Adobe Target - Automating Hot Tricks of Data Science

Summary

---

The Digital Marketing Podcast

Episode Title: Adobe Target – Automating Hot Tricks of Data Science
Hosts: Kieran Rogers, Daniel Rowles
Guest: Kevin Lindsey, Director of Product Marketing, Adobe Target
Date: September 11, 2016

---

Episode Overview

In this episode, Kieran Rogers interviews Kevin Lindsey about how data science is transforming digital marketing, specifically through Adobe Target's automation capabilities. The discussion dives into the difference between analytics and data science, the automation of personalized customer journeys, and how marketers can leverage machine learning without dedicated data scientists. The central theme is demystifying data science’s “sexiness” and making its benefits tangible and accessible for all marketers, regardless of their technical expertise.

---

Key Discussion Points & Insights

The Data Deluge and Data Science’s Role

• Data is Abundant, Actionability is Scarce:
  Kevin notes the central challenge is not the lack of data but making it actionable for marketers.
  “There’s no shortage of data. We have so much data that it’s more of a challenge to figure out what we do with that data and how to make it work for us.” — Kevin Lindsey [00:53]

• Shift in ‘Big Data’ Conversations:
  Kieran comments on the cooling hype around big data, emphasizing the actual size of datasets has only grown.
  “The noise maybe has died down, but [data is] certainly just as big, if not tremendously bigger...” — Kevin Lindsey [01:33]

Defining Data Science’s Value in Marketing

• Data Science Beyond the Buzz:
  Kevin addresses the allure and practical applications of data science, highlighting its critical role in modern customer analytics and decision-making for marketers.
  “Data science is now sort of coming into its own in terms of a key area that the marketing organization needs to pay attention to...” — Kevin Lindsey [01:48]

• Do You Need a Data Scientist?
  Not every organization needs a full-time data scientist, but every business can benefit from data science principles—often by using accessible tools and platforms.
  “The nimble marketing team...still, however, can take advantage of data science in their practice. And I would actually challenge that. They probably need to be taking advantage of data science to some degree.” — Kevin Lindsey [03:43]

Analytics vs. Data Science

• Clear Distinction:
  Kevin draws a line between consuming analytics for insights (generating reports) and using data science for building predictive models and automating decision-making.
  “That distinction comes down to the difference between insights and an action that can be taken on those insights or that data.” — Kevin Lindsey [05:11]

• Automating Actions:
  Marketers can now leverage technologies (like Adobe Target) that bridge the gap between insight and automated action, even without hands-on data scientists.
  “There are technologies today...that expose data science through a marketer-friendly product so that marketer can rely on that decisioning that is happening...” — Kevin Lindsey [07:30]

How Adobe Target Automates Data Science

• Embedding Data Science in Tools:
  Kieran queries if technology can truly replicate a data scientist’s approach algorithmically. Kevin confirms this is now reality, describing lifecycle journeys and decisioning automation that help optimize customer engagement.
  “A data scientist can...build a model to accommodate and optimize against the business outcome that the marketer wants to see...by writing an algorithm...as part of a personalization strategy.” — Kevin Lindsey [08:14]

• Practical Example – Automated A/B Testing:
  Traditional A/B or multivariate testing exposes all variants to traffic until a winner is found. Advanced tools now automatically reallocate traffic as soon as a leading variant emerges, minimizing wasted impressions on subpar experiences.
  “As we start to feel the engine [is] confident that a winner is revealing itself, we’re going to start funneling more traffic over here...so we’re not going to continue to send people toward this dud...” — Kevin Lindsey [11:16]

  • Kieran’s Analogy:
    “The big hairy dog.” — Kieran Rogers, referring humorously to the weakest variant in a test [12:03]

• Surprising Outcomes—What Really Works?
  Kevin shares a story where a big media brand assumed a celebrity-led campaign would drive the most lift, but data-driven testing revealed the opposite.
  “The celebrity appearances were actually not creating lift. It was the one without the celebrity that created lift.” — Kevin Lindsey [13:10]

Machine-Led Optimization and Marketer Trust

• Data Science Yields Superior Results:
  Across Adobe’s client base, machine-driven experiments consistently outperform manual, rules-based targeting.
  “Nine times out of ten, the more data science-driven, machine learning-based approaches are beating the manual...decisions.” — Kevin Lindsey [14:23]

