Take your budget further.
Speak with us today
How we’re different from other MMM solutions
Marketing mix modeling was first developed in the 60s—and most companies in MarTech are using the same old models. We thought it was time for a higher standard.
Don’t call it a “black box.”
Here’s the thing about a “black box”—it doesn’t mean what our competitors want you to think it means. A model is a “black box” if the people using it can’t see what’s happening under the hood and don’t understand how it gets to its outputs. MMMs aren’t a black box because how they work has been public for decades. The problem? All of that documentation is technical—we mean really technical. It’s hard to understand without training as a data scientist.
Our models? Glass boxes, as our CTO calls them. We know exactly how they work and exactly what’s happening. And, yes, we can tell you a bit about it—but not everything. The recipe for our secret sauce is, well, a secret.
| Prescient AI | Recast | Northbeam | Rockerbox | |
|---|---|---|---|---|
| Onboarding | 10–20 minutes | 4—6 weeks | Days to weeks | 4—6 weeks |
| Speed to insights | As little as 48 hours (!) | 5—7 weeks | Days to weeks | 5—7 weeks |
| Model applied | MMM—daily | MMM—weekly | MTA—daily MMM—daily |
MTA—daily MMM—daily |
| MMM type | Proprietary; no research paper available | 60s research paper (*) | 60s research paper (*) | 60s research paper (*) |
| Depth of MMM reporting | Channel & campaign | Channel & tactic | Channel & tactic | Channel & tactic |
| Pixel dependency | None | None | Pixels & APIs | Pixels & APIs |
(!) Some brands have years of data that can take longer to sync before modeling can occur. Custom connectors excluded. ( * ) Borden, N. H. (1964). The Concept of the Marketing Mix, J. of Advertising Research, 2, 7-12.
So you’ve heard daily insights aren’t possible…
That was true—when using models from the 60s. But this isn’t 1960 anymore. We’ve done our best to launch the MMM into 2024, and beyond. We’re constantly thinking about what’s next and building the tools marketers will need down the road.
So why are our competitors telling you that daily modeling isn’t possible? To keep this from turning into a lecture, we’ll keep it short. With older models, the more you zoom in, the more the models had to generalize what was happening, leading to data that was unhelpful at best and inaccurate at worst. So we evolved the models and layered them with advanced machine learning techniques—how exactly we do that is proprietary, but we’re really proud of the advancement. The result? Specificity that wasn’t possible before.
But there’s another factor at play here. Yes, our system is sophisticated enough to produce this level of specificity. But this also comes down to the cost of compute. We’re willing to make a larger investment here to empower our users to have the most up-to-date data. Our competitors may not be.
We’re “future-proof” because we’re future-focused.
Prescient isn’t just a name to us. It’s our mission, and we’re obsessed with living up to it. We saw the writing on the wall back in 2019. Cookies and pixels would become unreliable—we called it two years before Google announced it was depreciating cookies. Marketers needed something better and a way to marry targeted advertising with data privacy. We know you see the future, too, and we’re committed to building exactly what you need to enjoy seamless attribution and optimization of your paid campaigns.
Let’s address the elephant in the room…
That big question is about capital “A” Accuracy. How do you know that what our dashboard says is accurate? First things first. Someone needs to say it, and it might as well be us: There is no single source of truth in attribution technology. Not in marketing attribution. Not in forecasting. Not us, not our competition. Not the channel’s own reporting. That doesn’t mean MarTech or the SaaS platforms we offer aren’t truthful. It doesn’t mean you can’t get valuable insights from them. But it does mean that anyone promising to deliver “The Truth” about your marketing spend and its effects is, well, stretching the truth. Attribution is uniquely difficult not just because the math and the data make it hard, but also because it’s nearly impossible for a human to answer the very simple question “why did I buy this?”
To cut to the heart of the matter: SaaS solutions like Prescient can offer eCommerce brands a tool that helps them triangulate what happened and what might happen if you change your spend on paid campaigns. Yes, there are insights that help your brand become more profitable—but they’re directional and predictive, not prescriptive. And that’s a great thing. A prescriptive solution would eliminate the need for a marketer. Our predictive analytics let you pair your honed instincts as a marketer with data in a way that can amplify your results.
So why do we feel confident about the numbers you see in our dashboard? In a word, backtesting. When we test how our models perform on sample sets of your own brand’s historical data, they’re over 90% accurate. We wouldn’t settle for anything less.
FAQs
No, we get all the allowable data from every platform you connect. We collect at least one (1) year—more on that in the next question—but more if you have it. We know our onboarding time is considerably shorter than what you see elsewhere—and that’s on purpose. We built more robust and flexible data pipelines to ensure you could get all of your data into our dashboard and start seeing actionable insights fast. Marketing moves quickly, and we knew marketers needed a platform that could keep up and provide value ASAP.
Prescient uses at least one (1) year of data to ensure we capture the impacts of seasonality, holidays, promotions, product consideration periods, and other organic activities that may impact the credit that should be allocated to paid marketing activities—which in aggregate we define as “trend.” These trends have a considerable impact on your paid performance, and we wouldn’t be doing our jobs if we didn’t account for that