Scene: The Quarterly Business Review (QBR). Director of acquisition: “We crushed it! Facebook ROAS is 4.5! Google ROAS is 6.0! Overall we generated $5 Million in revenue from ads!” CFO (Looking at P&L): “That’s great. But our total revenue for the quarter is only $3 Million. How did you generate $5M?” Silence. This is the Attribution Crisis. Every platform (Meta, Google, TikTok) is greedy. They claim credit for every sale they touched.

User clicks Facebook Ad (Day 1).
User clicks Google Search Ad (Day 2).
User buys (Day 3). Facebook claims the sale. Google claims the sale. The CFO pays double the commission. This article explains how to stop the “Voodoo Math” and start measuring Incrementality.

Why Maison Code Discusses This

We are engineers. We deal in absolutes. 1 + 1 = 2. Marketing attribution is probabilistic. 1 + 1 = Maybe 3. We build the data pipelines that feed the CFO’s dashboard. We see the data before the Agency massages it. We discuss this because Bad Attribution leads to Bad Engineering. If you over-invest in a channel that isn’t working, you can’t afford to fix the website. We want your P&L to be healthy so you can invest in Innovation.

1. The Death of Last-Click Attribution

For 20 years, we relied on “Last Click”. “Whatever the user clicked last gets 100% of the credit.” This biases heavily towards Google Search (Branded).

User sees Facebook Ad (Discovery).
User sees Influencer Post (Consideration).
User googles “Brand Name” (Intent).
User clicks Google Ad -> Buys. Google gets 100% credit. Facebook gets 0%. If you fire Facebook based on this, the funnel dries up. No one enters the top. You cannot measure a soccer game by only looking at the striker who scored the goal. You must value the midfielder who passed the ball.

2. The Blended ROAS (MER) - The North Star

We advise clients to ignore platform-reported ROAS (Return on Ad Spend) for high-level budgeting. Focus on MER (Marketing Efficiency Ratio). Also known as Blended ROAS. Formula: Total Revenue / Total Ad Spend.

Total Revenue: $100,000.
Facebook Spend: $10,000.
Google Spend: $10,000.
Total Spend: $20,000.
MER = 5.0. This is the only number that matters to the CFO. “For every dollar that leaves the bank account for ads, 5 dollars come back.” It doesn’t tell you where to spend, but it tells you if you are profitable.

3. The Incrementality Test (The Gold Standard)

How do we know if Facebook is really driving net new revenue? Maybe those people would have bought anyway? We run Lift Tests (Geo-Holdout).

Select Markets: Choose two similar cities. (e.g., Dallas and Houston).
Holdout: Turn OFF Facebook Ads in Houston (The Control).
Active: Keep Facebook Ads ON in Dallas (The Test).
Measure: Run for 4 weeks.
- Did Dallas sales drop? Did Houston sales stay flat?
- If Houston sales dropped by 10% when ads were off, then Facebook contributes 10% incremental lift. This is scientific. It removes the “Attribution Voodoo”. It gives the CFO hard data to approve huge budgets. “We know for a fact that $1 in FB = $0.50 in purely incremental profit.”

4. Server-Side Tracking (CAPI)

(See Google Tax). The pixels are dying. iOS 17 strips tracking parameters. Use Conversion API (CAPI). Send data from Shopify Server -> Facebook Server. This recovers ~15% of “Lost Signal”. It improves the algorithm’s ability to find buyers. Better signal = Lower CAC.

5. View-Through Conversions (The Ghost)

Agencies love “View-Through”. “The user SAW the ad, didn’t click, but bought later. We claim credit.” Be very careful. Display Ads (Criteo, AdRoll) often claim massive View-Through revenue. But they are just retargeting people who were already buying. Rule: Value View-Through at 10% of Click-Through. Don’t let an agency bill you on View-Through revenue without a lift test.

6. The “New Customer” CAC vs “Blended” CAC

Another trap. “Our CAC is $20!” Is that for New customers, or Returning customers? If you spend $20 to get a Returning Customer to buy… you failed. Email is free. You should measure ncCAC (New Customer Acquisition Cost). Total Ad Spend / New Customers Acquired. This is the true cost of growth. If ncCAC > LTV (Lifetime Value), you are dying.

7. LTV:CAC Ratio (The Holy Grail)

This is the ultimate distinct metric of business health. LTV (Lifetime Value): How much profit a customer brings over 3 years. CAC: Cost to acquire them.

Ratio 1:1: You are losing money (OpEx eats you).
Ratio 3:1: Healthy SaaS / E-com business.
Ratio 5:1: Money printing machine. Scale aggressively. If your Ratio is 5:1, and you are “saving budget”, you are making a mistake. You should spend more until the ratio drops to 3:1. Gain market share while it’s cheap.

8. Offline Attribution (The Store)

“I saw the ad on Instagram, walked into the store, and bought it.” Facebook sees the cost. The Store sees the revenue. The E-com P&L looks bad. Solution: Offline Conversions API. Upload your POS (Point of Sale) data to Meta every night. Match via Email/Phone. Meta matches the user. “Ah, the user who saw the ad bought in-store!” Suddenly, your Facebook ROAS jumps from 2.0 to 4.0. The Marketing Team gets a bonus. The Store Team gets traffic. Everyone wins.

9. Zero-Based Budgeting

Every year, start from zero. Don’t say: “We spent $1M last year, let’s spend $1.1M.” Say: “What do we need to Enable?” Just because Google Brand Search worked last year doesn’t mean it works now. Question every channel. Force them to re-prove their Incrementality.

10. The Agency Incentive Problem

Most Agencies charge a ”% of Ad Spend”. If you spend $1M, they get $100k. If you spend $500k, they get $50k. They are incentivized to make you spend more, not more efficiently. They will push for “Broad Targeting” and “Awareness Campaigns” because they scale spend easily. Strategy: Move to “Performance Fee” pricing. Base Fee + % of Incremental Profit. Align their wallet with your wallet.

11. The Creative Strategist (The New Unicorn)

Since tracking is dead, Creative is King. The Algorithm can find the buyer if the creative calls them out. “Calling all New Moms in Paris!” This creative does the targeting. You need a Creative Strategist. Someone who looks at the data (retention, thumb-stop rate) and tells the designer what to build. “This video held attention for 3 seconds. The drop-off was at the pricing reveal. Let’s move price to the end.” This is where 80% of your optimization effort should go.

12. The Rule of 3 (Channel Diversification)

If you rely on 1 channel (Facebook) for >50% of revenue, you are fragile. If Zuck changes the algorithm, you die. The Diversified Portfolio:

Meta: 40% (Discovery).
Google: 30% (Intent).
Experimental: 30% (TikTok / Influencer / Direct Mail). Always have a third leg of the stool. This gives you leverage. If Facebook CPMs spike, you can shift budget to TikTok. Liquidity is safety.

13. Conclusion

Marketing Spend is an Investment Portfolio. Some assets are “Safe” (Google Brand Search). Some are “Growth” (Meta/TikTok). Some are “Experimental” (Influencers/Connected TV). Manage risk. Measure return. And never, ever let the fox (The Ad Platform) count the chickens (The Attribution). Own your data.

CFO asking tough questions?

We implement Server-Side Tracking (CAPI), Marketing Data Warehouses (Snowflake), and Incremental Lift Testing.

Hire our Architects.