TLDR;
- Our Strategic Fix: We built a multi-layered Meta Ads funnel (TOFU/MOFU/BOFU) to systematically test audiences and creatives, moving beyond a scattered approach to create a predictable system for acquiring new software users.
- Verified Results: We generated 4,622 B2B software trial registrations at an average cost of just $2.38 each, spending a total of $11,013.42.
- Key Takeaway: Consistent, low-cost acquisition for SaaS isn't about finding one secret audience. It’s about engineering a full-funnel system that systematically finds and converts users at different stages of awareness, then ruthlessly optimising based on hard data.
Most B2B software companies I talk to are stuck. They have a solid product, but their user acquisition is a mess of unpredictable spikes and troughs. They'll have a good week, then two bad ones. They'll throw money at a new campaign that fizzles out, with no real understanding of why. The goal is always the same: a predictable, scalable way to get new trial users at a cost that makes sense. But achieving it feels like chasing a ghost.
The core problem isn't a lack of effort; it's a lack of system. Without a structured approach, you're just gambling. This was the situation for our client, Threecolts, with their product FeedbackWhiz—a powerful software suite for Amazon sellers looking to manage feedback, reviews, and orders. They needed a reliable engine to drive trial signups, specifically one that could consistently bring in new users for under three dollars a pop. A tough ask in the competitive B2B software space.
We took on the challenge. The result? We built them that engine. It generated 4,622 new trial registrations at an average cost of just $2.38 per signup. This wasn't down to a lucky guess or a single magic ad. It was the result of a deliberate, multi-platform strategy, a meticulously structured funnel, and a relentless process of testing and optimisation. Here's a detailed breakdown of how we did it.
It All Starts With a Plan, Not an Ad Campaign
The biggest mistake I see is people jumping straight into Ads Manager and starting to build campaigns. That's like a builder starting to lay bricks without a blueprint. You're guaranteed to build something wobbly that will eventually fall over. Before we spent a single dollar of our client's budget, we went back to basics and created a detailed campaign strategy plan.
This isn't just a document that gathers dust. It's the operational playbook for the entire project. It forces you to answer the tough questions upfront.
First, we dug into competitive research. What were other Amazon seller tools doing? What angles were they using in their ads? What offers were they pushing? This isn't about copying them; it's about understanding the landscape so you can find a unique position. We saw a lot of feature-focused ads, which gave us an opening to focus more on the core problems of the Amazon seller: saving time, protecting their seller rating, and increasing sales.
Next, we mapped out the entire user funnel. A typical user journey isn't a straight line from ad to purchase. It's winding. People see an ad, get distracted, visit the website later, forget, see another ad, then maybe sign up. Our campaign structure had to account for this reality. We designed a multi-stage funnel to address users at every point of their journey:
- Top of Funnel (TOFU): This is for brand new audiences—Amazon sellers who had likely never heard of FeedbackWhiz. The goal here is simple: grab their attention and get them to the website.
- Middle of Funnel (MOFU): This was for the warm audience. People who visited the website or engaged with an ad but didn't sign up. They're interested, but not yet convinced. Our job here is to stay on their radar and give them another reason to come back.
- Bottom of Funnel (BOFU): This targeted two groups. First, people who started the signup process but abandoned it. Second, users who had already signed up for a trial but weren't active, with the aim of encouraging them to engage with the software and ultimately convert to a paid plan.
Finally, the plan detailed our entire testing and optimisation strategy. We outlined exactly what we would test: audiences, locations, creatives, ad copy. And we defined how we would measure success, looking beyond just the cost per registration to metrics like click-through rates (CTR) and link cost-per-click (CPC) to diagnose performance at every stage. This systematic approach meant our decisions would be driven by data, not guesswork.
Our B2B SaaS Funnel Architecture
Top of Funnel (TOFU): Prospecting Cold Audiences
Goal: Drive new, relevant traffic. We targeted cold but high-potential audiences of Amazon sellers using Lookalikes of existing customers and detailed interest targeting (e.g., 'Small Business Owners', 'E-commerce').
Middle of Funnel (MOFU): Retargeting Warm Interest
Goal: Re-engage interested prospects. This layer targeted people who had visited the FeedbackWhiz website or engaged with our ads but hadn't yet started a trial. The aim was to bring them back to complete the registration.
Bottom of Funnel (BOFU): Activating Trial Users
Goal: Convert trial users to paid customers. Ads at this stage reminded users of key features, offered help, and encouraged them to explore the software to see its full value before their trial ended.
Finding the Right People: A Deep Dive Into Targeting
Once the funnel structure was in place, the real work began: finding the audiences that would fill it. For a niche product like FeedbackWhiz, targeting is everything. Showing ads to the wrong people is just a fast way to burn money. We set up a rigorous testing framework to identify the most profitable audiences on Meta (Facebook and Instagram).
