Bold Play: Reinventing Marketing Measurement with Predictive Intelligence and Data

Marketers are always looking for a new tool to have in their toolbox. But what if it’s not a case of using the right tool at all?

Henry Ford famously once said that if you asked people what they wanted to get around, they’d answer ‘a faster horse.’ They’d never realize they wanted a car, until one came along.

Similarly, Marketers are working to grow their business by using different tools. But the typical approaches are outdated, they haven't evolved with the buyers’ journey.

The customer journey is more complex than ever. It’s non-linear, laws and expectations around privacy make it near impossible to understand the customer journey with as-is tools, let alone optimize against it. Because measurement techniques and tools haven’t evolved much in a decade, marketers wind up making assumptions in an effort to understand their customers better and design a campaign they believe fits.

The current measurement tools offer an incomplete and inaccurate view of your buyer’s journey due to issues with data quality, consistency, etc. Challenges already exist in getting a read on offline channels, or visibility into closed platforms (aka walled gardens) further complicates things.

Digital campaigns have a match rate anywhere between 29%-62%, per Google Ads or topping out at 63% per Nielsen. 

Layer in the fact the customer is on an increasingly complex journey - 61% of companies say that they are trying to measure more channels than ever before (The CMO Survey p15), and this figure is as high as 77% in the B2C space - and it’s even more challenging to get a view of your buyer journey

It’s an evolving privacy landscape - not just from a regulatory perspective, also from a consumer’s perspective: while many companies are reacting to the impending cookieless world by enriching their first party data, the truth is that 43% of consumers don't feel they can effectively protect their data and point to being forced to accept how a company will use their data in exchange for a service.(Source: Cisco Consumer Privacy Survey).  

Many marketers are using two types of tools for their marketing campaigns. The first and more commonly used tool is Multi-Touch Attribution (MTA), which collects individual, user-level data for addressable (trackable) media and conversion events to determine the impact each media event has on a customers’ path to conversion.  The second, Media Mix Modeling (MMM), uses Collective Correlation (Regression Analysis) to identify and measure the relative sales contribution of Marketing Mix and Market environment factors. 

There are dozens of other tools in a marketer’s tech stack, all variations on a similar path of measurement. Accordingly, they offer various outcomes or measures for the same underlying questions.  But they all rely on the similar uses of the available data for analysis. 

Additionally, these tools are reliant solely on the data that is available to marketers, and offer only a limited or abstracted picture of a customer journey.

Some of the limitations of Mixed Media Modeling (MMM) include:

• The results are very opaque, results are viewed only at the aggregate level, with detail beyond (campaigns, users, creative, messaging) averaged out

• They don't consider that people differ in their exposure to marketing, they differ in their response to marketing, and that their responses to  marketing message, creative and tactics differs

• Message, creative and tactics are often correlated with sales. Which measurements that exclude them are biased

• Decisions made without considering these differences in exposure and response to message, creative and tactics are suboptimal, which is doubly jeopardizing

• Resource and time-intensive process, taking weeks or months to complete, and is usually only done annually

• There is an assumption that variables are static, when in reality people, economy, etc. are in flux

• Attribution models are outdated, with single sources of information

Some of the limitations of Multi-Touch Attribution (MTA) include:

• A reliance on user-level data tracking (including cookies, which are being sunsetted), which is highly incomplete 

• An omission of relevant factors that are also correlated with sales biases the measurements severely 

• Most MTA methods also ignore unobserved factors leading to a sale.  They can't be readily used for actionable insights for the journey to conversion for most marketed products and services

• Every customer is treated the same, they fail to capture differences in impact across different customer types

• First and last touch are measured, but there is a time delay

• MTA requires tracking and connecting all media at the device, or hashed identifier level, it does not account for non-addressable media (print, radio, and traditional/linear TV), which cannot be traced to individuals

We are living in a post-cookie world and companies are using more channels than ever before to connect with their customers. Marketers need a system that truly puts customers—not their devices—at the center of measurement. 

Additionally, the weakness of the current tools in a marketer’s tool box is that they don’t accurately predict what the customer will do next - they offer partial insights into what has happened in the past.

Artificial intelligence provides marketers an opportunity to throw out their box of outdated tools, and think differently about the entire system of measurement. It is an opportunity to change how we use data to plan, deploy and optimize campaigns. 

An Opportunity to Test, Learn and Earn by Doing and Validating -  Just Like Human Intelligence

Whether we realize it or not, we are living in the Between Times, a period of time between the introduction of a new technology, and the mass adoption of it, as described by Avi Goldfarb, Ajay Agrawal, and Joshua Gans in their book Power and Prediction: The Disruptive Economics of Artificial Intelligence

As a predictive machine, artificial Intelligence offers a completely new way for marketers to plan their campaigns. 

AI is not a new tool for marketers, but a new way of how marketers work. It is a completely new system. AI will revolutionize the process marketers use, and the value they can bring to their businesses, propelling its growth. 

Artificial Intelligence evolves an opportunity for the marketer to decouple prediction and judgment. Traditionally, prediction and judgment have coexisted, and marketers have used tools to get better at understanding their customer. We've used our judgment to inform our predictions instead of using our judgment to think about how to act on predictions generated by understanding human experience.

By using AI for its predictive intelligence, marketers can sharpen the resolution to create more robust, complete views of customer journeys. Artificial intelligence can work back from an outcome (ie. a purchase), and analyze huge swaths of data - augmenting when needed. It uses prediction to discern patterns of behavior.

