Project Atlas: B2B SaaS Performance in 2026

For digital advertising professionals seeking to improve their paid media performance, the relentless pursuit of efficiency and impact defines our daily grind. We’re constantly refining, testing, and iterating—but how often do we really dissect a campaign, warts and all, to extract genuine, actionable insights?

As a seasoned paid media director, I’ve seen countless campaigns—some soar, others sputter. The difference often lies not in the initial budget, but in the rigor of the campaign teardown process. It’s about more than just reporting numbers; it’s about understanding why those numbers appeared and what to do next. Let me walk you through a recent B2B SaaS campaign we managed, “Project Atlas,” highlighting its strategic underpinnings, the creative execution, and the data-driven adjustments that ultimately salvaged its performance.

The Genesis of Project Atlas: Strategy and Initial Goals

Project Atlas was designed to drive sign-ups for a new enterprise-level AI-powered analytics platform, “InsightEngine Pro,” targeting marketing leaders in companies with 500+ employees. Our client, a burgeoning tech firm based out of the Atlanta Tech Village, had aggressive growth targets. They needed to establish market presence quickly and generate high-quality leads for their sales team.

Our core strategy revolved around demonstrating the platform’s unique ability to predict customer churn with 90%+ accuracy—a significant pain point for our target audience. We decided on a multi-platform approach, focusing heavily on LinkedIn Ads for its B2B targeting capabilities and Google Search Ads to capture intent-based traffic.

Our initial goals for the 8-week campaign were ambitious:

• Budget: $150,000
• Duration: 8 weeks (July 1, 2026 – August 26, 2026)
• Target CPL (Cost Per Lead): $75
• Target ROAS (Return On Ad Spend): 1.5x (based on projected sales cycle and average contract value)
• Target CTR (Click-Through Rate): 0.8% (LinkedIn), 3.5% (Google Search)
• Target Conversion Rate (Lead Form Submission): 5%

We knew these were tight, particularly the ROAS for a new, high-ticket SaaS product, but the client was firm. My team and I developed a comprehensive plan, including detailed audience segmentation and a content strategy that mapped specific platform features to pain points at different stages of the buyer journey.

Creative Approach: Balancing Authority and Urgency

For Project Atlas, our creative strategy was two-pronged. On LinkedIn, we opted for a mix of single image ads and video testimonials featuring early adopters. The imagery was clean, professional, and often depicted data visualizations or confident business leaders. Our ad copy emphasized the “90% churn prediction” statistic, leveraging authority and a clear call to action: “Predict. Prevent. Profit. Request a Demo.” We also experimented with thought leadership carousel ads, positioning InsightEngine Pro as the solution to complex data challenges.

On Google Search, our approach was more direct. We focused on highly specific keywords like “AI churn prediction software,” “enterprise analytics platform,” and “customer retention AI.” Our ad copy highlighted free trials, demo requests, and whitepaper downloads, ensuring strong alignment between search intent and landing page content.

We spent a significant amount of time pre-testing these creatives. Using a small, geographically isolated test audience in the Midwest (specifically targeting marketing managers in Columbus, Ohio, for a week prior to full launch), we ran A/B tests on headline variations and primary visual assets. This early testing, though an extra step, always pays dividends. It revealed that a slightly more empathetic tone (“Stop Guessing, Start Knowing”) resonated better than overly aggressive sales language, increasing our preliminary CTR by 15% on LinkedIn.

Targeting: Precision Over Volume

Our targeting was meticulously defined. On LinkedIn, we combined job title targeting (CMO, VP Marketing, Head of Analytics) with industry filters (Technology, Finance, Healthcare) and company size (500-5000 employees). We also uploaded a custom audience list of decision-makers who had attended relevant industry webinars in the past six months, provided by the client. This first-party data was a goldmine.

For Google Search, beyond the core keywords, we implemented negative keywords aggressively from day one to avoid irrelevant traffic. Terms like “free analytics tools” or “small business CRM” were promptly added to our exclusion lists. We also geo-targeted major business hubs: New York, San Francisco, Chicago, and, of course, Atlanta.

Initial Performance: A Mixed Bag

The first three weeks of Project Atlas were a whirlwind. Here’s how the numbers looked:

The CPL was a disaster. At $165.71, we were more than double our target, and the estimated ROAS was nowhere near what we needed. While impressions and clicks were decent, the conversion rate was abysmal, hovering around 0.6% instead of our 5% goal. This was a red flag, screaming for immediate intervention.

What Worked (and What Didn’t)

What worked:

• Google Search Ads CTR: Our Google Search campaigns actually performed well on CTR, averaging 4.2%, exceeding our target. This indicated strong keyword targeting and compelling ad copy for high-intent users.
• First-Party Data Audiences: The custom audience list on LinkedIn, though small, generated a CPL of $85—still high, but significantly better than other LinkedIn segments. This reinforced my long-held belief: your own data is your most powerful asset.
• Video Testimonials: On LinkedIn, the video ads had a 0.9% CTR, outperforming static images (0.7%). People want to see real users.

