Data has never been more critical to marketing success. Companies are striving to employ the latest analytics to drive improvement in all areas of operations.
The good news is that the majority of consumers are willing to provide data, given the right conditions. In fact, a 2018 survey by Acxiom found that 58 percent of Americans are willing to give some data to a company if they know what they’re getting for it. Coupled with the rise of artificial intelligence approaches to analysis, we should expect a steady gain of business insight and savvy, data-driven decisions.
Making the most of data
First off, if you aren’t using AI yet, you should be. According to Demandbase, back in 2016, 80 percent of B2B marketers thought AI would play a significant role in the industry by 2020.
However, not all companies are maximizing an AI approach to data collection and analysis. Only 26 percent of marketers are confident in using AI, and a mere 10 percent of marketers have implemented AI into their strategies at this point.
Yet AI opens a whole new realm of possibilities for marketers. For example, retailers can use AI to analyze weekly SKU performance data, allowing stores to fine-tune their promotions for the best results. And because AI can analyze customers’ behavior in real time, it can prompt automated systems to respond rapidly enough to influence consumer behavior favorably. This has important applications not only in retail, but in banking, telecommunications, entertainment, and many other areas.
Data points to ponder
For these reasons and more, making the most of your data is critical to enhancing overall performance. These three strategies should drive your approach to data and AI integration.
1. Task your team to do the tough thinking.
AI can provide an adaptive and customized user experience with a speed and precision that cannot be achieved by human customer service reps. Particularly because software can connect customer data from a variety of sources — social media and website use being two of the biggest — AI can help place and modify advertising in a way that matches well to particular users and shapes their purchasing decisions.
But that’s actually good news on the personnel front because you’ll need your human marketers to allocate their energy to things like determining long-term strategy and designing customer experience. As Norm Johnston, global CEO and chief digital officer of Mindshare, observes, “The rapid evolution of AI in media will enable our people to focus on innovation and intelligence rather than repetition and reports.”
2. Build data collection into your product.
In the initial design of your product, be it a physical object or an application, you should already be thinking about ways to collect data from the end user. Some websites, for instance, require visitors to sign up and share personal data (such as connecting a LinkedIn or Facebook account) to make full use of their sites.
Remember that data isn’t always just consumer behavioral data. AI machines also collect data like speech patterns, which are critical for making iterative improvements to speech recognition software. CMOs, in turn, use speech recognition products such as Alexa and Siri to direct and improve the customer experience. For example, Hyundai recently teamed up with Google to integrate the Google Home voice assistant into its luxury Genesis brand. Drivers can start and lock their cars and request navigational help, all without pressing any buttons. It’s a customer’s — and a marketer’s — dream.
3. Keep your data squeaky clean.
It’s easy to get carried away and simply throw as much data as one can, as fast as one can, at AI machines. In some cases, marketers are collecting too much of the wrong data; in others, they haven’t collected nearly enough data. But quality is often more important than quantity.
As Sourav Dey, managing director and head of machine learning at Manifold, notes, “Remember the AI uncertainty principle. When data is missing, incomplete or dirty, you won’t get much value from your AI.” Manifold, an AI engineering services firm, was able to capitalize on this approach to help a leading U.S. baby registry become more customer-centric. Ultimately, the firm created customer prediction tools using more than 300 potentially predictive features from its data sets, and these tools allowed the registry to estimate the lifetime value of its customers just a few days after they signed up.
For AI systems to excel, they must be trained on relevant data. To produce the most useful results, humans should sort some of the data to ensure it’s representative of whatever they’re attempting to analyze. Anything extraneous, such as HTML tags or random gibberish, needs to be removed.
Marketers that commission their teams to do the deep thinking, help build the pipelines for incoming data right into their products and services, and make sure to cleanse the data they collect will enjoy significant advantages in the marketplace. Those who start early and optimize these systems continually can look forward to encouraging returns in the years to come.