While the marketing tactics are interesting enough, what's more revealing is the underlying data methodology and the statistical foundation that makes the approach viable.
The Statistical Edge in Audience Selection
The core insight that drove the campaign wasn't just that Overjet had a clinical product for dentists. Rather, it was the data-driven identification of a genuine statistical arbitrage opportunity at the intersection of three variables:
Geographic diabetes prevalence (sourced from CDC surveillance data)
Practice sophistication (inferred from technology adoption markers)
Periodontal disease under-diagnosis rates (citing 67% miss rate in the general population)
This three-variable approach is particularly noteworthy because most marketing campaigns operate on only one or two dimensions, creating targeting that's either too broad or riddled with false positives.
Why Diabetes Data Creates Marketing Leverage
The relationship between diabetes and periodontal disease provides a classic example of what statisticians call a "proxy variable" for targeting. When we look at the data:
Research indicates diabetic patients are 2-3x more likely to develop periodontal disease
Areas with high diabetes rates (particularly in the Southeast) show disproportionately high periodontal disease prevalence
Technology-forward dental practices in these regions represent a high-probability audience for AI diagnosis tools
What makes this approach particularly effective is that it doesn't just identify where prospects are located, but why they would be receptive to the message. The campaign isn't just "Dentists in Alabama" (a common marketing oversimplification) but rather "Forward-thinking practices in regions where their patient population is statistically more likely to need the specific diagnostic capability this technology provides."
The Uncertainty Challenge and Probabilistic Approach
Like any statistical model, this targeting approach contains inherent uncertainties. The CDC data provides diabetes prevalence at the county level, but individual dental practices will inevitably see variations in their specific patient populations. Technology adoption markers (like equipment financing records) are imperfect predictors of AI receptiveness.
This is where the campaign's multi-variable approach becomes crucial. By combining three partially correlated signals, we increase the signal-to-noise ratio in our targeting. It's similar to how political forecasters might use a combination of polling averages, economic indicators, and historical trends rather than relying on any single predictor.
From Data to Message: The Non-Intuitive Finding
The most counterintuitive element of the campaign involved the messaging. Most marketing wisdom would suggest leading with the revenue opportunity ("find more billable conditions!"). However, our data analysis pointed in a different direction.
The message testing revealed that hyper-specific, data-driven communication performed significantly better:
"CDC data shows that in [City], [X%] of patients have diabetes with two-thirds having undiagnosed periodontal disease. Our FDA-cleared AI detects bone loss as small as 3mm, helping identify approximately [Y] new cases per month in your practice."
This precision (we like to call it a permissionless value prop) messaging works because it demonstrates knowledge of the specific local context the dentist operates in, making the generic promise of "better diagnostics" suddenly concrete and relevant.
The Scale Question
One valid criticism of this approach is scalability. While the campaign clearly works for Overjet's specific use case, does the methodology generalize? The data suggests it does, but with important caveats.
The key elements that make this approach viable are:
A product with clear affinity to a specific, measurable population characteristic (in this case, diabetics)
Public data that allows targeting by that characteristic (CDC diabetes prevalence by county)
Professional audiences concentrated enough to make the targeting efficient (dental practices)
This combination doesn't exist for every product, but occurs more frequently than many marketers realize. Healthcare technologies, financial services, education services, and specialized B2B software often have similar data-driven targeting opportunities hiding in plain sight.
The AI Multiplier Effect
What makes this approach particularly relevant in 2025 is the dramatically reduced cost of implementation. Five years ago, a campaign like this would have required:
Data scientists to analyze and join multiple datasets
Engineers to build custom integrations
Copywriters to personalize thousands of variations
Today, everyday AI tools enable small teams to execute similar campaigns in hours rather than weeks, at a fraction of the cost. This doesn't just make the approach more efficient—it fundamentally changes the ROI calculation, making highly targeted campaigns viable for smaller audiences.
How FIND Methodology Balances Data and Creativity
The Cannonball GTM methodology (Focus, Investigate, Narrate, Deploy) provides a structured approach to balancing data-driven targeting with creative execution:
Focus: Narrowing to the most statistically promising audience segments
Investigate: Identifying data sources that create genuine targeting advantages
Narrate: Crafting messages that leverage specific, localized insights
Deploy: Using programmatic tools to deliver personalized content at scale
What's notable about this framework is how it moves beyond traditional "persona-based" marketing to focus on measurable characteristics that actually predict product-market fit and receptiveness.
For those interested in seeing how this data-centric approach translates into practical campaign execution, the full livestream recording demonstrates each step of the process. You'll see not just the resulting campaign, but the statistical thinking that drove each decision—valuable for marketers who understand that the best campaigns are built on sound data foundations rather than marketing intuition alone.
Share this post