To boost customer retention and enhance the online experience, machine learning is a must-have in your lean manufacturing tool set.
By Shamir Duverseau, Managing Director/ Smart Panda Labs
Machine learning permeates every industry and if you aren’t using it, you may feel as though you aren’t keeping up. In fact, it may seem like an insurmountable challenge to keep pace with the latest technologies when you’re focused on meeting strategic business goals, improving profit margins, and driving cash flow. This is true of digital marketing – in particular – how customers experience your website and campaigns and how those programs continually speak to your customers –makes the difference for industrial companies large and small.
How do you leverage machine learning in your digital marketing repertoire, not just as a should-have but rather as a must-have to drive quantifiable growth?
Demystifying Machine Learning
First, let’s start by demystifying this computer science calls machine learning. It’s not all that new and is based on a computational approach, called neural networks, which is modeled on the human brain. It was first described in 1943 by Warren McCulloch and Walter Pitts on how neurons might work. Neural networks are a specific set of machine learning algorithms that generate rules based on inferences from data. With an estimated 1.7MB of data created every second for every person on earth, machine learning is taking off.
These computer algorithms are managing huge amounts of data based on a certain set of instructions and rules, which then fuel desired outcomes. Self-driving cars, facial recognition or Amazon’s Top Picks for You are great examples of how these rules play out in daily life. The Amazon example, however, is more aligned with the machine learning model best suited for marketing as its studies data to identify patterns. The same applies to manufacturing and industrial company websites – your customers need to know that you care. And being responsive and recognizing consumers when they return to your site – is key.
Machine learning is a self-learning process for computers, which doesn’t require computer programming. At its core, it’s a tool, not unlike a screwdriver or a hammer, which relies on the human element to be effective. And, it’s a human at the other end influenced and affected by the outputs of machine learning across the customer journey.
Machine learning is as good as the data it’s given. And, using machine learning to improve your digital marketing requires a diagnostic approach to fully flesh out the business challenges you want to solve and then design specific solutions to meet those challenges. Only then, you will have clearly defined your goals and a solution design that can effectively leverage machine learning in your digital strategy.
Laying the Groundwork
To implement machine learning as part of your digital marketing strategy requires you to:
- Articulate organizational vision, laying out business goals, and defining a business strategy for meeting those goals.
- Determine the digital strategy that rolls up to the overall business strategy.
- Detail the tactics needed to execute the digital strategy and the metrics you will use to measure tactic effectiveness.
- Ensure metrics map back to higher-level business goals you laid out to start.
For example, a manufacturing company looking to be acquired may have business goals tied directly to increasing the value of the company for prospective buyers. These business drivers are the building blocks to a digital strategy for improving cash flow in the shortest amount of time with promotional tactics aimed at online customers. A version of these tactics could also be applied to help move prospective buyers more quickly through their customer journey.
Next, examine your tactical challenges by looking at those that are: a) not performing the way you need them to move you forward; b) performing well but need to be iterated upon to adapt to the changing customer, market, and competitive landscape. In the manufacturing company scenario, there is a business need to accelerate the movement of prospective customers through the site’s eCommerce experience, as well as a to adapt to specific customer demographics. Identifying these challenges fuels machine learning, and the key here is learning. You want to apply lean process thinking, a methodology built upon building, measuring, and learning; then using that learning to make decisions on what comes next.
Tools such as Google Analytics and Optimizely may be used to gather customer data. Data is then iteratively used to personalize various touchpoints in the customer journey. Machine learning can be used to accelerate this process and immediately identifying buying patterns in real-time, helping you to quickly get to the endgame of meeting business goals and moving closer to your vision. Ironically enough, this machine learning approach goes a long way in personalizing, humanizing, and optimizing your customer’s online journey. More and more, your website not only serves to pave the way for customers to your doorstep, but it’s also a critical tool to maintaining the ongoing customer relationship. And that can be automated.
While machine learning can be used to shorten the iteration process in gaining measurable results, experimentation can be used to increase the percentage of prospective customers that move through the manufacturing company’s online customer experience. Experimentation can show people multiple variations of desired products and measuring those that generated the most interest. Insights are gathered on the variations and based on those insights the company tailors the customer experience in real-time to fuel higher revenue.
There are several technologies using machine learning from experimentation to recommendations to marketing automation and beyond. New players have entered the market while existing players are incorporating elements of machine learning into their product offerings. As for the data source, it comes from the brand’s data on their customers’:
- Personal Information: name, age, zip code, IP address, etc.
- Engagement: site visits, most viewed pages, email opens, etc.
- Behaviors: purchase details, average order value, cart data, etc.
- Attitude: customer satisfaction rates, preferences, purchase criteria, etc.
How the data is stored so that it can be leveraged to apply machine learning is a critical element to consider as well. The challenge here is many companies are not gathering data in any way that is useful. Using data is easy given the technology landscape. Having data in a supply and format that is useful is the challenge. Marketing organizations must have a strategy for long-term data storage and retention such as the public cloud and that strategy must address high performance computing platforms for processing the tsunami of data that can easily overwhelm typical computing resources.
The longer you wait to harness the power of machine learning, the harder it will become, and the longer it will take to organize and structure the data. And, the longer and harder it is, the greater the challenge of even beginning to use machine learning to determine its value for your brand. While machine learning, integral to your IoT projects, is a science, the art of applying an iterative learning process drives successful digital marketing strategies. And for today’s customers, being recognized can make all the difference between staying on your website or clicking to a competitor’s that is more responsive.
About the Author
Shamir Duverseau is co-founder and managing director of Smart Panda Labs, a digital consultancy that uses data and creative intelligence to drive customer lifetime value. For more information, email firstname.lastname@example.org or go to http://www.smartpandalabs.com.