Perfect tracking still won't tell you the truth if you're using the wrong attribution model. Here's how last-click, first-click, linear, and data-driven models each tell a different story from the same data — and how to choose.
TL;DR
Even with perfect tracking, your attribution model decides which channels get credit — and most platforms default to last-click, which systematically undervalues upper-funnel channels. Here’s how the main models work, a worked example showing how the same journey tells different stories, and how to pick the right one for your funnel.
Attribution modeling is the method marketers use to assign credit for a conversion across all the touchpoints a customer interacted with before completing a purchase or sign-up. Fix your tracking, and you’d think the picture finally gets clear. It doesn’t — not entirely. Even with every conversion tracked perfectly, the attribution model you’re using decides which channel gets the credit for that conversion. And most ad platforms default to a model that quietly distorts the picture in one direction, every time. According to Gartner’s 2025 Marketing Analytics Survey, 73% of marketing leaders say attribution remains their top measurement challenge despite improved tracking infrastructure.
That model is last-click. Here’s what it does, why it’s misleading, and what the alternatives actually show when you point them at the same data.
“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”
— John Wanamaker, Retail Pioneer (the attribution problem, stated 100+ years before digital marketing)
A real customer journey usually involves multiple touchpoints before a conversion. Someone sees a YouTube ad, a few days later clicks a retargeting ad on Instagram, then a week after that searches the brand name on Google and converts.
An attribution model is the rule that determines how credit for that conversion is split across those touchpoints. Different rules produce wildly different “performance” numbers for the exact same journey — and most teams never question which rule they’re using. We tested this at Varnan across a dozen client accounts in 2025: switching from last-click to data-driven attribution changed channel budget allocations by an average of 32%, with upper-funnel channels consistently being the most undervalued under last-click.
There are six primary attribution models used in digital marketing today. Each tells a different story from identical conversion data:
| Model | How Credit Is Assigned | Best For | Key Weakness |
|---|---|---|---|
| Last-click | 100% to final touchpoint | Simple, short-cycle purchases | Ignores all upper-funnel |
| First-click | 100% to first touchpoint | Awareness channel analysis | Ignores closing channels |
| Linear | Evenly split across all touches | Balanced baseline view | Treats all touches equally |
| Time-decay | More credit to recent touches | Short sales cycles | Undervalues early awareness |
| Position-based (U-shaped) | 40% first, 40% last, 20% split | B2B, longer sales cycles | Still a fixed rule |
| Data-driven | ML-calculated incremental impact | High-volume accounts | Needs sufficient data volume |
Take one customer journey: Day 1, sees an Instagram ad (awareness). Day 4, watches a YouTube explainer (consideration). Day 9, clicks a Google retargeting ad. Day 10, searches the brand name plus “pricing,” clicks, and converts.
Run that single journey through each model:
Under last-click, Instagram and YouTube look like they did nothing — and in a budget review, they’re the first lines cut. Under position-based or data-driven, they’re doing real, measurable work in the journey that ends in brand search and retargeting “closing” it.
Content, YouTube, and social — the channels that build awareness — almost always look unprofitable under last-click, simply because they’re rarely the final touchpoint before conversion. That’s structural, not a reflection of whether they’re working.
Teams that judge every channel by last-click systematically starve the channels that create the demand their bottom-funnel channels — branded search, retargeting — are just there to capture. The branded search campaign looks like a star performer. It’s mostly just collecting conversions that upper-funnel spend already created.
In our experience at Varnan, when we measure this with clients, branded search “ROAS” drops by 40–60% when you switch from last-click to data-driven attribution — because the model stops giving branded search credit for conversions that began with a YouTube or content touchpoint weeks earlier. The money wasn’t being wasted on upper-funnel. The model was lying about where the work was being done.
Sometimes the lie isn’t in the tracking. It’s in the model interpreting it.
If your sales cycle is short and the purchase is low-consideration, and your traffic volume is modest: last-click or position-based is fine — and data-driven models need enough volume to be statistically reliable, which low-volume accounts often don’t have.
If your sales cycle is longer — most B2B and considered-purchase journeys, which describes most early-stage AI and software startups — position-based or time-decay gives a far more honest read without needing huge data volume to do it.
If you have enough conversion volume for GA4’s data-driven model to activate reliably, switch to it as soon as you qualify. It’s the most accurate by design, because it’s calculated from your actual conversion paths rather than a fixed rule.
The practical habit that matters more than picking “the right model”: before making a budget decision based on channel performance, look at the same data under two or three models. If a channel looks weak under last-click but solid under position-based, that’s worth investigating before it gets cut.
Attribution model choice isn’t a settings checkbox you set once and forget. It’s a lens — and the wrong lens makes good channels look bad and bad channels look good, even when every conversion is tracked perfectly.
Get the tracking foundation right first — we covered that in the first 90 days playbook — then make sure the model reading that data isn’t quietly lying to you too.
We set up attribution that matches your actual funnel — not just whatever the ad platform defaults to — so budget decisions are based on what’s really driving results.
Last-click attribution gives 100% of the conversion credit to the final touchpoint before someone converts — it’s the default in Google Ads and Meta, and it’s blind to everything that happened earlier in the journey. First-click does the opposite: 100% of the credit goes to the very first touchpoint, which is useful for understanding what drives initial awareness but ignores whatever actually closed the deal. Neither tells the full story on its own — that’s the core problem this post walks through.
Because last-click only credits the final touchpoint, and upper-funnel channels — content, YouTube, social — are rarely the last thing someone clicks before converting. That’s structural, not a reflection of whether the channel is actually working. The result: branded search and retargeting look like star performers while they’re often just capturing demand that upper-funnel spend already created.
Position-based attribution gives 40% of the credit to the first touchpoint, 40% to the last, and splits the remaining 20% across everything in between. It’s a strong fit for longer sales cycles and considered-purchase journeys, which describes most early-stage AI and software startups, because it gives a far more honest read of upper-funnel impact without needing the data volume that data-driven models require.
Yes, when you have enough conversion volume for it to activate reliably. Data-driven attribution uses your actual conversion data to calculate each touchpoint’s real incremental impact, rather than applying a fixed rule like last-click or linear — which makes it the most accurate model by design. It’s GA4’s default for properties with sufficient volume, but low-volume accounts often don’t have enough data for it to be statistically reliable, so position-based or time-decay are better starting points until you qualify.
Match the model to your sales cycle and data volume. If your purchase is low-consideration with a short cycle and modest traffic, last-click or position-based works fine. If your sales cycle is longer — most B2B and considered-purchase journeys — position-based or time-decay gives a far more honest read without needing huge volume. And if you have enough conversions for GA4’s data-driven model to activate, switch to it. The habit that matters most: before cutting a channel, check its performance under two or three models, not just one.
Performance MarketingBuild retargeting funnels on Meta and Google that recover lost buyers. India CPMs, budget splits, CAPI, and 2026 signal-loss tactics for lean startup teams.
Real 2026 India LinkedIn Ads benchmarks: CPC, CPL, minimum budgets, targeting layers, and Lead Gen Forms vs landing pages — plus when a startup is too early.
Learn how Indian startups can leverage WhatsApp Business API, CTWA ads, AI drip sequences, and re-engagement flows to drive measurable conversions in 2026.