicated tool earns its cost. Start manual anyway. It teaches you what a clean touchpoint looks like before you pay to automate it.
What should I actually show my CFO or board about attribution?
Not a full conversion-path breakdown. Show one number that changes a budget decision: revenue or pipeline by channel, under one clearly named model, tracked quarter over quarter. A board wants a consistent, improving story, not the underlying methodology. Save the model debate for internal marketing conversations, not the boardroom.
How long does it take to get attribution working cleanly?
Tagging and CRM hygiene: a few weeks. Trusting the output: longer, usually one full sales cycle, because enough closed deals need to pass through the corrected system before the numbers are credible rather than a first guess. For a five-month B2B cycle, budget a full quarter before making a spending decision off the new data.
What are the different types of marketing attribution models?
The main models are first-touch, last-touch, linear, U-shaped, W-shaped, full path, and algorithmic attribution, and HubSpot’s own attribution documentation lays out the exact credit-split math behind several of them. Which one fits depends on your sales cycle length and how many people sit on the buying committee. We break down all seven, with worked numbers, in the next post in this series.
Can I trust the attribution numbers my ad platforms show me?
Not on their own. Every platform reports only what it can see and is structurally biased to claim credit for itself. Add up what Google, LinkedIn, and your CRM each report and the total will regularly exceed 100% of actual closed revenue, because the same buyer gets counted more than once. Treat platform-reported attribution as one input, never the final number.
Attribution sits underneath every other number in this series. Your CAC is only as trustworthy as the attribution behind it, and your pipeline coverage is only actionable once you know which loop is actually filling it. Once the basics here hold, the real work starts: choosing which model fits how your buyers actually decide.
MRKTRS We run engineered demand systems for B2B founders who’ve outgrown referral-only growth and guesswork budget calls. Attribution, pipeline coverage, and sales process all live inside the same system, not three separate reports that disagree with each other. |