Posted by If you had conversations about media mix modeling or predictive analytics before, those conversations probably didn’t last very long. In pre-pandemic days, the promise of multi-touch attribution (MTA) was considered the Holy Grail of marketing. When done right, it would help marketers understand the value of each user touchpoint across devices, platforms, etc.
MTA was a lofty and worthy concept, but it never really delivered to its full potential even before iOS14. Lots of privacy-focused regulations and cookie deprecation knocked it off its moorings. In its wake, marketers looking for an analytics and efficiency edge are seeing predictive analytics with new eyes — and finding a more viable solution than ever.
In this article, we’ll dig into predictive analytics: what it does, why it works now (and didn’t really work before) and why it just might deliver on its promise well into the foreseeable future of marketing.
What is predictive analytics?
Predictive analytics has been around for a long time (since the 1940s and Alan Turing!). For marketing, it uses statistical models to predict what to expect for Customer Lifetime Value (LTV) by customer segments, how to score leads based on demographics and behavior, where to apply the advertising budget and what the expected outcomes look like.
Years ago, from a marketing perspective, predictive analytics was only truly viable for big-spending, slow-moving channels like TV and radio. Today, machine learning, artificial intelligence and huge advances in accessible data storage and modeling solutions have helped predictive analytics become more agile and functional. Vendors like Pecan.ai and Channel Mix have emerged to provide high-growth-oriented solutions to supplement enterprise offerings from giants like IBM, Microsoft, Neustar, Oracle, and more.
Related: Why Industry Leaders Are Turning Towards Predictive Analytics
Why didn’t predictive analytics take off before now?
It was big before for all sorts of applications beyond marketing (weather forecasting, financial market modeling, etc.) and for enterprise spenders in the marketing world who could afford to spend time and money on huge, slow-moving data models. But as digital marketing took hold, with speed and agility at the forefront of campaign optimization and customer understanding, it got deprioritized in favor of the idea that you could get all the data from user engagements on every channel (think Google, Facebook, LinkedIn, Pinterest, TikTok, Snapchat, programmatic, etc.) and every device, and that you could understand the specific buyer journeys and the related LTV of individual users.
You don’t need to be a CMO to have heard that Apple’s iOS 14 solution and a slew of current and future privacy regulations (GDPR, CCPA) have changed the data picture. Today, what marketers are lamenting as “signal loss” means platforms like Google and Facebook have lost a lot of ability to track user behavior (and conversions), and the cookie-based tracking that allows marketers to understand individual user behavior is facing extinction.
In short, the pipe dream of MTA that promised a better way forward is going up in smoke. Meanwhile, the rise of machine learning and artificial intelligence has lowered the barrier of entry for predictive analytics and made the practice more nimble, which is essential for reacting quickly and effectively to real-time digital media data.
Related: How Predictive Analytics Can Help Your Business See the Future
Why is it a good bet to make?
As an agency founder, I’ve worked with dozens and dozens of SMB-to-enterprise brands that consistently over-invest in Google and Facebook (an industry-wide issue). These brands suffer from the law of diminishing returns that eventually curtails performance (I’d estimate that many brands our agency audits end up underperforming by 30% or more by refusing to diversify their spending). Perhaps just as important, and as iOS14 showed, they leave themselves highly vulnerable to events that disproportionately affect their main advertising channels.
On the other hand, brands that flex into a healthy mix of channels and touchpoints, including emerging platforms like digital out-of-home (DOOH) and connected TV, are setting themselves up to reach their audiences wherever they prefer to engage and at lower engagement costs.
Related: CMOs, Predict Your Wins With AI And Predictive Analytics
As far as customer understanding is concerned, predictive analytics doesn’t use individual behavior or demographic information (which was seen as a shortcoming not so long ago). Its models use big data to assess potential LTV and lead-scoring value for traits and actions at scale, which happens to align well with many digital targeting options.
The last point I’ll make on predictive analytics is a big one: because it doesn’t rely on personal identifiers or cookies, it will be a durable analytics option even as more privacy-focused regulations take shape at state, national, and international levels.
How do you get started with predictive analytics?
Although the barrier to entry is much lower than in years past, brands still need to go beyond simply signing up with a third-party vendor to harness the potential of predictive analytics.
The model’s power comes from understanding existing CRM (customer relationship management) data from platforms like Salesforce, HubSpot, and Adobe. Getting your customer and lead data in shape will enable predictive analytics to understand the attributes of your most valuable customers and help you find more of them at scale.
Predictive analytics relies on internal (or agency, if you don’t have that capability in-house) resources to turn data into insights and visualizations that lead to actions. You can get the most accurate data, but if you don’t have the right talent to interpret it and translate it into the next steps in campaign expansion and optimization, you’ll be paying for wasted potential.
Do your research to understand the state of your internal data, talk to several vendors to see what they offer, and gauge the resources you need to bring on board to bring your data to life in your campaigns. It’s still a decent amount of lift, but getting it right before your competitors do will provide a distinct and durable advantage in your marketing campaigns.