Inquiry Attribution Models: Choosing Between First-Touch, Last-Touch, and Linear

Comparing Three Attribution Models and Choosing by Sample Size

Inquiry Attribution Models: Choosing Between First-Touch, Last-Touch, and Linear — and Their Limits

Bottom line first: There is no perfect attribution model, only the trade-off that fits your current stage. Choose by annual closed deals: under 100 deals a year, use first-touch to find new traffic sources; 100 to 300 deals, use last-touch to optimize the conversion path; over 300 deals, use linear attribution to optimize your channel mix. A counterintuitive point: B2B deal cycles run 3 to 6 months, so last-touch systematically overrates retargeting ads — the buyer’s final visit is usually a returning visitor, not a new channel bringing them in.

1. What Attribution Is Actually Solving

Attribution models answer one question: when an inquiry finally closes, who gets the credit? A buyer might first find you through a Google search, read your blog post the second time, come back through LinkedIn the third time, and place an order by typing your domain directly the fourth time. Of those four steps, which contributed the most? Different attribution models give different answers.

This sounds academic, but it directly affects budget allocation. If attribution gives all the credit to last-touch, you will pour budget into retargeting ads; if it all goes to first-touch, you will chase new channels relentlessly. Pick the wrong model and your budget goes wrong. But here is the counterintuitive point: no attribution model is “right” — they are all approximations of the truth. What you need to pick is the approximation that best fits your current business stage.

Attribution model	Credit allocation	Core assumption	Typical bias
First contact	100% to the first visit	The first impression matters most	Underrates the role of the conversion stage
Last-touch	100% to the last visit	The final push matters most	Overrates retargeting and branded keywords
Linear attribution	Split evenly across every touchpoint	Every step contributes equally	Does not distinguish awareness from decision stages
Time decay	The closer to the deal, the higher the weight	Recent touchpoints have more influence	Ignores the long-term value of early nurturing

Each of these three common models fits certain scenarios and has its own blind spots. The point is not to find the “most accurate” one, but to know which direction your chosen model skews you, then use another model as a cross-check. For the underlying technology behind attribution tracking, see How to calculate lead attribution correctly, and for the Click ID system, see that article.

2. First-Touch: The Choice for the Cold-Start Stage

First-touch attribution gives all the credit to the buyer’s first visit to your site. This model fits the cold-start phase of a business: you’re still hunting for new traffic sources and don’t know which channels actually bring in inquiries. At this stage, first-touch helps you figure out “which channel brought the person in the first time.”

But its limits are obvious: it assumes the first touch decided the sale and ignores every nurturing action in between. A buyer finds you through Google Search, takes one look, and leaves. Three months later they come back via LinkedIn and finally convert. First-touch credits Google with the sale, but that LinkedIn return visit was the real closing touch.

Best-fit stage	Annual deal sample	Decision use	What not to use it for
Cold start, 0 to 1	<100 deals	Find new traffic sources	Don’t use it to optimize landing pages
New channel testing	New channel <50 deals	Judge whether a channel brings new customers	Don’t use it to cut budgets for older channels
Content awareness phase	Mostly blog traffic	See which content drives first visits	Don’t use it to evaluate sales follow-up

Example: you launched your own store this year and closed 60 deals. Using first-touch attribution, you find 40% of deals first came from Google organic search, 20% from LinkedIn, and 15% from direct visits driven by industry trade shows. This tells you Google is your biggest source of new customers, so your budget should go to SEO first. That decision is reliable at a sample of 60 deals. But if you use this data to optimize landing page conversion rate, you’re wrong: first-touch doesn’t reflect landing page performance.

3. Last-touch: the default, but it distorts over long B2B cycles

Last-touch attribution is the default in most CRMs and GA4: it gives all the credit to the buyer’s last visit. This model assumes “the final step, the closing touch, matters most.” That’s reasonable for short ecommerce cycles, but it distorts long B2B cycles.

Here’s the counterintuitive part: with a 3 to 6 month B2B sales cycle, last-touch systematically overvalues retargeting ads and branded search. Buyers near the end of their decision often search your brand name directly to come back, or return through retargeting ads you planted earlier. Last-touch credits the branded search or retargeting, but what pulled them into the awareness stage might have been a blog post six months ago.

Buyer path	Last-touch attribution	Touchpoint that actually contributed most	Direction of bias
Google Search → blog → LinkedIn → branded search, direct visit, close	100% credited to branded search	Google Search and blog	Overvalues branded search, undervalues SEO
Trade show → email nurturing → retargeting ad → close	100% credited to retargeting	Trade show and email nurturing	Overvalues retargeting, undervalues trade shows
YouTube video → website → quote page → close	100% credited to quote page	YouTube video	Overvalues quote page, undervalues video

When should you use last-touch? When your conversion path is stable and you want to optimize the final step. Say you close 100 to 300 deals a year and your channel mix is clear, and you need to know which landing page or quote page drives the most closed deals. Last-touch is a reasonable way to optimize landing pages here, but don’t use it to decide whether to keep investing in SEO, because it systematically understates SEO’s contribution. For landing page optimization, see Landing Page Conversion Rate Optimization.

4. Linear Attribution: Optimizing the Mix at Maturity

Linear attribution splits credit for a closed deal evenly across every touchpoint on the path. The model assumes each step contributes equally. It’s simple, but at a mature stage of the business it’s closer to reality than the two single-touch models above. It requires you to track the buyer’s full journey, from first visit to last, recording every touchpoint in between.