The more exposure and conversion data a marketer provides the artificial intelligence, and the better human judgment is encoded into the AI, the better a predictor it becomes at making recommendations that drive desired outcomes. Artificial Intelligence can create a cheaper prediction model, allowing for the constant updating, and can constantly model and shift with the market. It can ingest data and different levels of granularity (eg. user vs. aggregate) in real time, which allows it to overcome the limitations of traditional tools used by marketing analytics.

AI is the prediction machine that allows a marketer to build a campaign based on what customer behavior will be, by reviewing what it learned combined with updating and encoding the judgment framework that makes it more powerful and dynamic way than the aforementioned tools.  

Using AI to Drive ROMI

How One Company's Strategic Re-Deployment of Marketing Communications Led to Business Success

This Direct-to-Consumer (DTC) healthcare provider had traditionally dedicated a substantial portion of its marketing budget towards Facebook. However, the return on marketing investment (ROMI) had been less than satisfactory. Despite numerous attempts to optimize its advertising on Facebook, the company's Customer Acquisition Cost (CAC) remained persistently high.

The Company turned to Plus AIOS, a marketing measurement and optimization solution fuelled by artificial intelligence that enables businesses to leverage big and wide arrays of data and predictive analytics to guide their media allocation decisions.

With the help of Plus AIOS, The DTC Healthcare company began analyzing its marketing strategies and campaign effectiveness by collecting a wide range of data, including audience demographics, ad impressions and information about them like message, creative, tactics, timing, as well as customer interactions with owned assets, ie. website visitation, and conversion rates from all of the company’s  advertising platforms.

One particularly insightful piece of intelligence emerged. The data showed that the Company's Facebook campaign was less effective in comparison to YouTube. Not only were engagement rates lower on Facebook, but the cost of acquiring a new customer was also higher - where the Company also had a higher engagement rate.

Predictive intelligence led to campaign decisions.

Armed with this actionable insight, the Company made a strategic decision in the last half of 2022 to reallocate a significant portion of its marketing spend from Facebook to YouTube, while it sharpened its creative campaign strategy on Facebook.

The shift was executed iteratively,  carefully, ensuring a smooth transition. A powerful aspect of AIOS is that customers can also see in right-time and validate the results of their slow reallocation in the market which helps fortify the original decision to do so. Using Plus AIOS, the Company began by slowly dialing down its advertising budget on Facebook while simultaneously increasing spending on YouTube. 

The company leveraged YouTube's various ad formats, optimized for mobile users (who make up the majority of YouTube's user base), and created high-quality video content relevant to their target audience.

The Results

The strategic shift paid off. Within a short period, The DTC Healthcare Company saw an increase in conversions, followed by a decrease in their CAC. Engagement rates improved significantly, and they noticed a considerable increase in the rate of conversions from YouTube ads.

Moreover, The Company’'s brand recognition improved as a result of its more prominent presence on YouTube. This led to organic traffic growth on its website, further reducing their reliance on paid advertising and lowering the overall CAC.

Lessons learned

1. Understand analytics to tell the story you want to tell: 

    Decisions backed by data often lead to improved business results. By leveraging Plus AIOS, which decouples predictive analysis from judgment, this DTC Health Care Company gained insights that led to a more effective marketing strategy.

2. Be adaptable: 

    Despite having invested significantly in Facebook, The Company was willing to shift its strategy based on data insights. Their adaptability was crucial to their success.

3. Test and learn:

     The strategic reallocation of funds allowed The Company to experiment. Plus AIOS enabled the company to test synthetically with different marketing platforms. The successful shift to YouTube demonstrates the importance of a 'test and learn' approach to marketing investments.

So, How Does Plus AIOS work?

Plus AIOS is the intelligent All-in-One System for Marketing. Built with predictive intelligence, AIOS connects all touchpoints, and continuously measures creative performance to grow your business.  

AIOS unlocks what's possible. It doesn’t start by looking at channels to project conversions. Instead, it looks at the end goal or outcome (for example, a customer making a purchase), examines all the customers who bought that product, then reverse-engineer their shopping journeys. It's like figuring out all the steps a shopper took before finally making a purchase, from seeing an ad, to read online reviews.

You don't need to provide AIOS with years of detailed sales results or every data point about customer behavior. Just like an observant salesperson notices common interests among shoppers, the predictive intelligence of AIOS begins to pick out patterns from the information you have, and fills in the blanks where data may be missing. The more data you provide, the more adept AIOS becomes at predicting what will lead to the outcome you’re looking for (such as a sale).

AIOS presents a dynamic view of the market.  Much like the salesperson who knows what influences shoppers’ buying decisions, it factors in seasonality and economic conditions in the recommendations it makes. It's also smart enough to understand that repeated exposure to a campaign might increase certain customers’ propensity to buy. By continuously adapting to these changing factors, AIOS guides marketing and sales efforts to connect with customers in the most effective way. 

Recommendations:

Use Artificial Intelligence as a Prediction Machine:    Decouple prediction and judgment as you are using them now in Marketing campaigns. Allow artificial intelligence to analyze huge amounts of data to see patterns and provide insight, which, with human judgment, can create winning campaign strategies.

Put the Customer First:    Be relentlessly focused on the customer. Rather than map the success of your business platforms, start from where the customer made a purchase and use predictive analytics to understand better how they got there. 

Test and learn in Real Time:   Understand that the modern customer’s journey is different from models created years ago that track the success of their campaigns only once or twice a year.  Implementing an all-in-one measurement system based on predictive analysis allows marketers to test and learn, gaining a more fulsome understanding of their customers’ journeys in real time.