My thoughts: if you're running ads for a B2B software with a 14-day or 30-day free trial, you're likely setting your ad account up to fail from day one. I see founders constantly trying to optimise their Meta campaigns for the final 'Purchase' event. It sounds sensible enough on paper. You want paying customers, so you tell the algorithm to find paying customers. But here is the dodgy bit nobody tells you: Meta's attribution window cuts off after 7 days from the click.
If a user clicks your ad, signs up for your 14-day trial, and pays you on day 14, Meta never sees that purchase. The pixel starves. The algorithm assumes the ad is rubbish, your costs skyrocket, and your campaigns grind to a halt. You can't scale further because the system simply doesn't have the signals to learn who your best buyers are.
I've run quite a few campaigns for B2B SaaS, and bridging this attribution gap is always the main hurdle we have to clear. You need to feed the pixel enough conversion volume within that 7-day window to keep costs low, whilst ensuring those cheaper leads actually convert into paying users on the backend. To show you exactly how this works in practice, I'm going to tear down the data from a B2B software campaign we ran. We drove 4,622 registrations for an Amazon seller feedback tool at just $2.38 per registration.
I'll walk you through the methodology, the exact campaign structures, the targeting that worked, and how you can engineer your own ad account to bridge the gap between a free trial and a paid conversion.
What does the actual dataset look like?
Before we crack on with the analysis, let's be crystal clear about the data we are looking at. This isn't some aggregated industry average or a mashup of hundreds of accounts. This is a forensic look at a single Meta ads account export for one B2B software client over a specific date range.
The product is an Amazon feedback tool designed to help eCommerce sellers streamline their review requests and improve their feedback scores. The objective was to get Amazon sellers to register for the software.
Here are the raw facts from the account export we will be analysing:
- Total Ad Spend: $11,013.42
- Total Completed Registrations: 4,622
- Average Cost Per Registration: $2.38
- Total Ad Impressions/Reach: 490,084 reach across the main campaigns
- Campaigns Tested: 12 distinct campaigns spanning TOFU (Top of Funnel), MOFU (Middle of Funnel), and BOFU (Bottom of Funnel).
Because this data comes from a single account, the findings represent real-world observations from this specific dataset. They may not generalise to a consumer mobile app or a high-ticket enterprise sales cycle. But if you are selling B2B software to SMEs, the structural mechanics of how we got these costs are highly relevant.
The B2B SaaS Attribution Gap
Why does regional targeting dictate your frontend costs?
When we look at the raw data for this account, the first thing that jumps out is the extreme variance in Cost Per Acquisition (CPA) based entirely on location targeting. If you just look at the blended average of $2.38, you are missing the mechanics of how the algorithm prices inventory.
We ran several Top of Funnel (TOFU) campaigns aimed at getting new, cold traffic to complete a registration. Let's look at the breakdown:
- TOFU (Worldwide) - Complete Registration: 1,083 registrations at $0.31 each.
- TOFU (Worldwide) - Trials: 1,013 registrations at $0.21 each.
- TOFU (Developed Countries) - Complete Registration: 232 registrations at $5.76 each.
- TOFU (US) - Complete Registration: 24 registrations at $38.57 each.
You can see exactly what happens when you constrain the ad platform. When we asked Meta to find registrations strictly in the United States, the cost skyrocketed to $38.57 per registration. The inventory there is fiercely competitive, especially for B2B audiences. When we opened it up to all Developed Countries, the cost dropped to $5.76. When we went completely Worldwide, costs plummeted to $0.21 and $0.31.
It is very tempting to look at a $0.21 software registration and think you have cracked the code. But fair play, you have to be careful here. In developing countries, the CPCs are much lower (often £0.10 to £0.50), but the traffic can be of much lower quality. You might get thousands of trial signups, but if they don't have the purchasing power to pay your monthly SaaS subscription, your backend Trial-to-Paid conversion rate will be dismal.
Registration Cost by Location Targeting
Audited cost per completed software registration across isolated geographical targeting parameters.
My advice here is matter-of-fact: if you don't care about quality, run a worldwide campaign and let the algorithm gorge on cheap clicks. But if you need actual revenue, we usually start with the countries where most of your existing ideal customers are located. You will pay more for the trial signup, but your backend conversion rate will sort itself out.
