SalesTech Interview with Dean Abbott, Co-Founder and Chief Data Scientist at SmarterHQ

Know My Company

Tell us about your interaction with smart technologies such as AI and analytics platforms.

I’ve built solutions in analytics and predictive models using a wide variety of algorithms that I would put in general categories of statistics, Machine Learning, neural networks/AI, and pattern recognition. I’ve been doing this for over 30 years in a wide variety of vertical markets for more clients than I can easily count (I’m sure more than 100). Moreover, I’ve taught thousands of students and professionals how to apply these techniques in practical lecture-style and hands-on workshops.

How did you start in this space? What galvanized you to start SmarterHQ?

I started in this space right out of graduate school in the late ’80s, before it was called Data Science, Predictive Analytics, or Data Mining. We called it Statistical Learning or Pattern Recognition. My first job applied statistical learning to optimum guidance and control applied to smart bombs, medium L/D missiles, smart tank projectiles, and fighter aircraft.

SmarterHQ launched in 2010 after I met our other co-founder at a web analytics conference. We believed there would be a tremendous benefit to applying Machine Learning algorithms to marketing decisions in ways no one had done before.

How do you project the offering from SmarterHQ in the overtly crowded cloud-based Data Science and personalization tech landscape?

SmarterHQ’s mission is to make it easy for marketers to maximize revenue and customer relationships by powering highly personalized, cross-channel customer experiences. Marketing personalization and data science have been at our core since the very beginning, versus an ad-hoc feature thrown into our platform over time, which is the case for many other solutions.

In a world where it’s more complex and much harder to collect and activate customer data to successfully meet and exceed evolving consumer expectations, we make it possible for marketers without extensive help from their internal IT/BI/data science teams. SmarterHQ simplifies and elevates their ability to a) activate real-time, multi-channel data to reach more customers; b) identify audiences quickly based on behavior, history, and other key profile data; and c) automate personalized content across email, site, push, and more. The ability to leverage models to immediately and accurately understand behavior and predict future action that will produce the greatest impact is second to none.

Recommended: SalesTech Interview With Adam Mergist, Chief Sales Officer At Clearlink

Tell us more about SmarterHQ’s CDP and how your customers benefit from leveraging it.

The term Customer Data Platform (CDP) is tricky, and I believe a recent Forrester Report says it best: “CDPs are a band-aid solution to a much larger challenge… Data has become bigger, faster, and more complex, and marketers have to activate it on more channels.” CDPs mean well (and any solution that ingests and unifies data could be considered one, including SmarterHQ), but really marketers should be more focused on the advanced capabilities of data activation that many CDPs promise but don’t deliver on.

SmarterHQ provides much more than what a CDP can, allowing clients to leverage advanced identity resolution and cross-channel execution to not only manage and unify all of their customer data, but also activate it in real time. Marketers need to focus on choosing solutions that allow them to make data actionable through automation, personalization, and easy integration with internal systems and vendors in their marketing stack.

Why are CDP and Identity Resolution critical to earning accuracy from marketing data?

Customer Identity Resolution is critical to connecting shoppers across all channels (online and offline), multiple devices, and the various email addresses they use to engage with a brand. Identity resolution unifies all of these behavioral touchpoints into a single customer profile and history, allowing brands to deliver a better, more personalized experience no matter where their customers choose to interact.

If you don’t leverage solutions to actively maintain a single customer profile and pull in online and offline data, you’ll miss out on opportunities to engage top customers, win back disengaging customers, and also send messages that don’t make sense (such as an abandonment email after the customer already bought the item in-store).

How do you differentiate between technologies for AI and Machine Learning at SmarterHQ? Who are you competing within this landscape?

I don’t try to differentiate how technology is labeled as this changes rapidly. AI in the ‘70s, ‘80s and even the ‘90s meant very different things than it does now; in the ‘90s, AI was focused on symbolic learning and expert systems. In the 2010s, AI is focused on deep learning neural networks, applied most often to image processing (facial recognition, self-driving cars, and object recognition), signal processing (IoT streaming data), and Natural Language Processing (such as chatbots).

Many companies claim they do AI and Machine Learning—most, if not all, do. But it’s not just the technology that is applied that matters; it’s the data that undergirds the models that matter. After all, everybody uses the same algorithms: the same decision trees, neural networks, logistic regression, text mining. What differentiates SmarterHQ is the data that feeds the models. Putting this in more Machine Learning terminology, SmarterHQ has expended tremendous time researching features, meaning new, derived variables based on the raw data we ingest.

Read More: SalesTech Interview With Eugene Levin, Chief Strategy And Corporate Development Officer At SEMrush

How do you see the raging trend of including involving AI and Machine learning in a modern CIO/CMO’s stack budget?

