Tag Archives: machine learning

In The News: Ad Age Data Design & Alexa

I was recently interviewed by Ad Age discussing the efforts of my data design team and our work with Amazon and the Alexa Skills Kit.

screenshot-2017-02-11-10-46-11

When I first joined the Epsilon agency team I wanted to bridge traditional brand planning, strategy and data science to uniquely assess all of our data sources and build recommendations that leverage the right data to assist planning, strategy development and data-driven insights to support strategy and creative.

Now the agency data design group is comprised of 3 core components: 1) Mapping the data landscape 2) Storytelling through data 3) Consulting & training. My goal with this team is align intelligence from the data, regardless of source, that will inform how we communicate and message with consumers as technology and behaviors evolve and most importantly drive performance.

There are three primary areas of focus for the team:

1) Proprietary data sources & methodologies e.g. Leveraging Epsilon’s structured data

2) Unstructured data sources & methodologies e.g. Finding previously invisible insights by applying machine learning & artificial intelligence to unstructured category data

3) New data sources & methodologies e.g. Uncover new types of data sets that we call affective datasets and how it will impact and reshape how we connect across the consumer journey

screenshot-2016-12-02-14-47-49

Unstructured and New Data sources combined with Epsilon’s proprietary data began to accelerate our processing and analysis capabilities to uncover consumer truths with unstructured data to further fuel our agency’s strategic storytelling and data driven creative leading to an evolution of brand planning.

For the past 12 months my data design team has focused on aligning emerging artificial intelligence systems and algorithms with our structured data assets to combine all of the following elements.

screenshot-2016-12-02-14-52-09

Data Design is the bridge between planning and bleeding edge tools like cognitive computing, artificial intelligence and natural language processing. Ad Age highlighted our approach with Amazon and how we leverage machine learning on amazon.com down to the product SKU level to further inform communication and engagement strategy as well as our team being one of the early adopters of the Alexa Skills Kit (ASK).

screenshot-2017-02-11-10-46-27

Here is an example of data design concepts in action.

screenshot-2017-02-11-10-46-47

Follow Tom Edwards @BlackFin360

Tom Edwards Innovation To Reality

In The News: iMedia 7 Ways AI Enhances Marketing Cover Story

This morning my new article 7 ways artificial intelligence will enhance marketing was the cover story for iMedia Connection.

The article reviews seven subsets of artificial intelligence from machine learning, cognitive computing, natural language processing, deep learning, predictive API’s, object recognition and dynamic content generation and how brand marketers can better uncover insights, connect with consumers, and redefine customer experiences using this innovative technology.

screenshot-2017-01-17-16-00-30

Follow Tom Edwards @BlackFin360

Trends To Watch in 2017

Technology is now essential to our daily lives. Accessibility and empowerment has transformed how we connect and communicate. This has led to new forms of user interaction that will usher in the business models of the future.

2017 will be comprised of new types of conversational experiences to connect with consumers. It will see the continued evolution of artificial intelligence and connected systems as well as the rapid rise of third-party ecosystems supporting virtual, augmented and mixed reality.

The following trend deck outlines the evolution of marketing in 2017 through the consumer centric filters of connection, cognition and immersion and is now available for download.

Screenshot 2016-12-02 15.00.44.png

  • CONNECTION – Trends that reimagine how we connect, enable and empower consumers.
    • Examples include: Simplified Conversational Experiences, Pervasive Voice-Based Interfaces, Search and Retrieval to 1:1 Prediction, Affective Datasets and eSports

screenshot-2016-12-02-14-47-49

  • COGNITION – Trends where machine based intelligence will disrupt and redefine data assets and how we work.
    • Examples include: Machine Learning as a Service, Centaur Intelligence, Blockchain & AI

Screenshot 2016-12-02 14.52.09.png

  • IMMERSION – Trends that align technology and presence to evoke emotion, entertain and power commerce.
    • Examples include: Democratization of VR, VR Commerce, Social VR, (Re)Mixed Reality

Screenshot 2016-12-02 14.57.24.png

  • ZONE OF CONVERGENCE – Trends that align elements of connection, cognition and immersion that will redefine consumer engagement.
    • Examples include: Cars as the next Mobile Platform, Holographic Computing, Ambient Computing.

screenshot-2016-12-02-14-55-53

How we consume and interact via digital channels is about to be absorbed and redefined. We believe that 2017 will begin the convergence of connection, cognition and immersion toward an ambient computing future built on new data types that will simplify complex tasks and predict need states vs reacting.

Download the 2017 Trend Predictions Today!

epsilon-top-trends-for-2017

Follow Tom Edwards @BlackFin360

Tom Edwards Innovation To Reality

IIeX 2016 Machine Learning + AI = Data Driven Creative

Yesterday I had the privilege to speak at the Insight Innovation eXchange or IIeX North America 2016 discussing our approach to getting data driven creative via Machine Learning and artificial intelligence.

Photo Jun 14, 5 14 25 PM

This presentation was a joint effort between my Epsilon team and one of our strategic partners Oculus360. I work closely with Raju Kattumenu, the founder of Oculus360 and we have engaged on numerous initiatives over the past year.

Screenshot 2016-06-15 07.14.21

This has allowed us to leverage the power and reach across public domains leveraging their technology combined with our proprietary data assets to validate consumer truths or find new connections based on occasions, attributes, perception & demand signals.

