NexusAI’s 3.2x ROAS: Ad Tech Trends for 2026

Ad tech moves fast, so founders and marketing teams have to keep up. The only real way to see what’s coming next is to look at how good campaigns get built, run, and tweaked in the real world. So, let’s break down a recent six-week campaign for “NexusAI,” a B2B SaaS tool for e-commerce predictive analytics, and figure out what made it work.

Campaign Overview: NexusAI Launch

For NexusAI, the goal was simple: get quality leads from mid-market e-commerce shops in North America. We had a six-week window, from February 1 to March 15, 2026, to make it happen. The NexusAI product itself is built to solve real problems, specifically for e-commerce companies struggling with stockouts or wasting marketing money on bad customer segmentation. With a total paid media budget of $75,000, we were aiming for a Cost Per Lead (CPL) under $150 and a Return on Ad Spend (ROAS) of at least 2.5x, which felt realistic based on the product’s projected customer LTV.

Strategy and Core Pillars

We built our strategy around two core ideas:

1. Generate Demand on LinkedIn: We had to show NexusAI as the expert solution for the headaches of e-commerce operations. For this, we leaned heavily on LinkedIn Lead Gen Forms to make signing up as painless as possible.
2. Capture Intent on Google Search: We needed to be there when people were actively looking for solutions to inventory management, predictive analytics, or customer churn problems.

Any new SaaS launch has to do two jobs at once: teach people what the product is and get them to sign up. The first few weeks were all about building awareness and grabbing those initial leads, knowing we’d then shift gears to retargeting and follow-ups in the back half of the campaign.

Creative Approach: Education Meets Urgency

Our creative had to be different for each platform. On LinkedIn, we ran a series of short-form video ads showing a simulated NexusAI dashboard in action, with voiceover testimonials from some early beta users. These 30-45 second clips were focused on specific pain points, with titles like “Dynamic Stock Level Alerts” and “Personalized Customer Journey Mapping.” We also tested carousel ads that pulled key stats from an internal whitepaper, like “Reduce stockouts by 20%.” Google Search was all about direct problem-solution copy. We ran Responsive Search Ads (RSAs) with headlines like “Stop Inventory Guesswork, Try NexusAI” and “Predict Customer Churn with AI.” Giving Google’s AI a bunch of different headlines and descriptions to chew on let it find the best-performing combinations, and one headline that consistently killed it was “AI-Powered Inventory Forecasts” paired with a direct CTA like “Book a Demo Today.”

Targeting Breakdown and Implementation

LinkedIn Targeting: LinkedIn’s B2B targeting is its main strength, so we used it to find our ideal customer. Our primary audience was built on:

• Job Titles: E-commerce Manager, Director of Operations, Supply Chain Manager, Head of Digital Marketing.
• Company Size: 51-500 employees (our mid-market sweet spot).
• Industry: Retail (Online), Apparel & Fashion, Consumer Goods.
• Skills: Inventory Management, E-commerce Operations, Data Analytics, Demand Planning.

We also uploaded a seed list of 500 existing trial users and created a lookalike audience from it. That lookalike group turned out to be gold for scaling up our reach. Google Search Targeting: We didn’t just throw money at one type of campaign. Our Google Ads setup was layered:

• Branded Search: Simple defense, making sure we owned our own brand name.
• Generic Keywords: We went after broad terms like “inventory management software” and “e-commerce forecasting.”
• Competitor Keywords: We bid directly on the names of established competitors. It was a gamble with a higher CPC, but you can’t argue with the high intent of someone searching for your competitor by name. Of course, we used a ton of negative keywords like “free” and “personal use” to filter out junk traffic.
• Custom Intent Audiences: This was huge for us. We created custom audiences by targeting people who had recently searched for our competitors or visited their sites. This gave us a powerful intent signal to layer onto our campaigns, even though we mostly stuck to search for this particular launch.

Campaign Performance: What Worked and What Didn’t

After six weeks, we’d spent the full $75,000 and brought in 600 qualified leads. Here’s how the numbers broke down between the first and second half of the flight.

What Worked Exceptionally Well

The video testimonials we ran on LinkedIn were the hands-down winner. Their average CTR was 1.8% compared to just 0.7% for our static image ads, and people were actually watching them, with a 75% average completion rate. This alone dropped our LinkedIn CPL from a painful $180 down to $110 by the end of week three. People just connect with seeing someone else talk about how the product actually helped them. And those LinkedIn lookalike audiences converted 25% better than the standard interest-based targeting. Over on Google Ads, going after competitor keywords with laser-focused ad copy paid off big time. Sure, the CPCs were higher, averaging $8.50, but the conversion rate was 18%, giving us a Cost Per Conversion of just $47. That blew away the generic keyword campaigns, which were converting at 9% for a $70 Cost Per Conversion. You can’t fake that kind of intent, these people were already comparing their options. The retargeting we ramped up in the second half of the campaign also really moved the needle, creating an audience of people who’d hit the demo page but didn’t sign up and showing them an ad with a “Limited-Time Free Trial Extension” offer. That small segment converted at 15%.

Challenges and Underperformers

At first, our generic keyword campaigns on Google were a mess. They were bringing in irrelevant clicks and driving up the CPL for the first couple of weeks. We saw a lot of traffic from broad searches like “e-commerce data analysis” from people who just weren’t ready to think about a SaaS demo. That forced us to move money around, fast. We also had one specific ad on LinkedIn that completely bombed. It was an infographic about “5 Ways AI Boosts E-commerce ROI.” The CTR was a miserable 0.5% and it generated almost no leads. My gut feeling? It was just a wall of text that people scrolled right past. It couldn’t grab their attention and just didn’t have the stopping power of a video.

Optimization Steps Taken

We had to optimize on the fly. Here’s what we changed mid-campaign:

1. Budget Reallocation: By the end of week two, we saw the writing on the wall. We pulled 20% of the budget from the failing generic Google Search campaigns and pushed it into the competitor keyword campaigns and our LinkedIn video ads. That one change immediately started bringing the overall CPL down.
2. Ad Copy Refinement: On Google Search, we paused a lot of our broad match keywords that were pulling in junk queries and built out our exact and phrase match lists. We also rewrote our RSA headlines to focus on specific pain points like “Reduce Stockouts” instead of vague benefits.
3. Landing Page A/B Testing: We ran a simple A/B test on the lead gen landing page. The winning version had a 90-second explainer video embedded at the top and a simplified form, we cut it from 7 fields down to 5. That alone gave us a 12% lift in conversions, so we rolled it out everywhere by week four.
4. Negative Keyword Expansion: We were glued to the search query reports in Google Ads. All in all, we added over 200 negative keywords during the six weeks to stop burning cash on bad searches for things like “e-commerce AI free” or “shopify plugins.”
5. Creative Refresh: We killed the underperforming LinkedIn infographic and replaced it with a new video that focused on a single, powerful stat about inventory optimization. The new video hit a 1.1% CTR right away, which just proves people want to see something move.

The campaign worked because we didn’t just set it and forget it. We made data-backed changes every week. It just goes to show that your best-laid plan is useless if you’re not watching the numbers and ready to change course when an infographic bombs or a video testimonial takes off. In the end, the NexusAI launch hit a 3.2x ROAS with an average $104.17 CPL, proving that a specific, flexible plan is what gets results. The main lesson here is that you have to test everything, all the time. Your initial assumptions are just a starting point for the first A/B test. Ad tech founders have to do two things: build good platforms and prove they work with smart, measurable marketing.