Stop letting branded search cannibalize your marketing ROI. Here is how to architect your data pipelines to separate vanity metrics from actual growth drivers.

Attribution Hell: Why I Stopped Trusting Blended ROAS and How You Can Too

A few years back, I was sitting in a war room after a major traffic spike on prod-ingress-controller-01. Our Marketing Director was high-fiving everyone because their new “Performance Max” campaign was showing a 10x ROAS. I pulled the raw logs from analytics-db-prod and felt my stomach drop. We weren’t “reaching new customers.” We were paying $4.25 a click for people who were literally searching for “TechResolve Login.” We were subsidizing our own bookmarks. It’s the ultimate DevOps nightmare: paying for resources you already own because your monitoring is too abstracted to see the waste.

The root cause is what I call “The Halo Effect.” Google Ads likes to blend your high-converting branded terms (people who already know you) with your non-branded discovery terms (people who don’t). This creates a blended average that looks healthy but hides a non-brand campaign that might be burning cash at an unsustainable rate. If you aren’t separating these, you aren’t marketing; you’re just paying a tax to Google for your own brand equity.

Solution 1: The Quick Fix (Campaign Level Segmentation)

If you need a fix before the next board meeting, stop trying to be clever with scripts. You need to physically separate your budget. Create a strict naming convention and enforce it. In my experience, if it’s not in the name, it’s not in the report. This allows you to create a “Master Brand” filter in your reporting dashboard.

• Step 1: Rename all campaigns with a prefix: [BRAND] or [NON-BRAND].
• Step 2: Add your brand name and all variations (even the common typos we see in search-logs-01) as Negative Keywords in your Non-Brand campaigns.
• Step 3: Build a simple pivot table to compare the two.

Pro Tip: Don’t forget your competitors’ names. I usually bucket those into a third category: [CONQUESTING]. The CPCs are higher, and the conversion rate is lower, but it keeps your brand data clean.

Solution 2: The Permanent Fix (BigQuery Data Pipeline)

As an architect, I don’t trust the Google Ads UI. It’s built to encourage spending. The “Permanent Fix” involves piping your raw data into BigQuery using the Data Transfer Service. This lets us run our own attribution logic without the “Google bias.”

SELECT 
    campaign_name,
    CASE 
        WHEN REGEXP_CONTAINS(LOWER(campaign_name), 'brand') THEN 'Brand'
        ELSE 'Non-Brand'
    END AS campaign_type,
    SUM(cost) / 1000000 AS spend_usd,
    SUM(conversions) AS conversions,
    SAFE_DIVIDE(SUM(conversions), SUM(cost) / 1000000) AS roas
FROM `your-project.google_ads.p_CampaignStats_123456789`
GROUP BY 1, 2
ORDER BY spend_usd DESC;

This SQL snippet is a simplified version of what we use on reporting-srv-04. By pulling this into a DataStudio or Looker dashboard, we can see the real cost of acquisition (CAC) for new users versus the maintenance cost of existing ones. If the “Non-Brand” ROAS is below your break-even point, you have an efficiency problem, regardless of how good the “Total” number looks.

Solution 3: The “Nuclear” Option (The Brand Lift Study)

This is the one that makes the Marketing team sweat, but it’s the only way to find the “Incrementality” of your spend. We call it the “Dark Period.” You turn off all branded search ads for a specific geographic region for two weeks while keeping it on for the rest of the world. Then, you look at your organic search traffic in that same region.

Is it hacky? Maybe. Is it risky? Definitely. But I’d rather take prod offline for 10 minutes of scheduled maintenance than run a system that’s leaking five figures of company cash every month into a black hole of non-incremental conversions. As a Senior Engineer, my job isn’t just to keep the servers running; it’s to make sure the data we’re serving actually means something.

Darian Vance

Lead Cloud Architect & DevOps Strategist

With over 12 years in system architecture and automation, Darian specializes in simplifying complex cloud infrastructures. An advocate for open-source solutions, he founded TechResolve to provide engineers with actionable, battle-tested troubleshooting guides and robust software alternatives.

🤖 Frequently Asked Questions