How AI is Redefining Marketing with Next-Level Data Analytics
6 min read
The Gist
- AI-powered insights. AI is revolutionizing marketing by providing deeper insights into customer behavior through advanced data analytics.
- Personalization at scale. AI enables marketers to create personalized experiences for customers, improving engagement and conversion rates.
- Streamlining decision-making. With AI-driven data analytics, businesses can make smarter, faster decisions that drive better marketing outcomes.
Spotify launched a data-driven marketing campaign in 2016 that mined user data to spotlight highlights and trends in specific markets. The campaign went viral and included many memorable creatives, such as one for the New York City market that read: “Dear person in the Theater District who listened to the Hamilton soundtrack 5,376 times this year. Can you get us tickets?”
Spotify’s campaign demonstrated that data could engage users, create personalized experiences, and drive effective marketing campaigns. What was true in 2016 is still true today. According to Marketing Week, nearly half of marketers have identified data analytics as a priority investment for the remainder of 2024.
What is often overlooked is that generative AI can play an important role in data analytics. Since ChatGPT launched in 2022, companies and especially marketers have looked to generative AI to create engaging content. Recent developments have shown that this technology can also improve decision making through three important capabilities: summarization, autonomous agents and synthetic data generation.
We’ll look at each of these capabilities in turn.
Summarization: Generative AI’s Role in Modern Marketing
Generative AI is adept at summarizing large amounts of information and identifying trends and anomalies that might otherwise be overlooked. This has been especially true in regulated industries where analysts have access to filings and reports but may not have sufficient time to sift through all the information. Nasdaq, for example, developed a generative AI assistant that increased the productivity of its analysts by up to 33% by summarizing stock trades and regulatory filings.
Marketers can benefit from generative AI’s summarization capabilities in similar ways. Large language models can distill user surveys, product reviews, marketing reports, social media posts, and other datasets to detect trends and user preferences. L’Oréal, for example, is using generative AI to drive its marketing efforts. Among its initiatives, it uses large language models to analyze social media comments, images, and videos to identify new product opportunities.
Autonomous Agents: Breaking Down Data Silos
While data is a key asset for decision making, most businesses are plagued by data silos that make information inaccessible. For marketers, data comes from websites and mobile devices, email campaigns, social media platforms, digital advertising networks, call centers and many other systems. Marketers know this information could be useful, but rarely have the time or technical expertise to access, format and run analytics on the data.
Generative AI is changing the situation dramatically. Large language models can power autonomous agents that interpret user requests and integrate with multiple data sources to produce meaningful answers. Many technology companies already offer autonomous agent solutions that can handle multiple tasks, including supporting analytics.
To use an autonomous agent, users first grant the agent access to tools such as databases, APIs or other business software. The AI-powered agent can then break down a user’s request into a series of steps that it maps to specific tools. The agent orchestrates calling the tools to produce a final response that fulfills the user’s request.
For example, suppose a company has a database with email campaign data (campaign names, descriptions, email addresses, open rates, click throughs, etc.), as well as an API to access web analytics. A marketer could ask an autonomous agent: “Give me a chart for the last month showing the effectiveness of our email campaign named X.”
The generative AI-powered agent would analyze this request and recognize that it can use the campaign database and web analytics API to solve the problem. It could then write and run a query against the database to find all the users who had received a campaign email and clicked a link in the message in the last 30 days. The agent could then call the web analytics API to track what those users did on the website, including whether they completed a purchase. It could then generate a graph with all this information using data visualization software.
Generative AI and autonomous agents represent a significant change in how users access information. Large language models can now serve as a bridge that connects users to data sources and analytics tools. One researcher at MIT recently noted that with generative AI, users can “move from just querying data to asking questions of models and data.” This is AI in marketing at its finest.
Related Article: Stay Ahead: Embrace Generative AI in Marketing Now
Synthetic Data: The Power of Market Research
A major problem with business analytics is the lack of good user data. Conducting user surveys, for example, is expensive, time consuming and response rates are generally low. But do you really need to conduct user research? What if you could use generative AI to produce so-called synthetic data?
In April and December 2023, researchers published papers on using generative AI to conduct market research. Specifically, they prompted large language models to answer user survey questions to determine brand perceptions and preferences. In both studies, the results correlated very closely — as high as 90% — with the answers provided by human respondents. One group of researchers even demonstrated that they could prompt the models to provide information for specific demographic groups.
The results of this research were so promising that several marketing firms and startup companies have begun offering market research services that generate synthetic data. In a matter of hours, these companies can produce marketing data for target demographics that would have taken weeks to compile through traditional means. As one Marketing Week writer enthused: “The era of synthetic data is upon us.”
While some expectations may be overblown, there is strong evidence that synthetic data can significantly change market research. Synthetic data is already widely used in fields like biomedical research, insurance and financial services. It also makes sense that generative AI can produce marketing data because the underlying large language models have been trained on datasets created by real humans.
By ingesting product reviews, social media posts, blogs and other writings, the models have learned people’s perceptions of and preferences for specific brands and products. The models can thus use this information to predict how real people would respond to marketing questions.
Of course, people’s opinions shift over time. In order to work well, large language models need be fine-tuned periodically with new information for successful AI in marketing. Rather than conduct expensive user surveys, however, marketers can refresh their models with information from more readily available sources, such as social media and other first party data that captures the behaviors and preferences of users.
Related Article: Top 10 AI Marketing Analytics Tools
Building a Data-Driven Culture for AI Success
To utilize generative AI for data analytics, you need to establish a foundation for success. First, companies need to adopt a data-driven culture, where they are willing to make decisions based on data and invest in the right roles and tools to be successful.
Once resources are available, companies need to develop a coherent data strategy. What questions do you want to answer? What is the business value? Where is the data you need to answer those questions? Who is responsible for managing and collecting that data?
The last step is implementation, which is when you invest not only in generative AI systems, but all the other tools needed to collect, store and distribute data within your organization. During implementation, companies should start small with well-defined goals. They should also plan for rapid iteration. Any project is only the first phase in a continuous process of refinement and improvement.
Generative AI can help you analyze your data and improve your decision making through its ability to summarize, power autonomous agents and create synthetic data. But it is not a magic wand that will instantly solve your problems. It is just a tool that augments your abilities.
For generative AI in marketing to be effective, you need to define the business questions you want to answer and invest in the data foundation that will support the new technology.
Learn how you can join our contributor community.
link