• Transparency and Control:
  The real challenge is helping marketers trust machine learning by providing clear explanations of what’s working and why—bridging the comfort gap in ‘relinquishing’ some control to algorithms.
  “The only way to do it...is by ensuring that the marketer has a lot of insights into why some of this is happening, the variables that...are the most predictive variables, and giving them as much opportunity to interact and really understand...” — Kevin Lindsey [15:06]

---

Notable Quotes & Memorable Moments

• “Data science and sexy, those don’t seem to be very compatible words...But if you look at the hottest careers right now...data scientists are right up there at the top.” — Kevin Lindsey [03:03]

• “Sometimes it’s really hard to determine which [test variant] is going to be the obvious winner and sometimes we’re surprised. And so letting the data figure this out for you… it’s like a content traffic cop.” — Kevin Lindsey [12:58]

• “Often the big hairy dog is the winner.” — Kieran Rogers [12:46]

---

Timestamps for Important Segments

• 00:53 – The central challenge: Making data actionable
• 01:48 – Data science’s current role in marketing
• 05:03 – Analytics vs. data science explained
• 07:30 – Marketer-friendly data science tech (Adobe Target)
• 08:14 – Algorithmic personalization and journey modeling
• 11:16 – Automated traffic allocation in A/B testing
• 12:58 – Entertaining “traffic cop” and “big hairy dog” moments
• 13:10 – Celebrity campaign case study
• 14:23 – Machine learning outperforming manual approaches
• 15:06 – The trust and transparency challenge for marketers

---

Key Takeaways

• Data science isn’t just hype—it’s increasingly central to digital marketing, but using it doesn’t require you to hire data scientists.
• The distinction between analytics and data science is pivotal: analytics produces insights; data science produces automated, scalable actions.
• Tools like Adobe Target automate complex experimentation, optimizing in real time and frequently delivering surprising (and often superior) results compared to human intuition.
• Marketers need transparency and insight into what algorithms are doing to trust and fully leverage machine learning-based optimization.

---

Further Resources

• Adobe Digital Marketing Solutions: adobe.com
• Adobe Digital Marketing Blog: blogs.adobe.com/digitalmarketing
• Target Internet: targetinternet.com

---

Transcript

[Podcast Host]

Welcome to the Digital Marketing Podcast, brought to you by targetinternet.com.

[Kieran Rogers]

Hello, and welcome back to the Digital Marketing Podcast. My name is Kieran Rogers, and today I'm here with Kevin Lindsey. Kevin, tell us, who are you and where are you from?

[Kevin Lindsey]

That's a really good question. A good place to start. So I am the director of product marketing from for a product called Adobe Target. I've been with Adobe for close to eight years, and I am very involved in optimization and personalization. It's been a passion of mine, and it is the key focus of the product that I look after at Adobe.

[Kieran Rogers]

Okay, so you're kind of into data science, then?

[Kevin Lindsey]

Absolutely. I am by no means a data scientist at all, but data is something that clearly drives us all in this industry. It's something that we've been talking passionately about for many years. I think the main thing to sort of acknowledge is that there's no shortage of data. Obviously, we have so much data that it's more of a challenge to figure out what we do with that data and how to make it work for us.

[Kieran Rogers]

It's a big problem, isn't it? That was the whole big drive a few years ago. Everybody was talking about big data, and I think a lot of the noise has died down on that. But hey, the data's just gotten bigger, right?

[Kevin Lindsey]

Right. The noise maybe has died down, but it's certainly just as big, if not tremendously bigger, than it was when we first started throwing that term around.

[Kieran Rogers]

So let's talk a little bit about data science. Has data science found ways of taming the beast?

[Kevin Lindsey]

Data science is one of those topics that has got a lot of allure and sexiness kind of around it at the moment. And data science isn't by any means new. Data scientists have exist within all sorts of industries for many years, but data science is now sort of coming into its own in terms of a key area that the marketing organization needs to pay attention to, marketing analytics. And the data that marketing is looking at around customer analytics is becoming so important to how we do our business and how we move the needle and be successful, that we really have to look at the role that the data plays and how it all can really, you know, come together in interesting ways and guide our decision making. And it really is a science to understanding that and making it work.

[Kieran Rogers]

I guess maybe you've just answered one of my questions, which is I was a little bit skeptical at first.

[Kevin Lindsey]

Really.

[Kieran Rogers]

Data science is alluring and sexy, but actually, if it can help us do this Very cool stuff. It actually has a really hot context in terms of marketing, right?