Why do lookalikes of trials outperform lookalikes of purchases?
When we dive into the specific ad sets running within these campaigns, we can see exactly what audiences drove the results. A lot of folks assume that Lookalike audiences are dead on Meta, or that broad targeting is the only way to go. The data here tells a slightly different story.
We tested 9 different ad sets inside one of the main campaigns. Here are the registration volumes they generated:
- Broad: 232 registrations
- LLA 1% Complete Registration: 150 registrations
- Online Retail Interests: 38 registrations
- LLA 1% Trials: 27 registrations
- Small Business Owners: 13 registrations
- LLA 1% Purchases: 4 registrations
Notice that the "LLA 1% Purchases" only generated 4 registrations, while the Lookalikes built on top-of-funnel events (Complete Registration and Trials) generated 150 and 27 registrations respectively.
This brings us right back to the core problem. Because the SaaS trial period is longer than 7 days, the Meta pixel simply doesn't capture enough 'Purchase' events to build a robust Lookalike audience. If you feed the algorithm a custom audience of just a handful of tracked purchasers, the 1% Lookalike it creates will be heavily skewed or outright broken. Conversely, because we are feeding the pixel thousands of 'Complete Registration' events, it has a massive, highly accurate dataset to build a Lookalike audience from. In this dataset, Lookalikes built on mid-funnel volume completely outclassed Lookalikes built on low-volume backend events.
We also tested interest groupings. Targeting "Small Business Owners" yielded a measly 13 registrations. Why? Because it is too broad. If you are selling an Amazon seller tool, targeting generic small business owners means you are mostly paying to show ads to local plumbers, cafe owners, and consultants who have zero use for Amazon software. The "Online Retail Interests" ad set performed nearly three times better (38 registrations) because it contained people actually interested in eCommerce and marketplaces.
How do you calculate the unit economics of the attribution gap?
So, we know we have to optimise for the trial registration to get enough volume. But how do we know if a $2.38 or a $5.76 cost per registration actually makes sense for the business? You have to model the drop-off.
Let us say you take the "Developed Countries" campaign cost of $5.76 per registration. If your software converts 15% of free trials into paying users, you are not paying $5.76 for a customer; you are paying $38.40 ($5.76 / 0.15). If your software costs $49/month and users stay for an average of 12 months, your Lifetime Value (LTV) is roughly $588. Paying $38.40 to acquire $588 is an absolute no-brainer.
But if you are running the US-only campaign at $38.57 per registration, and your trial-to-paid conversion is only 5%, your actual Cost Per Acquisition (CPA) for a paying user jumps to $771.40. Suddenly, you are losing money on every customer you acquire.
I have built a multi-stage diagnostic calculator below based precisely on the metrics from our dataset. You can toggle the initial registration volume and see exactly where the value leaks out of the funnel.
SaaS Trial to Paid Bottleneck Diagnostic
Input your trial registration volume to model how backend conversion rates impact your final yield of paying users.
How should you restructure your campaigns into a three-stage funnel?
You can not just throw up one campaign, optimise for trials, and call it a day. The people who sign up for your trial need to be pushed to actually use the software and eventually buy it. This is where your campaign structure comes in.
In the dataset, you will see we did not just run TOFU campaigns. We ran a strict, layered funnel to mop up the leakage:
Budget Allocation by Funnel Stage
How the $11,013.42 ad spend was distributed across Top, Middle, and Bottom of funnel campaigns in the dataset.
The vast majority of the budget goes into the TOFU campaigns to find new people who have never heard of the product. The goal here is just to get them to the destination page.
Then we have the MOFU (Middle of Funnel) campaign. In our dataset, the "LH - MOFU - Complete Registration" campaign spent $987.93 and drove 1,122 registrations at an incredibly cheap $0.88 each. These are people who visited the website or engaged with the ads but did not sign up initially. Retargeting them is incredibly cheap and highly effective because they already have brand context.