Screenshot 2016-06-15 07.07.58

The technology is a combination of natural language processing, an artificial intelligence neural network and machine learning systems that combine to unlock various themes & trends associated with demand signals created by consumers.

Screenshot 2016-06-15 07.12.22

The key to this approach is that instead of starting macro across all facets of the web or social conversation, this approach looks at specific domains and can go incredibly deep down to the product sku level.

We then take the results from this approach and combine it with our connection planning process and data assets to unlock consumer truths that will define our approach to creative and strategic territories. It truly is an approach where data fuels creativity.

Screenshot 2016-06-15 07.15.48

This approach of combining machine learning + AI with Epsilon’s data assets allows us to truly identify contextual moments to create personalized experiences.

Context is key as this informs whether we should use storytelling vs. storymaking moments. We then align moments with personalized elements of the story based on our data findings and use cross device identity to create personalized story delivery at scale.

Screenshot 2016-06-15 07.15.59

I also discussed how we have realigned our approach to planning with data science to inform creative territories & strategic themes as well as how this approach supports innovation initiatives by informing and validating consumer readiness when it comes to emerging storytelling mediums.

Screenshot 2016-06-15 07.16.09

We then showed an example based on the mini-van category. Traditionally mini-van advertising has stayed very close to the “family” approach to connecting.

Screenshot 2016-06-15 07.36.58

We wanted to either validate the approach or find new consumer truths based on all the factors outlined above. What we found was very interesting as key attributes and occasions began to surface that outlined new demand signals that could be used to shift perception.

Screenshot 2016-06-15 07.18.51

This also allows us to take a look across brands and see which brands align with specific occasions which can lead to differentiation among competitors.

Screenshot 2016-06-15 07.19.05

We can also look at specific features associated with each of the brands to identify new territories or areas to focus on driving awareness, engagement or advocacy.

Screenshot 2016-06-15 07.42.03

In order to maximize contextual connections with consumers it is important to not only have qualitative data tied to consumer insights. It is also critical to leverage the power of machine learning and artificial intelligence combined with strong data assets to unlock demand signals that can fuel the creative process.

Follow Tom Edwards @BlackFin360

Follow Raju Kattumenu @RajuKattumenu

Tom Edwards BlackFin360

MMS Upfront @ Internet Week 2016

This week I had the opportunity to speak at the Modern Marketing Summit Upfront @ Internet Week 2016 discussing the topic of the transformation of storytelling. Below is a recap of my key talking points.

Screenshot 2016-05-17 10.03.52

The discussion covered 4 territories tied to how storytelling is being transformed.

1) How do you define storytelling from your company’s perspective? What is a story? 

I’d define Epsilon’s approach to storytelling as finding a compelling difference through data that leads to a consumer truth then completing a narrative around that truth. The truth/data will often reveal the correct medium by telling who, what, where and when of the audience. We then use traditional tools of persuasion formatted by channel to reach the audience.

For us a story can be any type of format that creates a connection with a consumer. This can come through brand created, co-created or user created content. We further delineate storytelling and storymaking by working with our brand partners to make them the catalyst for the stories consumers are making for themselves.

2) How does the context of where and when the story is being told affect the way you choose to tell it? 

Context is key as this informs whether we should use storytelling vs. storymaking moments. We then align moments with personalized elements of the story based on our data findings and use cross device identity to create personalized story delivery at scale.

We partner with a major sports speciality retailer to generate and optimize 1 million versions of the brand story that aligns with key contextual moments.

This is approach is based on transactional and online click stream data and that in turn continues to drive actionable insights across all of our initiatives to inform and optimize our creative process in near real time.

The key is consistency of message across various formats and having the ideal understanding of cross-device behaviors to deliver a message at the right time.

MMS Upfront

3) How do you coach your clients to think from the consumers perspective? What do you hope a good story will achieve for your client? 

We show them what consumers expect, how they perceive their brand and category and align strategy where there are opportunities to create new points of connection. This includes understanding when to leverage branded content vs. co-created vs. integrated vs. user created content to tell the story for the brand.

We focus a lot on consumer behaviors. This comes in the form of machine learning and artificial intelligence that looks at specific domains and mobile ethnography studies. We also leverage our proprietary assets that highlight key behavioral, transactional and affinity based data that allows us to demonstrate how we find unique ways to tell or make a story.

We not only show them what consumers expect, how they perceive their brand and category we focus on the potential outcomes tied to our storytelling and story making efforts.

Our goal is to start or change a conversation, create advocacy and ultimately drive purchase and business outcomes.

4) How important is data in informing your decisions about your storytelling methods? 

Data is the fuel of our creative process. It enhances the creative, it does not replace or stifle creativity.

We realigned our planning with data science to inform creative territories & strategic themes we then use this to map the story as well as deliver audiences all with an eye towards outcomes and building models that show the impact of our storytelling efforts

One of the other great assets is our ability to map to individuals across devices. This makes it easier to deliver highly personalized and dynamic creative. Mapping high level themes through to relevant micro-moments. This allows us to connect with consumers regardless of where they are in a heavily fragmented media landscape.

Data also informs consumer readiness when it comes to emerging storytelling mediums. Whether it’s the shift towards conversational user experiences to immersive experiences such as Virtual, Augmented and mixed reality. data is a foundational element to our approach to creativity and innovation.

2016 Header

Follow Tom Edwards @BlackFin360