3x Pipeline ROAS: How DevRev Saved a Google Ads Channel That Was About to Be Shut Down

The challenge

When DevRev came to us, they were close to shutting down their Google Ads account entirely.

Not because the budget was gone. Because the return wasn’t there. Prior spend had generated pipeline at barely above break-even, and the sales team was spending time on leads that were never going to close. The account was running on inertia, not strategy.

The audit surfaced the structural problems fast:

• No tracking infrastructure. Without it, there was no way to tie ad spend to actual pipeline. Everything downstream was a guess.
• Disorganised campaign structure. No alignment between targeting, messaging, and DevRev’s actual ICP.
• The keywords were pulling the wrong audience. Broad targeting was driving early-stage startup traffic. DevRev sells to Enterprise. Those two audiences are not the same conversation, and the account was treating them like they were.
• No visitor intelligence layer. No signal on who was actually landing on the page, or whether any of them were the right companies.

DevRev’s ICP is large-scale enterprise organisations. The account was generating traffic from companies that would never become customers.

Our approach

The fix wasn’t more spend. It was precision. We rebuilt the account around one question: what does it actually take to get an enterprise buyer to convert?

Pillar 1: Enterprise-first messaging

We started on the sales call recordings. Not the ad account.

We listened to calls with qualified enterprise prospects and compared them against startup calls. The difference in language, pain points, and objection patterns was immediate. Enterprise buyers cared about security, compliance, scalability, and integration depth. Startup buyers cared about price and time-to-value. The account was running one message for both.

We rewrote everything from that insight:

• Headlines, descriptions, and extensions reframed around enterprise pain points and compliance requirements
• Ad copy matched to the decision stage of an enterprise buyer, not a general SaaS searcher
• Landing page messaging aligned to what the enterprise searcher actually clicked on

When the message matches the buyer, the wrong buyers stop clicking. That sounds like a loss. It’s actually the goal.

Pillar 2: Visitor intelligence stack

To confirm the messaging was reaching the right audience, we needed more signal than Google Ads provides by default.

We integrated three tools:

• Factors.ai for firmographic visitor intelligence — company name, size, and industry for site visitors
• Microsoft Clarity for session-level behaviour analysis, watching how enterprise versus startup visitors engaged with the page
• RB2B for individual-level visitor identification on the US side

This gave us the confirmation layer we needed. We could see in near real-time whether enterprise-tier accounts were showing up, and which creative assets and keywords were bringing them in. The optimisation loop got much tighter.

Pillar 3: Intent-led media buying

We ran broad match at the start deliberately. Not because it’s efficient, but because enterprise campaigns need discovery data before you can get precise.

Broad match across relevant categories gave us the attribution map: which campaigns attracted enterprise accounts, which attracted startups, and which attracted nobody useful. That data was the point.

Once the signal was clear:

• Paused everything that was not generating enterprise-tier traffic
• Concentrated budget into campaigns with confirmed ICP reach
• Applied audience segmentation overlays to reduce impression waste from non-ICP firmographics
• Ran ongoing search term audits and negative keyword reviews to keep the signal clean

By the end of the engagement, every dollar in the account had a verified line back to an enterprise account that entered pipeline.

The results

The account that was weeks from being shut down became the top pipeline channel.

Prior period pipeline ROAS was 1.1x — spend barely covered what came back. Here is what the rebuilt account delivered, broken down by region.

Pipeline grew 7x versus the prior period, on 2.7x the spend. The efficiency gain was 2.7x, not just the volume. Every additional dollar worked harder than the last.

The US market ran a 58% ICP lead rate. More than half of every US lead was enterprise-qualified. That is what happens when the keywords, the message, and the audience layer are all pointing at the same buyer.

India produced equal pipeline share on equal spend, at a 33% ICP rate. Volume and precision each carried their weight by region, and both came in at nearly identical pipeline contribution at the end.

Why it worked

Three things drove the outcome:

1. We started with sales calls, not keywords. The gap between what DevRev was saying in ads and what enterprise buyers needed to hear was visible the moment we listened to the calls. Fixing that gap upstream fixed conversion rate downstream.
2. Visitor intelligence closed the feedback loop. Without Factors.ai and RB2B, we would have been optimising on click data and CPL. With them, we were optimising on whether the right companies were actually showing up. That is a fundamentally different game.
3. Broad match was a tool, not a default. We used it deliberately to build the attribution map, then paused everything that couldn’t show enterprise reach. The final account was precise because the discovery phase wasn’t — and we planned it that way from the start.