Google Ads Attribution Models: Why 'Last Click' is Costing You Money (2024)

Imagine a soccer team. The Striker scores the goal. But the Midfielder made the perfect pass. And the Goalkeeper prevented the other team from scoring.

If you paid your players based on "Last Click Attribution", you would give 100% of the salary to the Striker and fire the rest of the team. Then, you would wonder why the Striker stopped scoring. (Answer: He never gets the ball anymore).

This is exactly what happens in Google Ads when you rely on the default "Last Click" model. You give 100% of the credit to the final ad the user clicked, and you pause the "assist" campaigns (Display, YouTube, Top-of-Funnel Search) that actually created the demand.

In this guide, we break down why Data-Driven Attribution (DDA) is the only model that matters in 2024, and how switching to it effectively "unlocks" scale for your account.

---

Part 1: The Death of "Last Click"

For 20 years, Last Click was the standard because it was easy to measure.

• User clicks Ad A -> No sale.
• User clicks Ad B -> No sale.
• User clicks Ad C -> Sale ($100).
• Result: Ad C gets $100 revenue. Ad A and B get $0.

The Problem: The user wouldn't have clicked Ad C if they hadn't seen Ad A and B first. The journey is non-linear. If you look at your "Path Metrics" in Google Analytics 4, you'll likely see that your average customer interacts with your brand 3-7 times before buying.

If you optimize for Last Click, you inadvertently kill your growth because you stop filling the top of the funnel.

---

Part 2: Enter Data-Driven Attribution (DDA)

Google's Data-Driven Attribution uses machine learning to analyze thousands of conversion paths in your account. It looks at what happened when users saw an ad versus when they didn't.

It assigns fractional credit.

The DDA Scenario:

1. Touch 1 (Display Ad): Introduces the brand. (DDA assigns $20).
2. Touch 2 (Generic Search "Best CRM"): Educates the user. (DDA assigns $30).
3. Touch 3 (Brand Search "Acme CRM"): Closes the deal. (DDA assigns $50).

The Result: Instead of your Display Campaign looking like a failure (0 Conversions), it now shows 5 "Fractional Conversions". You realize it is actually profitable, so you scale it up.

Key Takeaway: DDA aligns your metrics with reality. It proves that marketing is an ecosystem, not a series of isolated events.

---

Part 3: The "Smart Bidding" Connection

This is the most critical part. Smart Bidding (Target CPA, Target ROAS) optimizes based on the data you feed it.

• Scenario A (Last Click): The algorithm sees that "Generic Keywords" have 0 conversions. It bids them down to $0. Your volume dries up.
• Scenario B (DDA): The algorithm sees that "Generic Keywords" are responsible for 30% of the value. It keeps bidding on them.

By switching to Data-Driven Attribution, you are effectively training the AI to value the entire funnel, not just the bottom. This allows the algorithm to bid aggressively on "Assisting" keywords that drive future sales.

---

Part 4: View-Through Conversions (The Controversy)

"View-Through Conversions" (VTC) happen when a user sees your ad (but doesn't click), and then converts later on another channel.

• Metric: Google Ads Reporting column "View-Through Conv."
• Debate: Many advertisers think VTCs are fake. "They would have converted anyway!"

Our Stance: For Display and YouTube, VTCs are essential context.

• If you run a TV ad, you don't expect people to click the TV. You expect them to remember. YouTube is digital TV.
• Strategy: We typically value a View-Through Conversion at 20-30% of a Click-Through Conversion when calculating true ROAS. It's a signal of brand lift.

---

Part 5: Window Shopping (Attribution Windows)

How far back should you look?

• Click-Through Window: Default is 30 days. For high-ticket B2B (e.g., Enterprise Software), we extend this to 90 days. Buying decisions take time. If someone clicks in January and buys in March, you want to credit the January ad.
• View-Through Window: Default is 1 day. We recommend keeping it tight (1 day). If someone sees an ad and buys 7 days later without clicking, the correlation is weak.

---

Action Plan: How to Switch & Audit

1. Check Your Setting:

   • Tools & Settings → Measurement → Conversions.
   • Click your Primary Goal (e.g. Purchase).
   • Look at "Attribution Model". Is it "Last Click"? If so, change it to Data-driven.

2. The "Model Comparison Tool":

   • Go to Attribution → Model Comparison.
   • Compare "Last Click" vs "Data-Driven".
   • Look for huge shifts. You will likely see that Non-Brand Search and Display campaigns gain credit, while Brand Search loses credit.
   • This confirms that your Non-Brand campaigns were doing more work than you thought.

3. Adjust Targets:

   • When you switch DDA, your conversion numbers will change.
   • Your tCPA targets might need adjustment. If a campaign suddenly "gains" 20% more conversions due to DDA, you can lower your tCPA target or increase the budget.

Conclusion: You are not smarter than the machine learning algorithm. Stop forcing it to look at a limited dataset (Last Click). Give it the full picture (DDA), and watch it find customers you didn't know you had.