Our hypothesis was that Lookalike audiences would be our strongest performers. Lookalikes are where Meta's algorithm finds new people who share characteristics with an existing audience of yours. They're incredibly powerful if you give them the right source data. We didn't just create one generic Lookalike; we tested a whole hierarchy of them, based on how valuable the source audience was:
- Lookalikes of past purchasers
- Lookalikes of users who started a trial
- Lookalikes of users who completed registration
- Lookalikes of all website visitors
The logic here is simple: someone who has already paid for the software is a much stronger signal to the algorithm than someone who just visited the homepage. By testing these systematically, we could see which signals produced the most cost-effective new signups. And the results spoke for themselves—Lookalikes were consistent winners throughout the campaign.
Alongside Lookalikes, we tested a range of detailed targeting options. This is where many advertisers go wrong. They target massive, broad interests like "Amazon" or "E-commerce." The problem is that these audiences include millions of consumers who just shop on Amazon, not the sellers we needed to reach. It’s too diluted.
We focused on interests and behaviours that were more specific to our ideal customer. The standout performer was an audience of "Small Business Owners." This makes sense; many Amazon sellers are exactly that. We also layered in interests related to the e-commerce world, such as 'Shopify', 'WooCommerce', and 'Online marketplace', to narrow the focus further. We found that these more specific, business-oriented interests dramatically outperformed the broader, consumer-focused ones.
Just as important as finding what worked was identifying what didn't. We tested broad targeting (letting Meta's algorithm run free with minimal input) and a variety of other detailed interest combinations. They just didn't perform. The cost per registration was too high. So, we did what any data-driven advertiser should do: we turned them off. We were ruthless about cutting the underperformers and reallocating that budget to the audiences that were delivering results. This constant pruning and feeding is the key to maintaining a low CPA as you scale.
Geography Matters: The Worldwide vs. Developed Countries Test
Another critical variable we tested was location. It's tempting to target only high-value markets like the US or the UK, but this often comes with much higher advertising costs due to competition. On the other hand, targeting worldwide can bring down your average cost per result, but you risk getting lower-quality signups from regions where customers are less likely to convert to a paid plan.
We set up separate campaigns to test this head-on:
- A campaign targeting the United States only.
- A campaign targeting a list of "Developed Countries" (including Western Europe, Canada, Australia, etc.).
- A campaign targeting Worldwide.
The data gave us a clear picture. The "Worldwide" campaign was by far the cheapest in terms of cost per registration. One of our most successful TOFU campaigns, targeting worldwide, brought in over 1,000 registrations at a staggering cost of just $0.31 each. In contrast, the campaign targeting "Developed Countries" had a cost per registration of $5.76. The US-only tests were even more expensive.
This presented a strategic choice. For the primary goal of driving trial signups at the lowest possible cost, the Worldwide campaign was the undisputed winner. This became the workhorse of our top-of-funnel strategy. However, it was crucial for the client to monitor the backend metrics—how many of these low-cost signups from different regions eventually converted to paying customers? This data allows for a more sophisticated, long-term strategy, where you might accept a higher upfront CPA from a country like the US if you know its users have a much higher lifetime value (LTV). For this campaign's objective, though, a low-cost entry into the funnel was the priority, and the data showed worldwide targeting was the way to achieve it.
Creative is King: Designing Ads That Convert
Even with perfect targeting, your campaigns will fail if your ads are rubbish. The creative—the image or video—is what stops the scroll, and the copy is what earns the click. We followed a strict testing methodology here as well, never relying on a single ad to do all the work.
Our creative strategy was rooted in the strategy plan. For a B2B software product, ads that show the product in action and clearly articulate its value proposition work best. We tested a variety of ad formats, including single images and short videos, that featured screenshots of the FeedbackWhiz dashboard. The goal was to make the solution tangible. Instead of just saying "manage your feedback," we could show them the clean, organised interface they would be using.
The ad copy was direct and benefit-driven. We tested multiple headlines and primary text versions for each ad, focusing on the core pain points of an Amazon seller. Taglines like "Automate your Amazon feedback" and "Unlock Amazon success" spoke directly to their goals. We used strong calls-to-action (CTAs) like "Sign Up" or "Learn More" to make it clear what the next step was. This combination of clear visuals and benefit-focused copy was key to driving high click-through rates and, ultimately, low-cost registrations.
Here's a quick walkthrough of the account structure and some of the ad creatives that drove these results.
Our optimisation process was continuous. We didn't just launch the ads and hope for the best. We used a process where we would test multiple ads within each ad set. After a few days, enough data would come in to see which ones were resonating and which were not. We'd look at the aggregate performance of an ad across *all* the ad sets it was running in. If an ad was consistently underperforming everywhere—lower CTR, higher CPC—we'd turn it off. The budget it was using would then automatically be allocated by Meta to the better-performing ads in that ad set. This iterative process ensures that the account becomes more and more efficient over time, as the budget naturally flows to the winning creative and copy combinations.