The limitation of linear attribution is that it doesn’t distinguish between the awareness stage and the decision stage. A first visit to read a blog post and a final visit to submit an inquiry carry the same weight in a linear model. In reality, those two steps contribute very differently. More refined models like time decay or U-shaped attribution (40% to first and last, 10% to each in between) can solve this, but they demand more tracking data.

Annual deal sample	Primary model	Reference model	Main decision
<100 deals	First contact	Last-touch	Find new channels
100 to 300 deals	Last-touch	First contact	Optimize the conversion path
>300 deals	Linear attribution	Time decay	Optimize the channel mix
>500 deals and multilingual	U-shaped or data-driven	Linear attribution	Fine-tune budget allocation

Example: you close 400 deals a year across Google organic search, Google Ads, LinkedIn, email nurturing, and retargeting. Linear attribution puts Google organic search at 25%, LinkedIn at 22%, email nurturing at 18%, and retargeting at 15%. Compare that with last-touch, where retargeting drops from 40% to 15% — meaning you had been overvaluing retargeting, and what actually brought buyers in was SEO and LinkedIn. At this point you should shift part of the retargeting budget into LinkedIn content.

5. Primary Model Plus Reference Model: Don’t Mix Decisions

The most common mistake is mixing numbers from multiple attribution models: first-touch says SEO contributes the most, last-touch says retargeting does, and then you don’t know which to believe. The right approach is to pick one primary model to make decisions and use the other as a reference — the primary model decides budget allocation, and the reference model flags where the primary model may be leading you astray.

Decision	Which model to use	Why
Should you open a new channel?	First contact	See how many first visits the new channel brings
Should you optimize landing pages?	Last-touch	See which landing page closes the most deals
How should you split budget across mature channels?	Linear attribution	Look at full-path contribution
Should you cut the retargeting budget?	Check first-touch as a reference	Retargeting is naturally inflated under last-touch
Is SEO actually worth it?	Check first-touch as a reference	Last-touch attribution systematically undervalues SEO

Another common mistake is using a complex model when your sample size is too small. If you close 50 deals a year and run linear attribution, each touchpoint gets too few data points and the numbers are pure noise. When your sample is small, a simple model is actually more reliable. For the technical side of attribution tracking, see How to calculate lead attribution correctly, and for GA4 event setup, see GA4 event setup.

FAQ

Which attribution model should a B2B site use?

Pick based on your annual deal volume. Under 100 deals a year, use first-touch to find new traffic sources. Between 100 and 300, use last-touch to optimize your conversion path. Over 300, use linear attribution to optimize your channel mix. There is no perfect model, only the trade-off that fits your current stage. Pick one primary model to make decisions and another as a control, and don’t mix them.

Why does last-touch break down for long B2B cycles?

Because B2B deals take 3 to 6 months to close, buyers often search your brand name directly or come back through retargeting ads you planted earlier. Last-touch credits that deal to the brand term or the retargeting ad, but what first brought them into awareness might have been a blog post six months ago. So last-touch systematically overvalues retargeting and brand terms while undervaluing SEO and content.

When is first-touch attribution the right fit?

It fits the cold-start 0-to-1 phase and periods when you’re testing new channels. It answers the question “which channel brought the person in the first time.” Use it to find new traffic sources when you close under 100 deals a year. But don’t use it to optimize landing page conversion rates, and don’t use it to cut budget from established channels — it assumes the first touch decides the deal and ignores every nurturing action in between.

What are the limits of linear attribution?

Linear attribution splits credit evenly across every touchpoint on the path, with no distinction between the awareness stage and the decision stage. A first visit to read a blog post and a final visit to submit an inquiry carry the same weight in a linear model, but their actual contribution is very different. A more precise U-shaped model (40% to the first and last touches) or time decay can solve this, but they demand better data tracking and only become reliable once you close over 300 deals a year.

What happens if you pick the wrong attribution model?

Your budget goes wrong. If attribution credits everything to last-touch, you’ll pour budget into retargeting ads; if it credits everything to first-touch, you’ll chase new channels relentlessly. But the channel that brought the buyer in and the one that closed the deal may be different. The right approach is to let your primary model drive decisions and your control model flag the direction of bias — for example, check first-touch to see SEO’s contribution before cutting your retargeting budget.

Can you use complex attribution with a small sample size?

No. If you only close 50 deals a year, linear attribution gives each touchpoint too few data points, and the numbers are pure noise. When your sample is that small, a simple model is actually more reliable. A good rule of thumb: under 100 deals, use first-touch; 100 to 300 deals, use last-touch; over 300 deals, then use linear attribution. At the small-sample stage, focus on getting your path tracking in place, not on calculating precise attribution.

Final thoughts

The biggest trap with attribution models is hunting for the “most accurate” one. But attribution is fundamentally an approximation of the truth, and no model is the truth. First-touch makes you overvalue new channels, last-touch makes you overvalue retargeting, and linear attribution won’t distinguish early from late touchpoints. Your job isn’t to find the perfect model, it’s to know which direction your chosen model will skew you.

Three steps to put this into practice: first, pick your primary model based on annual deal volume: under 100 deals, first-touch; 100 to 300 deals, last-touch; over 300 deals, linear attribution. Second, pick a comparison model specifically to flag which way your primary model is biased. Third, don’t mix them when making decisions: budget allocation follows the primary model only, and the comparison model is only there to challenge it. Do these three steps and you won’t blow your budget in the wrong place because you picked the wrong attribution model.

One line worth remembering: attribution isn’t about right or wrong, only about fit. The model you choose determines which channel your budget flows to, and knowing where it will skew you matters more than finding the “most accurate” model.