The last-click attribution model lies: which channel really brings in your B2B leads?

A marketing attribution model determines which channel gets the credit for a conversion. If you still work with last-click, you give all the credit to the last channel before the enquiry, usually direct, branded search or a retargeting ad. The channel that introduced the customer months earlier, a blog, LinkedIn or a webinar, gets nothing. In B2B with long sales cycles that is an expensive blind spot. In short: move away from last-click, look at the whole path to the deal and distribute your budget based on real contribution to revenue, not on the channel that happened to be in view last.

We are a Belgian B2B growth agency. We prefer to steer on customers and revenue rather than vanity numbers. Attribution is exactly the place where that distinction becomes visible, because the wrong model has you investing in channels that close deals that were already there anyway, and cutting on channels that fill the pipeline.

Work it out yourself: calculate what a single lead is worth with our free value-per-lead calculator.

What exactly is a marketing attribution model?

A marketing attribution model is the rule that decides how the value of a conversion is distributed across all the touchpoints that preceded it. A prospect first sees a LinkedIn post, reads a blog via Google two weeks later, comes back once more via a newsletter and finally requests a quote after a branded search. That is four touchpoints, one conversion. The question is: which channel deserves how much credit?

The common models:

• Last-click (last interaction): 100% of the credit goes to the last channel before the conversion. Simple, but it overrates everything at the bottom of the funnel.
• First-click (first interaction): all the credit goes to the channel that introduced the customer. The opposite problem.
• Linear: every step gets an equal share.
• Time-decay: steps closer to the conversion get more weight.
• Position-based (U-shaped): assigns 40% of the credit to the first interaction, 40% to the last, and the remaining 20% is distributed across the intermediate channels. This model recognises that both the introduction and the close matter.

No model is “true”. They are lenses on marketing attribution. But last-click is the most misleading lens, and coincidentally often the default that teams work with unconsciously.

Take the example path from this article: LinkedIn as the first contact, a blog and a newsletter in the middle, and a branded search as the last click. Position-based then recognises both the introduction and the close, while last-click only rewards that last step and sets the rest to zero.

Why does last-click attribution lie in B2B?

Because the last click is almost never the real work. In B2B with a sales cycle of weeks to months there are many intermediate steps. The prospect has known you for a long time before that last branded search. Last-click then gives the credit to direct traffic, branded PPC or organic search, while social, display or content marketing introduced the customer in the first place.

The result is a budget error. You see that branded search “converts” and pump more money into it, while those people had already found you. At the same time LinkedIn or your blog seems “not to work”, because they are rarely the last click, and you cut into precisely the channel that fills the pipeline. An honest look at your data analysis turns this around: you want to know which channel set the deal in motion, not which channel was standing by last.

If you want to set this up structurally, it comes down to proper conversion tracking and the right measurement setup. Our data analytics help often begins here: first make sure the paths are accurate at all, then compare models.

How do you see the full path to a lead?

In GA4 you do that with the attribution and advertising reports, in particular the Attribution paths report (in earlier versions known as the conversion paths). There you see, per conversion, the full sequence of channels that preceded it, not just the last one. You learn, for example, that a typical path looks like “Organic Social, followed by Organic Search, followed by Direct” instead of just “Direct”.

A second useful metric is the assist ratio per channel, readable in the old assisted-conversion reports. The interpretation (according to the Google Analytics documentation on these reports): divide the assisted conversions by the last-interaction conversions. If that ratio is close to 0, the channel mostly closes deals. If the ratio is far above 1, the channel is mostly an assistant that introduces the customer or keeps them warm. Both types are valuable, but you treat them differently: a closer you optimise for conversion, an assistant for reach and quality.

Practical: compare the same period under last-click and under position-based or data-driven attribution. If the credit shifts sharply towards your top-of-funnel channels, you know that last-click has given you a distorted picture until now.

Which attribution model should you choose for B2B?

No holy grail, but a workable line: use a model that weighs in the introduction, such as position-based or data-driven, and always read out the paths alongside it. For most B2B teams we see, position-based is a fair starting choice because it recognises both the first and the last contact. Data-driven is more powerful if you have enough volume, but with thin datasets it becomes a black box you can barely steer with.

More important than the exact model is what you do with it. The goal is to distribute budget based on real contribution to revenue, not on numbers. And that is where the biggest pitfall of all lies.

Why you should look at revenue, not at the number of leads

Because more leads does not mean more revenue. A telling example from our practice: a channel like organic delivered 2 leads, while social brought in 11 leads. At first glance social is the clear winner. Until you look at the lifetime value: the organic leads had a 5x higher lifetime value. Do the maths and those 2 “slow” leads are worth more than the 11 quick ones.

An attribution model that only counts conversions misses this entirely. That is why you always tie attribution to value: lay the deal size, the customer acquisition cost and the lifetime value next to your channels. Only then do you see whether a channel brings in cheap leads that never become customers, or expensive leads that become your best customers. This is exactly what we stand for: steer on customers and revenue, not on the number that looks nicest in a dashboard.

If you fix this in a standing report, it is smart to bring attribution and value together in one marketing dashboard, so you are not manually puzzling over paths every month. If you want to calculate the return of your site as a whole, read what your website really delivers and how you measure it.

What is a realistic step-by-step plan?

You do not have to overhaul everything at once. A workable order:

1. Check your current default model. Are you unconsciously working on last-click? Start there.
2. Set up reliable tracking. Without clean conversion tracking every model is guesswork.
3. Read out the paths. Open the Attribution paths report and see which channels are at the front of the path.
4. Determine the role per channel via the assist ratio: closer or assistant?
5. Tie revenue and LTV to your channels, not just numbers.
6. Redistribute budget based on contribution to revenue, and retest after a full sales cycle.

It is not a one-off exercise. Your sales cycle and your channel mix change, so you repeat this periodically. A small team that moves fast has an advantage here: you can shift budget the moment the data says so, without a quarter of approvals.

Frequently asked questions about marketing attribution models

Is last-click attribution always wrong? No. For a short, impulsive purchase last-click can be perfectly adequate. The problem arises with longer B2B trajectories with many touchpoints, where the model systematically wipes away the introduction and the intermediate steps.

What is the difference between first-click and position-based? First-click gives 100% of the credit to the channel that introduced the customer. Position-based distributes it: 40% to the first interaction, 40% to the last and 20% across the intermediate steps, so both introduction and close count.

How do I know whether a channel mostly assists or closes? Look at the assist ratio: assisted conversions divided by last-interaction conversions. Close to 0 means the channel mostly closes deals, far above 1 means it mostly assists. Both are valuable, you just adjust them differently.

Do I have to use the data-driven model? Not necessarily. Data-driven works well with enough volume, but becomes unreliable with little data. For many B2B teams, position-based is a fairer and more steerable starting choice.

Which model shows my real revenue contribution? No single model on its own. A model distributes credit across channels, but only when you tie deal value, CAC and lifetime value to it do you see which channel really brings in revenue.

Stop distributing budget on the wrong numbers

Last-click feels comfortable because it is simple, but in B2B it costs you money: you reward channels that did not make the deal and punish channels that fill your pipeline. Swap it for a model that weighs in the whole path, tie revenue and LTV to it, and distribute your budget on real contribution. Want to see that set up objectively, with tracking that is accurate and a model that fits your sales cycle? Plan your free intake.