Case studies / SaaS
Paid media & UGC · B2B SaaSB2B ads underperform because they look like B2B ads.
Smith.ai sells virtual receptionists to law firms. The accepted wisdom says reach that buyer with polished product screenshots and professional credibility. We ran creator video at them instead: and cost per lead fell from $1,500 to $100.
A virtual receptionist for firms that can't afford to miss a call.
Smith.ai virtual receptionist helps law firms and healthcare practices manage incoming calls with AI-powered receptionists and live agents. Calls are answered, screened, and qualified 24/7, helping businesses capture more leads and avoid missed opportunities.
It is a strong solution for professional services businesses, with customer acquisition driven largely by paid advertising. The challenge was finding creative that could turn a professional audience into qualified leads without increasing acquisition costs.
The attorney scrolling Instagram is just a person on Instagram.
B2B marketing assumes a professional buyer needs professional creative. So the feed fills with product screenshots, feature grids, and logo walls: assets that look like advertising, in an environment where anything that looks like advertising gets scrolled past. The buyer's job title doesn't change how their thumb behaves at nine in the evening.
Looks credible. Doesn't stop the scroll.
Product shots, feature callouts, and brand-safe design. It communicates competence to someone already reading, but it announces itself as an ad in the first frame, and it fatigues fast because there are only so many ways to photograph a dashboard.
- Stops the scroll Rarely
- Creative variations available Limited
- Fatigue Fast
- Cost per lead $1,500
Looks like the rest of the feed.
A real person, talking, in the format everything else in the feed uses. It earns the first two seconds before the viewer has classified it as advertising, and because it's people rather than product, the variations are effectively unlimited.
- Stops the scroll Yes
- Creative variations available 50+ produced
- Fatigue Managed by volume
- Cost per lead $100
$1,500 a lead. $5,000 a qualified one.
Smith.ai had a strong product and a real market. What it didn't have was a creative engine: and in paid social, creative is the whole account.
At $5,000 per sales-qualified lead, the acquisition maths only works if every customer is enormous and stays forever. For a subscription service, that's a difficult position to build on.
Four things were causing it.
-
Creative output was the bottleneck, not budget
Meta rewards volume and variety of creative. A small library of static assets means the same few ads run until the audience stops seeing them, and performance decays on a schedule you can predict but not prevent.
-
Static ads fatigue faster than anyone plans for
There are a limited number of ways to present a software product as an image. Once you've made them, you're re-running variations of the same idea and paying more each week for the same result.
-
The funnel wasn't segmented
Cold prospects, engaged visitors, and previous site traffic were being treated similarly. Someone who has never heard of the service and someone who abandoned a signup need different messages, and they were getting the same one.
-
Meta and Google weren't working together
Demand generation and demand capture ran as separate efforts rather than as two halves of one system, so the searches Meta was generating weren't reliably being caught.
A creative factory, then a media system around it.
The order matters. Structure without creative volume just fatigues more efficiently.
Build the funnel structure first
Full account architecture across Meta and Google, with audiences separated into cold acquisition, warm engagement, and retargeting, and segmentation built specifically for the legal and healthcare verticals rather than applied generically.
Each stage of the journey could now be addressed with its own message.
Source and vet creators who fit the vertical
End-to-end influencer sourcing and vetting, selecting people credible to a legal or healthcare audience. This is the step that decides whether creator video works at all: the wrong face makes the ad less believable than a static one.
A roster capable of producing continuously rather than once.
Produce 50+ videos built for paid, not organic
Performance-driven scripting focused on hooks and conversion rather than brand storytelling, produced as 30 to 45 second assets designed for a paid feed. Structured approval and revision kept the volume consistent without diluting the messaging.
Creative stopped being the constraint on spend.
Run UGC against static, honestly
Both formats deployed side by side across Facebook and Instagram lead generation campaigns, with Google Ads capturing the high intent search that Meta activity was generating. Every variation measured on the same terms.
The creator video won clearly, and the response rate jumped.
Scale winners, retire losers, keep producing
Behaviour analytics and performance monitoring driving a continuous loop: strong creative scaled, weak creative replaced, and new assets entering the rotation before the current ones fatigued.
Cost per lead fell to $100 and stayed there.
Fifteen times cheaper, at the same quality.
Cost per lead fell from $1,500 to $100. Cost per sales-qualified lead fell from $5,000 to $350. Across roughly $200,000 in managed media, that's the difference between about 133 leads and about 2,000.
- Cost per lead, before $1,500
- Cost per lead, after $100
- Cost per SQL, before $5,000
- Cost per SQL, after $350
- UGC videos produced 50+
- Leads generated from Meta ~2,000
The obvious objection, answered
Anyone who has run paid acquisition will read a 93% drop in cost per lead and assume the leads got worse. That is usually exactly what happens: you loosen targeting, volume arrives, and the sales team quietly stops trusting the pipeline.
Check it against the qualified number. At $1,500 per lead and $5,000 per SQL, 30% of leads were qualifying. At $100 per lead and $350 per SQL, 29% were qualifying.
The qualification rate didn't move. The leads weren't worse: they were fifteen times cheaper to reach. The entire gain came from acquisition cost, not from loosening the filter.
That's the part worth taking from this. Creative was never a branding question on this account. It was the variable controlling how much Meta charged to put the offer in front of the right person.
Three things worth taking to your own account.
Creative is a media cost, not a brand cost
On Meta, the creative decides what you pay for attention. Treating it as a design line item rather than a performance lever is why so many accounts plateau with budget still available.
B2B is not exempt from how feeds work
Your buyer's job title doesn't change their scrolling behaviour. Creator video works in B2B for the same reason it works everywhere else, and almost nobody runs it because "we're B2B."
Check cost per qualified lead, always
A falling cost per lead means nothing on its own. Track it alongside the qualified number. If the ratio between them holds, the improvement is real. If it widens, you bought volume.
We're now one of their partners.
The engagement ended in the least common way an agency relationship ends: we became a Smith.ai referral partner and now recommend the service to our own clients.
It's a small thing, but it's the endorsement we'd point to first. Running a company's paid media tells you exactly how good the product is and how well the business is run, and we chose to keep selling it after the contract finished.
As part of that partnership, our clients get 25% off their first two months of Smith.ai service with the code DMB25OFF.
What went into it.
Seven of our services, on one account.
How many creatives are running in your account right now?
If the answer is under ten, creative is almost certainly capping your spend rather than your budget. Give us read access to your ad accounts and we'll show you where performance is decaying, what it's costing, and whether the fix is media or creative. Findings are yours to keep either way.
Read-only access · No obligation · You keep your accounts