Project Echo: AI Drives 3.7x ROAS in 2026

In our “Project Echo” campaign, we used sophisticated AI to perform deep account analysis, and the results were immediate: better segmentation, a more effective campaign, and real financial gains. This kind of targeted approach, which dissects individual account interactions, is how you turn data into dollars.

Campaign Teardown: Project Echo’s Precision Targeting

We needed to sell more of our enterprise SaaS into mid-market companies, but our old broad-stroke campaigns were failing us on conversions. They generated leads, sure, but we didn’t have a deep enough read on individual account needs to actually close them. For Project Echo, we bet that a more granular approach, using AI intelligence for deep account analysis, would finally move the needle on our conversion rates.

We put $250,000 behind Project Echo for a six-week sprint in Q1 2026. Our targets were mid-sized companies (50-500 employees) in manufacturing and logistics, specifically ones whose digital footprints screamed that they had the exact problems our software solves. The goals were clear: keep cost per lead (CPL) under $75 and get a return on ad spend (ROAS) over 3.0x.

Strategy: Data-Driven Segmentation and Dynamic Messaging

Project Echo’s whole strategy was built on Amplitude, our AI analytics platform, to run exhaustive account analysis. We went way beyond simple demographic filtering by feeding the AI a mix of our own first-party CRM data (think past interactions, support tickets, product usage) and third-party behavioral signals like website visits, competitor research, and even industry news. The AI then churned through all of it to build out some incredibly specific customer segmentation profiles.

For instance, the AI flagged one segment of manufacturing firms in the Southeast that were not only searching for “supply chain optimization software” but also had a ton of open support tickets in our CRM about inventory management. That’s the kind of specificity that kills generic personas. We didn’t just know who they were. We knew what they were actively struggling with, right now.

Creative Approach: Hyper-Personalized Narratives

Once we had these hyper-specific segments, our creative team built a whole matrix of ad copy and visuals. Instead of just one message, we had dozens, each one hitting on a specific pain point we’d identified. For that manufacturing segment, the ads showed simplified factory floors with headlines like, “Reduce Inventory Overheads by 15% with Intelligent Supply Chain Automation.” That’s so much more powerful than the generic “Boost Efficiency” fluff you see everywhere. We ran it all through AdRoll, using its dynamic creative optimization to automatically tweak creative elements based on how each account was engaging in real time.

Targeting and Placement: Surgical Precision

Our targeting was surgical. We used a mix of LinkedIn Ads, Google Display Network (GDN) retargeting, and programmatic through The Trade Desk. We uploaded our segmented company lists right into LinkedIn for account-based targeting. On the GDN, we built custom intent audiences from the keywords our AI analysis spit out, so we’d show up the moment a decision-maker started researching. Programmatic buys were focused on industry publications our target accounts were already reading. We even layered on IP-based targeting to hit high-value companies even if their people hadn’t touched our content yet. Yes, it was more complex to set up, but the benefit was a huge reduction in wasted impressions.

What Worked: Unpacking the Success Metrics

The results from Project Echo blew past our goals. We hit an overall CPL of $62 (our target was $75) and, most importantly, a ROAS of 3.7x. That’s $3.70 in revenue for every ad dollar spent, a massive jump from our usual 2.1x. No one’s arguing with numbers like that.

The click-through rate (CTR) really showed the power of this approach. Our personalized ads for top segments hit an average CTR of 1.8%, double the 0.9% baseline from our more generic ads. That doubling in engagement sent much higher-quality traffic to our landing pages, and our bounce rates dropped accordingly. Better yet, our lead-to-qualified-opportunity conversion rate for this campaign was 12%, up 4 points from before. We weren’t just getting more leads. We were getting the right leads.

The manufacturing segment was our big winner. It drove 35% of total conversions but only ate up 28% of the ad spend. The personalized messaging around inventory control and supply chain visibility just clicked with them. We saw it in the numbers: the cost per conversion for this specific group was only $450, way below the campaign’s average of $535.

What Didn’t Work: Learning from the Edges

Of course, not everything was a home run. We had a segment of smaller logistics companies (50-100 employees) in the Pacific Northwest that just fizzled out, giving us a CPL of $98 and a paltry 1.9x ROAS. Diving back into the account analysis, we realized our mistake: they had the right pain points, but their budget cycles were way longer than we’d planned for. Our six-week campaign sprint and quick-sale messaging just didn’t fit. The creative also missed the mark. It didn’t do enough to calm their fears about implementation complexity. The lesson was clear: you can have perfect segmentation, but if you don’t understand the buying journey for that segment, you’re going to waste money.

Our retargeting strategy also had some weak spots. The GDN retargeting frequency caps were way too aggressive for some segments, and it was causing obvious ad fatigue. We could see it in the data, after the fourth impression in some high-volume groups, the CTR would dip and the CPC would start to climb. We clearly needed a more sophisticated approach to frequency management, one that would adjust based on how engaged a segment was and where they were in the buying cycle.

Optimization Steps Taken: Iteration is Key

We made a few key optimizations on the fly. For that underperforming logistics segment, we hit pause on the hard-sell conversion ads and reallocated that budget to nurturing them with webinars and whitepapers. That change dropped their CPL to $80 almost overnight, although we knew we were now playing a longer game with their conversion timeline.

To fix the ad fatigue, we dialed back our retargeting frequency caps, no more than three impressions in a 7-day window for highly engaged users, and started rotating in more creative variations. That simple change gave us a 7% bump in retargeting CTR for the second half of the campaign. We also ran an A/B test on landing pages for our star manufacturing segment and found that a page with a specific case study converted 9% higher than a generic product page. It just proved again how much a concrete proof point matters to a highly targeted audience.

Constantly checking the data through continuous account analysis and performance monitoring was what let us refine our approach and get the most out of the campaign. A set-it-and-forget-it strategy just doesn’t work. You have to be ready to adapt based on what the numbers are telling you. The speed at which you can pivot, especially when you have the insights from AI intelligence, is what makes a campaign truly successful.

Project Echo proved to us that deep, AI-driven account analysis is essential for winning in a tough B2B market. Its incredible precision in customer segmentation lets you create incredibly relevant messages and put your budget where it will do the most good, which flows directly to the bottom line.