[Kevin Lindsey]

Yeah, you know, it is, I mean, data science and sexy, those don't seem to be very compatible words or terms. But if actually, if you look at the hottest careers right now and the highest paying careers and the most in demand professionals, data scientists are right up there at the top. Data scientists are in very high demand and are very much at the center of what many, many companies that are data first, data driven organizations are bringing them to the table. Now I think the other interesting aspect of this is that it brings up that question, does every organization need to have a data scientist? And the answer to that is no. It is one of those perhaps luxuries. If you are a major airline, your major financial services institution, or you're, you know, a large retailer, you probably have data scientists within your broader team. But that doesn't mean that the nimble marketing team, the E commerce startup, or the, you know, the new content oriented site that's just trying to drive significant growth needs to have its own data scientist. They still, however, can take advantage of data science in their practice. And I would actually challenge that. They probably need to be taking advantage of data science to some degree.

[Kieran Rogers]

Okay, so I mean a lot of people listening to this will have analytics and think, well, you know, we've got a lot of data in there. How can we use data to really enhance the customer experience? I think for years big data's been looked at, right? How can we get more money from customers, how can we sell more stuff? But this is a different fresh approach, Right. You're kind of coming at it through the customer lens, through customer shoes. How can we make their experience better?

[Kevin Lindsey]

Sure. I think it's important probably to clarify first though the difference between analytics and analysts who are taking advantage of the data, exposing interesting insights, producing reports and servicing organizations like marketing and the role of a data scientist and data science. And that distinction really comes down to the difference between insights and an action that can be taken on those insights or that data. The data scientist is charged with doing something with the data, creating models, using algorithms and math and calculations to take that data and make predictions and do interesting things with the data with regard to using a model to serve a particular business function. So I think it's important to make that distinction. For quite a number of years now, most companies, the majority of companies, have been taking advantage of analytics. They've been doing the the former, they've been using analytics and acting on those insights. They've been saying, look what the analytics is showing us. Here's a really interesting observation, here's an anomaly we're seeing, let's take some action on that. But now there's definitely a need. Given, you know, we talked about big data, it's getting more and more challenging for us to act in as quick a fashion as we need to by just looking at the data, then taking action as human beings, as human being marketers. And so what role can the data science or the models play in sort of creating linkages between insights and the ultimate action that you want to take? And so at many companies where data scientists exist, they'll look at a business problem and they'll say, how can we apply to data science to create decisions and actions that can be rolled out at scale across my organization, across my digital properties? And that's really where the difference lies in my opinion, in the opinion of many of the people I work with. And so back to this challenge of, well, what if I don't have a data scientist on my staff? Can I still take advantage of, of data science capabilities? And the answer is yes. There are technologies today, such as the product, I'm responsible for that. Expose data science through a product, through a marketer friendly product so that that marketer can rely on that decisioning that is happening as a result of some very intelligent modeling that's happening on top of your data.

[Kieran Rogers]

So to a certain extent can technology take those processes that any good data scientist would work through and kind of recreate those algorithmically, is that what you're saying?

[Kevin Lindsey]

Absolutely, absolutely. So if you take a data scientist might look at a problem like, well, we have a particular customer journey here. We understand this customer journey by consulting with the business owners within the line of business. And so, you know, maybe that journey is, it's not a simple journey. It's one where, you know, people are interacting with the brand on their phone and then maybe they're logging in once they're at home and they're with their family and they say, we're going to take a look at our finances together. We're going to make a decision about how much house we can afford, given our budget. All of these different kinds of things that kind of go into a decision that a potential customer might make about a mortgage or a car loan or something like that. They're complex journeys. And it's one of these ones where a data scientist can kind of look at that and build a model to accommodate and optimize against the business outcome that the marketer wants to see that actually can be accomplished by writing an algorithm that you can deploy as part of a personalization strategy on a website. And by doing that, you're actually taking advantage of data science that would be embedded within a personalization product such as Adobe Target. And so what the marketer really just needs to do is say, well, who is my audience? What do I want them to do? How do I define success? What kinds of potential experiences do I think are part of the mix here that could drive them toward the. The decision and the outcome? I would like to see. And we throw those into the mix and then let the algorithms kind of do the work to figure that out. And that is an example of a marketer taking advantage of data science.

[Kieran Rogers]

Okay, so let's move on then. Tell me more about Adobe's products that bring data science and enable this kind of level of personalization online.