What didn’t work:

• Broad LinkedIn Targeting: Our broader LinkedIn segments (job titles + industries) were generating extremely high CPLs, some upwards of $300. The sheer volume of impressions wasn’t translating into qualified leads. This was a classic case of chasing reach at the expense of relevance.
• Landing Page Performance: A deep dive into Google Analytics revealed a high bounce rate (70%) and low time on page (under 45 seconds) for users arriving from LinkedIn. The landing page, while visually appealing, seemed to lack immediate clarity or a strong enough value proposition for colder traffic.
• Lack of Nurture: We realized too late that our initial strategy relied too heavily on a single conversion point (demo request). Many users weren’t ready for a demo after a single ad interaction.

Optimization Steps: The Mid-Campaign Pivot

Facing these grim numbers, we convened an emergency meeting with the client. It was clear we needed a substantial pivot, not just minor tweaks.

1. Aggressive LinkedIn Audience Refinement: We paused all broad LinkedIn campaigns immediately. Instead, we created hyper-focused segments combining multiple layers: specific job titles, company size, and skills relevant to AI analytics (e.g., “Predictive Modeling,” “Data Science,” “Business Intelligence”). We also created a lookalike audience based solely on the high-performing custom list, rather than broader website visitors. My experience tells me that tighter targeting almost always beats broader, even if it means sacrificing some impressions.
2. Landing Page Overhaul: This was critical. We redesigned the primary landing page for LinkedIn traffic, making the headline more direct and benefit-driven (“Unlock 90% Churn Prediction: See InsightEngine Pro in Action”). We added a concise explainer video (under 90 seconds) and introduced a secondary, lower-commitment conversion option: a downloadable “Executive Guide to AI-Powered Retention” behind a gated form. This provided a softer entry point for prospects not yet ready for a demo.
3. Introducing Retargeting Sequences: We implemented a robust retargeting strategy. Anyone who visited the InsightEngine Pro landing page but didn’t convert was placed into a 7-day retargeting sequence across LinkedIn and the Google Display Network. These ads offered the “Executive Guide” or highlighted specific features with case studies. Those who downloaded the guide were then retargeted with ads pushing the demo.
4. Bid Strategy Adjustment: We switched from a “Maximize Conversions” bid strategy to a “Target CPA” strategy on Google Ads, setting a more realistic initial target CPA of $100 and allowing the algorithm to learn. On LinkedIn, we moved to manual bidding for our new, tighter segments, giving us more control over spend.
5. Sales Team Feedback Loop: We established a daily sync with the client’s sales team to get qualitative feedback on lead quality. This was invaluable. They reported that leads from the new “Executive Guide” were more engaged and better informed during initial outreach.

The Turnaround: Weeks 4-8 Performance

The impact of these changes was dramatic.

By the end of the campaign, we had delivered 1,550 leads in total, with an overall average CPL of $96.77. While still above the $75 target, it was a vast improvement from the initial $165.71. More importantly, the lead quality, as reported by sales, was significantly higher, leading to a projected ROAS of 1.3x—much closer to our goal. We even saw a 25% increase in conversion rate on the revised landing page, a testament to the power of aligning content with audience intent. According to a recent report by HubSpot, companies that nurture leads generate 50% more sales-ready leads at a 33% lower cost (HubSpot). This campaign perfectly illustrates that principle.

One editorial aside: don’t let anyone tell you that “set it and forget it” is a viable strategy in paid media. It’s a lie. The platforms are too dynamic, audiences too fickle, and competitors too aggressive. Constant vigilance and a willingness to completely overhaul your approach are non-negotiable.

Key Takeaways for Digital Advertising Professionals

This campaign reinforced several critical lessons for me and my team:

1. Audience Segmentation is Paramount: Don’t be afraid to narrow your focus. Broad targeting often leads to wasted spend and poor lead quality. It’s better to reach fewer, highly qualified prospects than a mass of indifferent ones.
2. Landing Page Optimization is a Continuous Process: Your landing page is the hinge point of your campaign. If it’s not performing, no amount of ad spend will fix it. Always test and iterate, providing clear value propositions and multiple conversion paths.
3. The Power of the Nurture Funnel: Not every prospect is ready to buy immediately. Offering lower-commitment conversion options (e.g., guides, webinars) and then nurturing those leads through retargeting is far more effective than a single-stage approach. This is where a robust CRM integration truly shines.
4. Data-Driven Decisions, Always: The initial performance was disheartening, but by dissecting the data, identifying the root causes, and implementing decisive changes, we were able to turn the campaign around. Never ignore what the numbers are telling you, even if it means admitting your initial strategy was flawed. We use tools like Supermetrics to pull data into custom dashboards, allowing for real-time analysis and faster decision-making.

The campaign’s budget was ultimately $150,000, yielding 1,550 conversions (leads), for an average Cost Per Conversion of $96.77. The total estimated ROAS, factoring in the sales cycle, reached 1.3x. We didn’t hit every target perfectly, but the sheer improvement from the initial performance was a victory in itself.

To truly excel in paid media, you must embrace the campaign teardown as an essential, ongoing practice, because understanding what went right—and more importantly, what went wrong—is the only path to sustained improvement.