Powering up Personalization’s IQ

by Aseem Chandra

posted on 07-27-2017

With so many experience optimization technologies and personalization levers available today, enterprises are relentlessly focused on creating, refining and perfecting the most relevant customer experiences for each customer touchpoint or interaction. Yet, the scale of this undertaking quickly becomes insurmountable without the right tools and technology required to make it all flow seamlessly. AI technologies such as machine learning and data science hold the key to the most personalized experiences.

A recent HBR survey found that over 30 percent of companies are using AI for marketing and sales to drive competitive differentiation. Marketers leveraging machine learning to optimize customer experiences are increasingly attuned to what a customer wants and needs at the moments that matter.

Adobe already offers customers a best-in-class machine learning and AI framework with Adobe Sensei, which powers intelligent features across all Adobe products, and Adobe Target, our solution for marketers to personalize experiences and drive business impact. As a neutral third party working exclusively in the interest of marketers, Adobe partners with our customers by providing an open and transparent platform to help drive their businesses forward.

More Science = More Success
Today we’re introducing a new strategy to open up our data science capabilities within Adobe Target. The first step will be to offer a “bring your own algorithm” (BYOA) framework to help businesses better compete within an increasingly crowded business landscape.

Think of Adobe Target as a sandbox for data scientists, where they are able to bring in their own proprietary algorithms, leverage their own expertise and methods within a leading marketing platform – an industry first. Not only will they be able to leverage our finely-tuned algorithms to automate their experience optimization, but it will be possible to easily insert their proprietary algorithms into Target to run alongside ours to deliver personalized experiences in real-time and measure performance. Our customers will be able to continuously train and refine algorithmic models directly within Target, making adjustments and changes as needed.

Our BYOA approach will enable marketers to try modifications of current testing approaches or try completely new approaches in areas such as deep learning within multivariate testing. The idea here is to use neural networks to identify variants of the candidates to be tested, test, and then fine-tune the variants for the next round based on the results.

Evolving to a Hyper-Personalized Approach
Leveraging universal algorithms already available within Target is a great starting point for personalization, but the next step to further differentiate a brand is to plug in more specific, proprietary algorithms. Consider the ultra-competitive airline industry which uses models and algorithms to fluctuate pricing based on a variety of factors. While generic pricing models exist, it is those with dedicated data scientists who can refine existing algorithms further, making them more specific to their own businesses and distinct customer segments to increase conversion by a few percentage points. At scale, these increases can truly make a difference to a bottom line.

Our customers have domain-specific knowledge of their own businesses and industries that no marketing platform provider can provide. By injecting their data scientists’ expertise and influence into the process, brands will be able to speak to their customers in a hyper-personalized way.

We Bring the Data, you Bring the Magic
Today, one of the biggest bottlenecks for data scientists utilizing machine learning is the deployment of algorithms in the field. Much of their time is spent navigating the constantly-evolving big data frameworks and menial tasks such as data extraction. By opening up our platform and automating personalization through machine learning, we’ll free up data scientists to do more of what they care about – develop and refine new algorithms, a much more strategic and impactful use of their time.

An open platform also helps eliminate the anxiety of working with a third-party marketing provider, giving data scientists and marketers more control and influence over their experience optimization efforts. Instead of blindly trusting a black box approach, which can undoubtedly be effective in many situations, the flexibility of working with an open platform and a BYOA framework will bring brands more transparency and guidance over the process.

However, the most sophisticated algorithms are useless and machine learning systems cannot adequately improve over time without enough initial raw data to analyze and test against. This “cold start” challenge leads to the inability to take action on data, a barrier to adequately leveraging machine learning to draw reliable inferences and optimize experiences. Fortunately, businesses using Adobe Marketing Cloud will have built-in access to the raw customer data required to make intelligent and automated personalization decisions. Because Adobe’s platform manages over 100 trillion data transactions each year, we have a massive amount of content and assets available to help brands overcome some of the most daunting personalization challenges. Coupled with a brand’s own powerful algorithms, this presents a real differentiator.

Drinking Our Own Champagne
While we’re not only developers of our own products, we’re also devoted users. We know firsthand how powerful machine learning and AI is to a personalization engine. We have developed predictive methods to understand the current state of the customer and predict their transition to other relevant stages, which helps determine how we personalize our next engagement with them.

We’re also leveraging machine learning to better manage customers’ journeys with Adobe’s products. By using algorithmic segmentation and prediction, we can easily evaluate customers’ usage of our own marketing platform and identify who may need more assistance with our products. We can also determine which users and organizations are less active and why, and how to help them extract more value from our products. As a result, we can automate personalized interactions with them to encourage them to try new things within our solutions.

With access to our open platform, our customers will soon be able to leverage similar strategies to harness the power of their own data and take action on it.

Helping Our Customers Remain Competitive
As enterprises increasingly develop their own algorithms internally, imagine the advantages of leveraging these algorithms, coupled with Adobe Target and Adobe Sensei. No other company can offer brands anything comparable.

There’s no better time to up the ante in your personalization game and we’re committed to partnering with our customers in their endeavors. Expect to hear more from us as we provide more details in opening up our platform next year.

Topics: News, Personalization

Products: Target