The Final Tally: Analysing the Campaign Performance
At the end of the day, the only thing that matters is the results. The entire system—the funnel structure, the audience testing, the creative optimisation—was designed to produce one outcome: a high volume of trial signups at a very low cost. And the numbers show it worked.
Across all campaigns, we spent $11,013.42 and generated 4,622 completed trial registrations. This gives us our headline figure: an average cost per registration of just $2.38, well below the client's $3.00 target.
But the aggregate numbers only tell part of the story. Breaking down the performance by campaign reveals the power of the funnel structure we implemented. The table below shows the results from our main Meta Ads campaigns.
Meta Ads Campaign Performance Breakdown
This table details the performance of each campaign, showing how our multi-funnel and multi-location testing strategy contributed to the overall result.
| Campaign | Registrations | Cost Per Reg. | Amount Spent |
|---|---|---|---|
| LH - MOFU - Complete Registration | 1,122 | $0.88 | $987.93 |
| LH - TOFU (Worldwide) - Complete Registration | 1,083 | $0.31 | $336.75 |
| LH - TOFU (Worldwide) - Trials | 1,013 | $0.21 | $217.57 |
| LH - Promotions (Developed Countries) | 419 | $9.29 | $3,890.48 |
| LH - TOFU (Developed Countries) - Complete R... | 232 | $5.76 | $1,337.42 |
| LH - TOFU (Developed Countries) - Trials | 230 | $3.84 | $884.13 |
| Other Campaigns (BOFU, US Tests, etc.) | 523 | Varies | $3,359.14 |
| Total | 4,622 | $2.38 | $11,013.42 |
A few key insights jump out from this data. Our MOFU retargeting campaign was a massive success, bringing in 1,122 registrations from warm traffic at a cost of just $0.88 each. This proves the value of having a dedicated middle-funnel layer. These are people who were already aware of the product, and a simple, timely reminder was all it took to get them over the line. Leaving this audience untapped is like leaving money on the table.
The TOFU Worldwide campaigns were our volume drivers, bringing in over 2,000 registrations combined at an almost unbelievably low cost. As discussed, this was the key to hitting the overall CPA target. It's also worth noting the impact of a budget cut the client experienced towards the end of our main push. This is a real-world scenario many businesses face. It explains why some of the campaigns have lower spend than you might expect. Our structured approach allowed us to react quickly, consolidating the remaining budget into the absolute top-performing campaigns and ad sets to maintain efficiency even on a reduced spend.
Thinking Bigger: The Multi-Platform Acquisition Model
While this case study focuses on the outstanding results we achieved on Meta, our initial strategy was always designed as a multi-platform model. The brief was to build a predictable *acquisition model*, not just a predictable Facebook campaign. For a B2B SaaS company, relying on a single channel is risky.
Our strategy plan included a full build-out for Google Ads. Why? Because Google Ads captures a different kind of user: one with active intent. While Meta is great for generating demand and reaching people who *might* be interested, Google Search is for people who are *actively looking* for a solution right now. They're typing things like "best amazon feedback software" or "how to remove negative amazon review" into the search bar. Capturing that intent is incredibly valuable.
Our research for Google Ads identified several key themes to target:
- Competitor Keywords: Targeting the names of other tools in the space (e.g., Helium10, Sellerboard).
- Problem-Based Keywords: Targeting searches related to the core problems FeedbackWhiz solves (e.g., "how to get reviews on amazon", "amazon feedback removal").
- Solution-Based Keywords: Targeting high-intent searches from people already looking for software (e.g., "top amazon feedback tools", "amazon review software").
We planned to use a mix of Search campaigns to capture this direct intent, and Performance Max/Discovery campaigns to reach relevant audiences across Google's wider network (YouTube, Gmail, Discover feed). The plan was to launch on Meta first to validate the offer and messaging at a lower cost, then expand to Google to layer on high-intent traffic and build a more resilient, multi-channel acquisition engine. This phased approach allows you to use learnings from one platform to inform the strategy on another, making your overall marketing spend far more effective.
The Final Word: It's a System, Not a Secret
Getting to 4,622 software registrations at $2.38 a pop wasn't about finding a single "growth hack" or a magic audience that no one else knew about. If anyone tells you that's how paid advertising works, they're either lying or they don't know what they're doing. These results were the outcome of a methodical, disciplined process.
It was about building a logical funnel structure that respected the user's journey. It was about creating a rigorous testing framework to systematically discover the best-performing audiences, creatives, and locations. It was about being ruthless with the data, turning off what didn't work and doubling down on what did. It's a system of continuous improvement.
For any B2B software company struggling with unpredictable growth, the lesson here is to stop looking for a silver bullet and start building a machine. Define your strategy, structure your funnel, test everything, and let the data guide your decisions. That's how you move from gambling to engineering, and it's the only way to build a truly predictable and scalable customer acquisition model.
If you're tired of the guesswork and want to see how a systematic approach could work for your business, we offer a free, no-obligation strategy consultation. We can take a look at what you're doing now and map out a plan to get you on the right track.