Transparency in an Age of Machine Learning

Evergage Blog

Ideas and Strategies for Real-Time Personalization
Transparency in an Age of Machine Learning

April 20, 2017 by

In my update on March 22nd, when we introduced Evergage SmartBundle, I mentioned I’d be writing a more detailed blog post on why we believe, when it comes to marrying transparency with machine learning, Evergage stands apart from every other marketing technology vendor. For those of you anxiously waiting, this is that post.

From our customers and prospects to analysts and media, these days it seems like everyone is talking about machine learning. And while it is an undeniably popular topic, there’s a fair amount of uncertainty about it too. Not necessarily what machine learning is – although there are questions about that as well – but rather how it can be effectively harnessed to deliver great customer experiences (i.e., recommendations and more). After all, given that marketers are known to be quite vigilant when planning and executing campaigns, how can they trust machines to determine what’s best for their important customer relationships?

The reality is that no company can deliver on the promise of true 1:1 experiences without leveraging machine learning. What’s more, no human has the capacity to uniquely understand the true interests, preferences and affinities for each and every customer, and immediately apply those insights to deliver a relevant “in the moment” digital experience. Algorithms, complex data stores, computer processing and high-speed networks are all needed to sift through vast amounts information and deliver meaningful and impactful individualized experiences in real time.

Still, marketers are right to be cautious as they move from comfortable and controllable segment-based marketing campaigns – where reviewing each customer’s experience is still manageable – to full-scale machine learning.

At Evergage, we want marketers to be in control of their machine learning-driven initiatives – just like they would be with any other marketing campaign. After all, even though we have lots of experience working with clients, we still cannot know as much about your business as you do. So we aim to give you the tools and the means to optimize your personalization efforts. We believe there has to be a balance – a bridge of sorts – that offers you oversight and control, while still letting machine learning do its thing.

When it comes to providing transparency in an age of machine learning, here’s what companies can expect when they partner with Evergage.

Evergage Collects and Uses the Best Data

We talk about this quite a bit in our marketing, in sales demos and in conversations with analysts, but having access to deep analytics data is crucial for delivering effective, machine learning-driven personalization. By monitoring active time spent, on-page engagement and the context of everything a visitor looks at – in addition to basic clickstream data – and applying correlation analysis, Evergage develops a highly accurate understanding of what each visitor is actually interested in. And as visitors continue to interact with the pages, screens and content within your website or mobile app, their affinities are continually updated and refined, so the accuracy keeps improving. Add in outside CRM, email marketing and other attribute data, and you get a complete picture of each person across your digital channels. If you’re solely relying on referring source and clickstream data for personalization – as other vendors do – your program is already at a significant disadvantage. As the saying goes, garbage in, garbage out.

Before deploying an algorithmic personalization campaign, Evergage customers have access to the best possible customer data to power their machine learning initiatives.

A True “White Box” Solution

Within Evergage, end users are free to build and deploy machine learning strategies – or what we refer to as “recipes” – without vendor involvement. Using an intuitive interface, business professionals can add one or more algorithms to a recipe and then apply filters and boosters to hone the results. Doing so only takes a few minutes, and if something doesn’t look quite right, the recipe can easily be adjusted.

As an aside, recipes can be used to deliver 1:1 product or content recommendations, personalized search, page sorting and even navigation options.

Evergage provides customers with the “keys to the castle” when it comes to building and managing machine learning algorithms.

Take a Sneak Peek

To ensure desired results before deploying a recipe, marketers can preview what machine learning-driven experiences will be displayed for a particular individual. Simply select a known user profile (from within Evergage) and you’ll see what will be presented to the visitor – when on your site, in your mobile app or when opening an email campaign. Additionally, Evergage allows – and highly recommends – companies to run A/B tests on different recommendation recipes.

When deploying a machine learning campaign, you can preview what a visitor or customer will see when visiting your site, using your mobile app or opening an email.

Omnipresent Assistant  

With Evergage, machine learning isn’t just limited to delivering 1:1 customer experiences. Evergage Guardian constantly monitors thousands of business metrics and quickly identifies opportunities and issues that are performing above or below an anticipated patterns. It then surfaces metrics having the most significant impact on business goals. For instance, if a machine learning recommendations campaign is underperforming, your team will immediately know about it so it can take action.

Machine learning can also serve as a safeguard to notify end users of both new opportunities and potential issues.

If you’re looking for industry-leading, machine-learning personalization capabilities that you are able to understand and manage, we invite you to schedule a demo to learn more about Evergage’s user-friendly and transparent solution.

Share This
testingpage2