Breaking down $3.78M in attributed pipeline revenue

We built the pipeline I described in the last post. Adobe → BigQuery → Salesforce → Google Ads. Then we ran it for a full 90 days.

The number that came back: $3.78M in new-business pipeline revenue attributed to paid search.

That number saved my career. But more importantly, it broke open how we actually think about B2B Google Ads.

The context

Before the pipeline, Google Ads native reporting said we acquired $1.2M in new customer revenue. Salesforce said we closed $860K in actual new-business deals. We had no idea which number was real. The gap was a gap we couldn't explain.

With the pipeline running, we could finally match conversion events to closed deals. Not just by GCLID. By domain, by company, by close date, by opportunity amount.

The result: we were actually capturing $3.78M in pipeline. Which means Google's native reporting was off by a factor of 3. And Salesforce was showing us deals that the pipeline proved came from paid search, not "other" channels.

Where the $3.78M came from

This is where it gets interesting. Because the breakdown doesn't match what generic B2B SaaS marketers expect.

Tier 1 — Hyper-specific keywords (52% of attribution, $1.97M): These are the exact keywords that shouldn't work according to conventional wisdom. Part numbers. Specific model names. Exact pain-point terminology. "Industrial marking system specifications." "NFPA 70E arc flash labels." "Laser marking for medical device traceability."

These keywords have basically zero search volume in Google's Keyword Planner. They get clicked maybe 20-30 times a month. But when they convert, they convert to real deals. We're talking 35-40% conversion rate from click to actual pipeline opportunity.

Tier 2 — High-intent branded + category terms (31% of attribution, $1.17M): Brady safety labels. Seton marking systems. Emedco compliance signs. Plus high-intent category terms like "warehouse labeling system" or "chemical safety signage."

These have real search volume. 500-2000 clicks a month. But the conversion rate is lower (maybe 8-12% to opportunity). The volume makes up for it.

Tier 3 — Broad intent terms (12% of attribution, $450K): "Safety signs," "marking systems," "inventory labels." These get the most clicks but the lowest-quality conversions. Lots of DIY, hobbyist, and competitor traffic mixed in. We still included them because the sheer volume matters, but honestly we could trim these and barely miss the pipeline.

The other 5%? Attribution noise. First-click aliasing, domain matching edge cases, people who cleared their cookies mid-journey.

Why this matters more than the total

The breakdown tells you where to build.

In most B2B accounts I've audited, teams spend 60% of their effort on Tier 3 (broad terms). They're "optimizing" for volume, for click volume, for impression share on generic keywords. They're doing weekly bid adjustments on "safety" and "marking" and "labels."

But 60% of the actual revenue is hiding in Tier 1. In keywords that get 30 clicks a month. In part numbers nobody thinks to bid on. In specific terminology that looks worthless in Keyword Planner.

If you're not actively building Tier 1 and Tier 2, you're leaving millions on the table.

The bidding implication

Here's where I changed our strategy based on this data.

We flipped our bid structure upside down.

Before: High bids on broad terms. Low bids on specific terms (because they looked risky, low-volume).

After: Extreme bids on Tier 1 terms. Moderate bids on Tier 2 terms. Conservative bids on Tier 3 terms.

Tier 1 keywords now get $15-25 per click. Some part numbers get $30. Sounds insane until you realize a single deal is $50K-200K. A $20 CPC is free money.

Tier 2 gets $3-8 per click. Tier 3 gets $0.50-2 per click, and honestly we're thinking about cutting most of it.

Our overall CPC went up. Our click volume went down. Our CPL went up. And our pipeline went up 60%.

The algorithm hated it. Google's Smart Bidding recommendations screamed at us to rebalance toward broad terms. We ignored it. And made more money.

How we discovered this

Without the pipeline, we never would have known.

Google's native reporting would have shown broad terms "performing well." The algorithm would have optimized toward them. We would have been re-allocating budget away from what actually works.

Salesforce alone wouldn't have shown the pattern either. We could see that we closed a $100K deal, but not which keyword kicked it off. Not which click from 8 months earlier mattered.

BigQuery let us connect the dots. Run the cohort analysis. Ask "which keywords appear in the first 10 clicks before a company closes a deal?" and get a real answer.

Then we could make a real decision.

What this cost to build

About 4-6 weeks of engineering time (mostly SQL and Python scripting).

About $2K/month in BigQuery costs (that number has come down as we optimized the queries).

About $500 in Salesforce API calls.

Versus: leaving $2M+ on the table every quarter because we didn't know where our actual revenue came from.

The ROI math is almost insulting.

The hard part

Building the pipeline is technical. I can't sugar-coat that.

You need SQL. You need to understand GCLID matching. You need to work with your data warehouse team. You need to handle the edge cases (GCLID expiration, domain changes, bot traffic).

But here's the thing: if you're running a B2B account above $50K/month, you don't have a choice anymore. Native reporting is not enough. You're making optimization decisions on bad data. You're bidding wrong. You're pruning what works and scaling what doesn't.

The technical barrier feels high until you realize the cost of ignorance is higher.

The action

If you're in-house managing a B2B account:

1. Audit your actual attributed revenue using your CRM + Google Ads data, even if it's manual
2. Compare it to what Google's native reporting says
3. If they don't match by more than 20%, you have a data problem
4. Build a basic pipeline. Not perfect. Just real.

If you're using an agency:

1. Ask them: "What percentage of our new-business revenue did you attribute to paid search last quarter?"
2. If they don't have a number, they don't have a pipeline
3. If they do have a number, compare it to what your Salesforce says actually closed
4. If they don't match, ask why

The gap between "what Google says" and "what actually happened" is where all the money is.

The pipeline is how you find it.