Manual Data Science and External Dependencies Impact Predictive Analytics in Marketing
Discover the state of predictive analytics in marketing from Pecan’s latest study.
Predicting the future is essential for marketers to understand customers’ interests, behaviors, and trends. With technological advancements, AI-powered predictive analytics are helping marketers in this regard. Pecan recently conducted a study to understand where predictive analytics stands regarding marketing today. Check out the findings here.
The prospect of predicting the future seems enticing. But for businesses, especially the marketing department, it is necessary. Marketers must predict customers’ interests, behaviors, and trends to achieve their goals. With the advent of technology, artificial intelligence (AI)-powered predictive analytics are helping marketers anticipate customers’ behaviors and trends and plan their campaigns with greater accuracy. Overall, the global predictive analytics market is expected to touch $28.1 billion by 2026, according to MarketsandMarkets.
But where does predictive analytics stand today when it comes to marketing? To understand this, Pecan recently conducted a study. The following are a few findings.
AI Ambitions and Obstacles Are Universal in Marketing
The adoption of predictive analytics by marketers is growing. However, the degree to which marketers are implementing the methodology varies widely. Further, the implementation is not always focused on the most impactful applications.
According to the study, the two major uses for predictive analytics were customer-level predictions and future behavior (51%) and customer trend forecasts (50%). While there are other uses, only 40% use it to uncover predictive insights into their business. Interestingly, 49% do not use predictive analytics for individual-level predictions about customer behavior. Hence, they may be missing out on the core differentiating aspect of the technology. This also indicates missed opportunities for them to improve customer and revenue growth.
The study also found that marketing decision-makers have a vision for the AI-powered capabilities they like to gain. Among the chief capabilities are the abilities to predict churn and retention on a customer level (46%), predict lifetime value (40%), and identify upsell and cross-sell likelihood (40%). No marketing leader said they did not want additional AI-driven predictive capabilities.
AI-powered capabilities marketers wish to access
Source: State of Predictive Analytics in Marketing 2022
While everyone has AI ambitions, they also face obstacles when implementing AI-based predictive analytics. There are two obstacles marketers primarily face:
- The high cost of doing manual data science, with 40% citing it as an obstacle
- The challenge of working with unstructured or unorganized customer data (40%).
Besides these, marketers also face other obstacles, such as limited technical knowledge (39%), leadership not being convinced of the value (39%), insufficient or outdated data (35%), and siloed data (31%).
See more: Artificial Intelligence Delivers a More Enlightened Framework for Marketing
The Alignment Between Marketing and Data Analytics Is Evasive
Many organizations depend on technical people outside the marketing teams to implement predictive analytics. This requires the marketing and technical teams to be on the same page for the implementation to succeed. However, the study showed that the teams are not always aligned, making implementation a challenge.
Respondents cited multiple reasons the data projects failed to progress toward the expected marketing goals. The following are some of the reasons:
- Different understandings of business goals and needs: About 40% of respondents said individuals who developed the models didn’t understand the marketing goals. Further, 38% said data scientists don’t ask the right questions about customer behavior. About 37% also received models developed on partial or incorrectly selected data, resulting in flawed predictions.
- Flawed data selection and slow model building: According to the study, data scientists were overwhelmed and couldn’t meet the requests of 42% of respondents. For 35%, the models took too long to develop.
- Neglected models, which are useless and even dangerous: Models must be regularly monitored and updated to be useful for marketers. However, for 38% of respondents, data isn’t updated often enough. Further, 37% said models built for their teams were never updated. Not updating models can lead to inaccurate or, sometimes, dangerous predictions.
Updating data and models can be automated. However, most businesses may not have identified that automation yet, and they likely depend on manually run data science workflows.
The study also found that there is no universal approach to integrating AI into marketing for the time being. About 80% of respondents adopted various approaches to bring AI into marketing teams. This included collaborating with data science teams, using external consultants or vendors, or building data science resources within marketing.
However, these disparate processes may make creating alignment between business needs, timelines, expectations, and resources more challenging. On the other, full ownership of predictive analytics within the marketing team would likely provide far better results.
Marketers Feel Guesswork Still Guides Decisions
About 84% of respondents agreed that despite all the customer data the organization gathers, it is still challenging to make day-to-day data-driven decisions. Further, half of the respondents said their ability to predict customer behavior is usually guesswork or gut feeling.
Respondents also felt that their organizations couldn’t rapidly adjust acquisition and retention programs to the shifts in customer behavior. Only 28% felt they could make this adjustment within a week, while 72% needed more time. Despite constant data collection, marketers seem to struggle to put fresh data to work immediately to guide their programs.
In continuously changing market conditions and customer behaviors, the right timing is of the essence. Acting on the latest behavioral data helps improve customer experience throughout the buyer’s journey and customer acquisition. Yet, business leaders seem to be steering a slow-moving ship through the fog without knowing what the future holds.
Marketers Want Impactful Predictions Specific to KPIs
According to the respondents, the AI application that is most valuable to them is being able to put their data to use in decision-making (62%). Simultaneously, they also want more targeted information. According to 61%, AI should empower them to extract the most impactful analysis from their data. And importantly, it should be specific to key KPIs instead of scouring for potential useful insights (60%). This means that instead of diving down the data rabbit hole, they need targeted and relevant insights quickly that are useful to move the needle on their KPIs.
Marketers ranking the above in terms of what they think is most valuable to them
Source: State of Predictive Analytics in Marketing 2022
The priorities make sense in the context of the metrics businesses use to evaluate the value of marketing analytics tools. Some of the most frequently used metrics to measure the success of their investments are ROI on advertising spend (41%), churn and loyalty KPI improvement (39%), and lift in the acquisition of high-value/profitable customers (36%).
See more: The Problem With Relying on Your IT Department for Data Analytics
Demonstrating Quantifiable ROI Is Essential for Investments in Further Programs
How are marketers currently faring as they try to demonstrate quantifiable ROI for channel or campaign performance? For 83%, it is at least somewhat challenging. Yet, the ability to show the value of these efforts is crucial to assessing the teams’ success and convincing leadership to invest in further programs. Added to this daunting task is the economic downturn, which is compelling many companies to cut marketing budgets.
Fortunately, 63% of respondents expect only a little reduction in spending on marketing technology and data measurement. They will be better equipped to innovate and use data effectively.
The study also found that alternative approaches to data science can offer AI’s full potential. Potential investments for these times include alternative solutions for doing data science more efficiently and using skills already present in the marketing teams. For example, 93% agreed that low- or no-code tools could help automate predictive insights.
Conclusion
It can be seen from the study that marketing teams have huge ambitions regarding predictive analytics. Simultaneously, they also face obstacles, such as a lack of alignment between marketing and data science teams and a lack of buy-in from senior leadership. Further, despite massive data being available, many leaders still tend to go with gut feeling when making predictions. Organizations can overcome a few challenges by bringing full ownership of data analytics within marketing and introducing automation tools that can be used by people without advanced data science knowledge. Ultimately, leveraging predictive analytics to the fullest will help marketers achieve better predictions and revenues.
What challenges are you facing when using predictive analytics in marketing, and how are you overcoming them? Let us know on Facebook, Twitter, and LinkedIn.
Image source: Shutterstock
MORE ON AI IN MARKETING
3 Ways AI Will Change Revenue Technology in 2022
Personalization at Scale: What Synthetic Media’s Rise Means for Brands
AI-Powered MarTech: Moving from Potential to Indispensable?
How to Leverage AI To Win in Today’s B2B Selling Environment
Automation and AI Are Key for Manufacturers’ Go-To-Market Strategy
