4 ways to use machine learning to improve customer experience (original) (raw)

There are many ways to use machine learning to improve customer experience. Gartner's Bill Delrieu offers four and names specific machine learning-based tools to help with each.

In a digital business environment, providing a quality customer experience -- on multiple digital fronts -- is not only a crucial aspect in modern business strategies, but it's also becoming a key responsibility of the CIO.

AI and machine learning tools have a significant role to play. According to Gartner, customer experience (CX) represents the majority of AI business value through 2020. AI-driven customer experience projects are still nascent, however. A Gartner survey found that 50% of customer experience professionals are using digital analytics or big data in their CRM/CX projects, but only 26% are using AI or machine learning.

How can CIOs best incorporate machine learning to improve customer experience initiatives? At the recent Gartner Catalyst 2018 event, Delrieu, research director at Gartner, detailed how to use machine learning to improve customer experience by adding AI-rich tools to enterprise analytics platforms.

Editor's note: This transcript has been edited for clarity and length.

1. Augmented/predictive analytics

Delrieu: In traditional analytics, you have a theory and you go through all your data and workloads to either validate that theory or completely disprove it. Then you come up with an insight and repeat that process over and over again.

In the augmented or predictive analytics model, machine learning tools can provide insights without your having to repeat that process over and over again. These insights can be really useful in helping you understand what customers are doing and how to best serve up experiences to them.

This concept of using machine learning for segmentation is going to be really useful because it allows people to see segments that are not readily visible to humans going through the data.

Delrieuresearch director, Gartner

An example is the Google Analytics Intelligence feature that automatically generates insights. Google Analytics can go through all the different metrics and landing pages and see which performed better than others and give you practical recommendations right away.

Adobe Analytics also has a feature called Anomaly Detection, in which a machine learning algorithm goes through your data, creates a trend line and gives you an idea of where you should expect those values to be over time. It automatically points out areas that are anomalies for those expected values and provides specific details about that anomaly.

One interesting use case is from a company called Epson America Inc. It recently used this anomaly detection to pinpoint areas that had broken links or broken marketing landing pages and to update its site in real-time to provide a better user experience.

2. Segment discovery

The second use case is segment discovery. This is where you want to use machine learning tools within your customer analytics application to provide insights on who your customers are. This concept of using machine learning for segmentation is going to be really useful because it allows people to see segments that are not readily visible to humans going through the data.

These [machine learning] tools might say, for example, 'Did you know that you have a very high frequency of people who are mobile but using Safari on mobile or using Chrome on mobile?' You might never have realized that was a part of your segments. The beauty of these systems is that once you have these segments, you're able to have more tailored and optimized experiences by providing more relevant content or products to those specific segments.

One example of this is Redtag.ca, a Canadian travel site. What they found, using Adobe Analytics' Segment Comparison, is that their users on mobile were doing a lot more searches, but were not buying nearly as much. So, for the company to convert users, it was really important to better optimize its search function on mobile and then focus on the desktop.

IBM Marketing Insights is another example [of a segment discovery tool]. The tool can go through users who are opening your company's emails and identify those who don't buy much or don't necessarily convert, but who actively read the emails. A suggestion [the IBM tool] may provide is to entice these customers with a special offer that would maybe put them over the edge.

3. Experience personalization

Use case No. 3 is experience personalization -- leveraging machine learning tools to analyze sources of data and to optimize the individualization of customers' digital experience. This use case depends not only on machine learning tools, but also on the customer profile -- we want to know as much we can about the customer. Your end goal here is to actually provide an experience that's personalized for each user based on that profile. All of your analytics data should be able to help you set up a customer profile that you feed into the machine learning algorithms.

How are these enterprise applications doing this? An example is the auto target optimization feature in Adobe Target. There's a well-known mobile telco that has used this feature to test and optimize landing pages. The company used to have a rule-based system in which it showed a different ad or a different version of a landing page depending on whether someone is coming from Google or from Bing. But with the tool automatically providing the best ad or landing page based off the conversion data, the company saw a 13% improvement over time.

4. Customer journey orchestration

The last use case that I want to talk about is customer journey orchestration. The machine learning tools in this can provide the best next action to nudge customers along the journey to conversion -- [whether it is] reaching out to a person by email, sending a text, sending a push notification or giving them a phone call. In some cases, the system can even take the next action as well, without having any human interaction.

Adobe Sensei's next-generation Journey AI services let you know what the next best actions are and give you recommendations on when to perform these actions. One example of this is a record label that was trying to figure out the best times to promote their releases on their [social] channels. What they found through the tool was that 30 days before the release was the best time frame to promote it on Facebook, 15 days before on Twitter and only two days before on Instagram.