Atlanta Gear Works: Boosting ROAS in 2026

It’s 2026, and marketers are sitting on piles of data from powerful platforms, but their return on ad spend is stubbornly flat. Everyone’s chasing new AI tools, but the real gains come from getting more value out of your existing ones with smarter ad optimization. So how can your team actually do this without having to rip and replace their entire tech stack?

The Case of “Atlanta Gear Works”: A Story of Underutilized Potential

Let me give you a concrete example: “Atlanta Gear Works,” a regional e-commerce business selling custom-machined parts to industrial clients across Georgia. Their marketing manager, Sarah Chen, was using the standard stack for years: Google Ads, Meta Business Suite, and a solid Salesforce CRM. They were spending good money and driving traffic, but their new customer conversion rate was stuck at 1.8%. “We had all these platforms, all this data, but it felt like we were just scratching the surface,” Sarah recounted. “The dashboards showed us what happened, but not why, and certainly not what to do next with our ad optimization efforts.”

This story is incredibly common. So many businesses have a great tech stack but don’t actually use the AI and automation features that are already built in. It usually comes down to two things: a failure to properly integrate the platforms and a general hesitation to trust what the algorithms recommend. A Q4 2025 IAB report even found that nearly 40% of advertisers admitted they weren’t fully using the AI features in their main ad platforms.

Unlocking Granular Audience Segmentation

The first thing I had Sarah do was go way deeper into their audience data. Previously, their Google Ads campaigns were targeting really broad categories: “industrial parts,” “manufacturing equipment,” and “B2B suppliers.” While these brought impressions, they also attracted a lot of garbage clicks from people who’d never buy. Inside Google’s “Audience Manager”, a tool they weren’t using much, I had her import their customer lists straight from Salesforce to create Customer Match audiences. This was a huge step because it let them build lookalike audiences based on their absolute best, highest-value clients.

Over on Meta, we stopped targeting general demographics and started building segments based on actual website engagement. For example, instead of a generic “small business owners” audience, we created a custom audience of people who visited their “heavy machinery components” page three or more times in the last 60 days but didn’t buy anything. This took a small tweak to their Meta Pixel implementation to track these more specific events, but the results were immediate. Click-through rates (CTR) on those retargeting campaigns jumped from 1.5% to over 3.2% within a month.

Intelligent Bidding Strategies: Beyond Manual Adjustments

A huge chunk of their budget was being burned on manual bidding, a time-suck of a process that’s almost always too slow to react to what’s happening in the auction. “I spent hours every week adjusting bids, and it felt like I was always a step behind,” Sarah admitted. It’s a classic mistake. The AI in these platforms can make real-time adjustments at a scale no human, or even a team of humans, could ever hope to match.

So, we moved their Google Ads campaigns off manual CPC (cost-per-click) and onto Maximize Conversion Value with a target ROAS. This strategy is right there in the campaign settings and uses machine learning to predict not just if a click will convert, but what its value might be. For a business like Atlanta Gear Works, where the lifetime value of a customer is substantial, this was a big deal. We set an initial target ROAS of 300% (aiming for $3 in value for every $1 spent), and within eight weeks, their campaign ROAS hit 350% as their cost per acquisition (CPA) decreased by 18%.

On Meta, we did something similar by switching their prospecting campaigns to Lowest Cost with a bid cap. This let the algorithm hunt for the cheapest conversions while making sure a single bid couldn’t get out of control. That mix of letting the machine do the work while you set the guardrails is where you find the real power.

Creative Optimization Through A/B Testing and AI Insights

Creative gets stale fast, especially in B2B where the messaging can get repetitive. It’s a constant problem. Atlanta Gear Works had relied on a few static ad creatives for months. Their platforms already had great tools for A/B testing and getting AI-driven feedback on creative performance, but they weren’t using them.

In Google Ads, we started using the Responsive Search Ads feature properly. Instead of one or two headlines, Sarah started feeding the system 15 different headlines and 4 descriptions for each ad group, giving the AI plenty of material to test. The AI then mixed and matched them, quickly figuring out which combinations worked best and giving recommendations on specific phrases that were resonating, which boosted their ad strength scores by 10%.

For Meta campaigns, we turned on the Dynamic Creative option. This let them upload a bunch of different assets, images, videos, text, and calls-to-action, and Meta’s AI built the best-performing ad for each specific person viewing it. The system was learning in real-time instead of us just guessing what would work. Sarah committed to creating at least three distinct visual assets and three different primary texts for each ad set, and the AI quickly showed that photos of their actual machinery in action crushed generic stock photos, delivering a 20% higher engagement rate. You simply can’t iterate that fast and get that kind of data-backed feedback without the algorithm doing the heavy lifting.

The Power of Integrated Data: CRM to Ad Platform

But the biggest performance lift came when we finally connected their Salesforce CRM data directly to the ad platforms. Historically, they had tracked website form fills as “conversions,” but that data never made it back to the ad platforms to show which leads turned into actual sales in Salesforce. As a result, Google and Meta were optimizing for “website leads,” which is a completely different (and much lower-value) goal than “qualified sales opportunities” or actual paying customers.

We set up Enhanced Conversions for Web in Google Ads and Conversions API (CAPI) on Meta. These technologies let them securely send hashed customer data from their website and CRM back to the platforms, building a much more durable picture of the customer journey, which is critical now that third-party cookies are on the way out. By uploading offline conversion data from Salesforce, they were finally able to tell the ad platforms which specific leads resulted in a sale and for how much. This let the bidding algorithms chase actual revenue, not just cheap website form fills.

Think about it: a lead from a Google Search ad might look small at first. But if that same person later signs a $50,000 deal in Salesforce, feeding that final sale value back to Google Ads completely changes the algorithm’s definition of a “good lead.” It started proactively seeking out users who looked like their other high-value customers, and within six months, the average order value from their ad-generated leads went up by 15%.

A Continuous Cycle of Learning and Adjustment

This wasn’t a set-it-and-forget-it project for Atlanta Gear Works. Sarah built a weekly routine to check the AI-powered suggestions in Google Ads’ “Recommendations” tab and Meta’s “Ad Reporting” section. She learned to separate the genuinely useful suggestions from the minor, low-impact tweaks. “The key,” she observed, “is to not just blindly accept every recommendation, but to understand the underlying data and test changes incrementally.”

They also started using the Experiments feature in Google Ads to properly test big changes, like a new bidding strategy or audience segment, against their control campaigns. This gave them statistical confidence before they rolled anything out across the board. This disciplined approach, just using the AI they already had, turned their ad budget from a line-item expense into a real growth driver. After a year of this, their new customer conversion rate hit 3.1%, and their overall ROAS reached 420%.

If you’re a marketer feeling buried by your ad platforms, my advice is to just start small. Pick a single campaign and one feature you aren’t using (like dynamic creative or an automated bidding strategy) and just commit to testing it for a defined period. The tools you have are already powerful. The job is to learn how they think and start trusting them to do the math.

Putting AI to work in your ad platforms gives you more than just automation. It gives you insights that let you make better, data-backed decisions at a much larger scale. For Atlanta Gear Works, it all came down to using their existing tech to get real, measurable results, all without buying some new, complicated AI solution. Often, the solution is simply a better understanding and application of the tools you already pay for.

In the end, getting more from your ad platforms and their integrated AI means you have to adopt a mindset of constant experimentation and learn to trust the data. The huge jump in performance Atlanta Gear Works saw proves that applying these tools strategically can completely change your ad results and deliver more value from every single dollar you spend.