Finally, and most importantly to bridge the attribution gap, we ran BOFU (Bottom of Funnel) campaigns. Look at the "LH - BOFU - Purchase" campaign. It spent $205.49 and generated 36 logged results at $5.71 each. The targeting for this campaign? People who have already registered an account (the free trial users). The objective here is not to get net-new users; it is to stay on their minds throughout their trial and push them to eventually make a purchase.
This is how you solve the 7-day attribution problem. You use the cheap TOFU and MOFU campaigns to drive the high-volume trial events that Meta's algorithm needs to learn. Then, you use a dedicated BOFU campaign to nurture those trialists outside of the 7-day window. You track the final outcome in your own CRM (like HubSpot or Salesforce), accepting that Meta will never fully take credit for the final sale.
What kind of creative strategy actually converts B2B buyers?
You can have the best funnel in the world, but if your ads are rubbish, your costs will still be astronomical. In the dataset, we split-tested 7 different ads simultaneously. The performance was not evenly distributed.
The clear winner was "LH - Automate your Amazon feedback - Copy", driving a staggering 1,046 registrations. The runner up, "LH - Unlock amazon success - Copy", drove 633 registrations. Meanwhile, "LH - Email and alerts - Copy" only managed 2 registrations before we binned it.
Why the massive disparity? It comes down to pain. The winning ad focused specifically on a painful, manual process ("Automate your Amazon feedback"). It isolated an urgent nightmare the seller has—wasting hours manually begging for reviews while their seller rating tanks. The worst performing ad ("Email and alerts") just listed features. Nobody cares about features. They care about their problems.
Your ads need to speak directly to the problems of your ideal customers. You do not sell an Amazon seller tool; you sell a good night's sleep and protection from an Amazon account suspension. Ads that include screenshots or recordings of the product in action often work best because they prove the solution exists and is easy to use.
Paid Ad Diagnostic Audit & Action Blueprint
The exact strategic steps to restructure a failing B2B SaaS ad account based on our dataset findings.
| Audit Stage | Primary Failure Point | Diagnostic Action Item | Expected Impact |
|---|---|---|---|
| 1. Attribution Setup | Optimising for Purchases on a 14-day software trial (starves pixel) | Change TOFU campaign objective to 'Complete Registration' or 'Start Trial'. | Drastic CPA Drop |
| 2. Audience Targeting | Using small custom audiences for Lookalikes (e.g. 100 recent purchasers) | Build 1% Lookalikes using higher volume mid-funnel events (all trial signups). | Scaleable Volume |
| 3. Funnel Structure | Assuming trialists will convert automatically after the 7-day window. | Deploy a BOFU campaign retargeting trial users with product features and urgency. | Higher Trial-to-Paid |
Why should you stop fighting the ad platform?
The attribution gap is not going away. Meta is not going to extend its 7-day click window just because your B2B sales cycle is slow. If you continue to optimise for a purchase event that the algorithm cannot see, you will continue to burn cash on expensive, disjointed campaigns.
As the data from these 4,622 registrations proves, the logical move is to lean into the platform's strengths. Feed the pixel the high-volume registration events it needs to stabilise costs. Accept that your ad platform metrics will look like trial acquisition costs, not final customer acquisition costs. And rely on a structured, multi-stage funnel and your own internal CRM to turn those cheap trials into high-value paid subscriptions.
If you're running ads for a SaaS or B2B software and you're stuck hitting a ceiling with plateaued spending and rising costs, it's highly likely your account structure is fighting the algorithm rather than working with it. It's not just about setting up an ad and hoping for the best; it's about engineering a system that converts.
If you're interested in achieving similar results and want a strategy like this implemented for your business, we can definitely help. I'd like to invite you to schedule a free strategy consultation where we can look at your ad account together and figure out the exact path to profitable scale.
Cite this Research & Data
If referencing this research, data, or strategy in your publication, please use the attribution below.
Lukas Holschuh
Founder, Growth & Advertising Consultant
Great campaigns fail without expertise. Lukas and his team provide the missing strategy, optimizing your entire advertising funnel—from ad creatives and copy to landing page design.
Backed by a proven track record across SaaS, eLearning, and eCommerce, they don't just run ads; they engineer systems that convert. A data-driven partnership focused on tangible revenue growth.