There is good, bad, and ugly here. I’m very glad that the C-Suite is aware of Machine Learning and has a favorable view. Perhaps the first wave of positive expectations came from the 2012 Harvard Business Review article, “Data Scientist: The Sexiest Job of the 21st Century.” However, on the bad side, there is also a huge amount of hype that has surfaced, especially around AI and AGI (Artificial Generalized Intelligence).

The ugly part of the process is that AI is rarely the silver bullet in solving problems. Though there is a perception that if we “just buy me some of that AI and we will make lots of money!”— the reality is, AI is just math (complex math to be sure, but still just math) and can’t turn bad data into great solutions. We still need to roll up our sleeves and solve the right problems with the right data.

What are the biggest challenges and opportunities for businesses in leveraging CDP and Behavioral Intelligence technology to optimize customer support and customer success?

First, marketers want to use many different data elements collecting and processed in the CDP, including session/visit information, product interactions, category page interactions, internal site search, purchase funnel interactions, etc. Second, it’s not just one element that matters: customer segments are based on combinations of behavioral elements. Third, it’s not just current behavior that matters, it’s historical data plus current behavior — what a shopper did months ago, last season, last year, and even over their lifetime can be influential in forming the best marketing message for the client.

So, in short, the challenge (and opportunity) is scaling processing to handle huge volumes of historical data in near real time so that the marketing messages are relevant and timely.

How should young technology professionals train themselves to work better with automation and AI-based tools?

What I tell young professionals is the same, regardless of their degree. You need what I call a “Freakonomics” mindset. In that book, data science, predictive analytics, and Machine Learning are not mentioned. However, the mindset for how to solve problems oozes out every chapter:

1. Gain experience in a company building models that get deployed. Good model-building skills take the experience with business stakeholders to define problems to be solved, understand data infrastructure, and work with IT teams. This could take years before one is a good practitioner.
2. Learn from other, more experienced analysts within your company, at meetups, conferences, online. Read, listen, apply—a lot!
3. Degrees are nice, but not the most important part of the process. I contend that more important than statistics, mathematics, or computer science degree is understanding how the algorithms work and how one can tune them. Higher degrees help but aren’t essential.

What is the biggest challenge to digital transformation in 2019? How does SmarterHQ contribute to a successful digital transformation?

The top challenges (and opportunities) for 2019 include:

• Size and complexity of data and consumer expectations continue to rise on the personalization front;
• Inaccurate understanding of consumer intent so the messaging sent is most relevant based on their behavior and history;
• Lack of focus on more than just what the customer is doing right now, but also what they did last week, last month, and last year when trying to interpret what that abandonment means now.

With SmarterHQ, we solve all of these problems, and the breadth and depth of data are immediately actionable, bigger, and broader. We’re collecting granular behavioral, profile (CRM, persona, loyalty, scores), and offline and offline data and pull it all together in an actionable, cohesive, flexible way. Doing this without our platform is incredibly hard to do and it would be very time-consuming/resource-heavy for companies to build on their own.

How potent is Human-Machine Intelligence for businesses and society? Who owns the Machine Learning results?

“Own” and “results” are two key terms in your question. The company that processes the data owns the Machine Learning process and the results. The customer ultimately owns their identity and whether his or her personal data can be used by the company. GDPR provides powerful safeguards for consumers and will likely be expanded to regions outside of the EU.

Where do you see AI/Machine learning and other smart technologies heading beyond 2020?

Machine Learning will provide ever-increasing automated decision-making capabilities in every area of life.

The Crystal Gaze

What Cloud Analytics and SaaS start-ups and labs are you keenly following?

It’s impossible to know what others are doing behind the scenes, so, I stay focused on SmarterHQ and pushing our data science capabilities and innovation forward. I do continue to keep tabs on industry trends and have frequent conversations with colleagues and experts, including our clients.

What technologies within AI/NLP and Cloud Analytics are you interested in?

One in particular: Model deployments and automating the operationalization of Machine Learning.

As a tech leader, what industries do you think would be fastest in adopting analytics and AI/ML with smooth efficiency? What are the new emerging markets for these technology markets?

From my perspective, it really has nothing to do with industries—every industry has examples of companies that are operationalizing analytics and AI/ML well. At SmarterHQ, we are fortunate to work with leading brands in retail, travel and hospitality, and financial services verticals, so that is where I continue to focus my attention.

Also Read: SalesTech Interview With Brian Howe, Alliances Director at Skynamo

What’s your smartest work-related shortcut or productivity hack?

If you do something more than twice, automate it.

Tag the one person in the industry whose answers to these questions you would love to read:

There are many, but since I can only name one at this time, for these particular questions, I’d ask Bernard Marr.

Thank You, Dean. We hope to see you soon on SalesTech Star.