[Kevin Lindsey]

Sure. So I mentioned one of them, which is something we call lifetime value decision. Another example is even just within our testing capabilities. Everyone, all your listeners are probably familiar with a B testing or ABN testing, where you might want to expose traffic to one or two or three different variations of an experience and figure out which one is working best. Well, traditionally, a marketer would have to basically set it up and run this test. And then once it reached a certain level of confidence, they'd go, you know what? We know which one. One. Now, as we do this, we have to acknowledge that you're ultimately exposing traffic to a winner, and then you're ultimately exposing traffic to a loser as well. Like in any test, you can have the winners and the losers. And so what a lot of marketers like to do is, like, really keep a close eye on this. We don't want them to peak. It's like one of these things, like, don't lift the lid off the pot until it's boiled. Right. And so it's one of these things where we've actually built data science into this process to say, all right, don't worry about it, don't worry about it. You can take your eye off of this thing. Because what we're going to do is as we start to feel the engine starts to feel that we're confident that a winner is revealing itself, we're going to start funneling more of the traffic over here to that so that we're not going to continue to send people over here toward this dud, to this experience that's not as successful as this other experience.

[Kieran Rogers]

The big hairy dog.

[Kevin Lindsey]

Right, Exactly. And it can be hard. It can be kind of elusive. Because no one sets up a test with one that's obviously the big hairy dog and a loser. We put up our best possible experiences, our best possible ideas. So sometimes it's really hard to determine which one is going to be the obvious winner and sometimes we're surprised. And so letting the data kind of figure this out for you, I like to call it a content traffic copy. It's using data science to act as that traffic cop and start to funnel things, to funnel traffic and visitors to the right experience in an intelligent way. So that's another example of data science at work.

[Kieran Rogers]

The thing I love about it is you never know, I've been doing a little bit of this recently and very often the big hairy dog is the winner.

[Kevin Lindsey]

Right. It certainly can be.

[Kieran Rogers]

Seriously? Seriously.

[Kevin Lindsey]

That works? Yeah. We had an interesting experience in the US recently with a big media company who had signed a pretty major celebrity as the face of the brand.

[Kieran Rogers]

Like you do.

[Kevin Lindsey]

Like you do. And so one would think, oh, this is a very, very popular celebrity, this is going to win. And someone decided, you know, we really shouldn't just rest on our assumptions here. And the interesting thing was that the celebrity, all the celebrity appearances in these offers and content were actually not creating Lyft. It was the one without the experiences, without the celebrity that were creating Lyft. So it's interesting because, you know, how one comes to make that decision about an endorsement like that or bringing it can be fairly, you know, you didn't go out and ask all your customers for their opinion, but they ultimately will give it to you if you test.

[Kieran Rogers]

That's fantastic. So what outcomes can an organization achieve if they get this kind of level of testing? Right.

[Kevin Lindsey]

You know, it's really interesting. They can absolutely achieve significant levels of conversion lift. Every organization defines success in a different way. And you know, for many it's engagement, it's top of funnel driving people deeper into the experience. For, for many it's very practical. Did we increase a conversion rate? Did we improve revenue? And honestly, in the experiments that we're seeing among our clients, nine times out of 10, the more data science driven machine learning based approaches are beating the manual or rules based or marketer driven targeting decisions. And so it's one of these situations where the marketer does need to become comfortable with this and relinquishing a little bit of control to the machine. And that is, it's a bit of a controversial topic at the moment and it's something that we're working on in terms of just trying to help people get more comfortable with it. And the only way to do it that we're seeing so far is by ensuring that the marketer has a lot of insights into why some of this is happening, the variables that are having the most impact and are apparently the most predictive variables and giving them as much opportunity to interact and really understand why things are happening, and then they are more comfortable with relinquishing some of the control.

[Kieran Rogers]

Well, Kevin, thank you so much for your time and insight into this. If we want to find out more about Adobe's marketing suite and how they can start taking advantage or take an interest in the kind of functionality that Adobe provide, that how do we go about that? What should we look at?

[Kevin Lindsey]

Yeah, definitely. Adobe.com is a great place to start finding some of the information. Plus our blog, which is just blogs.adobe.com digitalmarketing and we publish a lot of great content. There's thank you so much for your time. Thank you.

[Podcast Host]

Thanks for listening to another episode of the Digital Marketing Podcast brought to you by Target Internet.

[Kieran Rogers]

If you'd like to get more information.

[Podcast Host]

On the show, get hold of back issues of this podcast, or get details on any of the links we mentioned, please visit our website at www.targetinternet.com. if you've enjoyed the show, we would love to read your feedback. Please rate us in itunes or even better, write us a review. Or if you have any questions, please get in touch. We'd love to help.