Dynamic Yield power user and CMO on the nuances, misconceptions, and foundations of personalization
An interview with our CMO on why successful personalization depends heavily on having the right people, expertise, and resources in place.
Dynamic Yield’s CMO, Yaniv Navot, recently joined the eCommerce Fastlane podcast with Steve Hutt. The following is an excerpt from Episode 119: “Gain a Competitive Edge Through Personalized Experiences That Turn Browsers Into Shoppers Into Loyal Customers.”
The discussion picks up with the host Steve Hutt, the Merchant Success Manager and Strategic Advisor at Shopify Plus (a valued partner of ours). He mentions that many of the merchants he works with are trying to figure out how to build authentic, one-on-one relationships between the brand and its customers. However, despite the need to better curate the shopping experience, many haven’t been able to effectively respond to consumer needs, going on to share a story of a friend who recently bought dog food from a Shopify merchant only to receive a follow-up email sequence with cat-related products.
Asserting personalization is about more than just inserting a customer’s first name into the body of an email but rather generating value, Yaniv Navot, the CMO at Dynamic Yield, dives into what it takes to move beyond these initial steps and into a truly customer-centric organization.
Yaniv Navot (YN): The success or failure of your personalization program depends heavily on having the right people, expertise, and resources. I’m not going to sit here and tell you the only thing you need is a magic software like Dynamic Yield that can work on its own. There are no magic tricks or shortcuts. Think about the marketing automation market – you have people who are experts at this and it’s what they’ve been doing for most of their career. And we see it happening all around us. We see personalization experts as job titles and our customers are looking to hire and fill roles within personalization and conversion rate (CRO).
Essentially, there are these three components that you have to solve internally: Do you have the right people? Do you have the right processes? And do you have the right resources in terms of design functions, creative resources, development resources, and all of these different functions that are going to be needed in order to execute your personalization program?
Now, it really depends on the scale and scope of use cases or your personalization vision. A company like Dynamic Yield is here to support you with that. Because what we can do is, first of all, help educate you. If you don’t have the right team or skills, we have a lot of knowledge that’s publicly available on the Dynamic Yield website. We also have training workshops and a lot of resources that are available for you to help you become better at what you do, independent of your work with Dynamic Yield.
We also have a huge partner network, and if you don’t have enough internal resources, they can do most of the work for you. If you don’t have a lot of development resources, we have what we call the templating engine that includes over a hundred out-of-the-box personalization templates you can deploy on the website with a click of a button. Things like exit-intent and social proof messages, dynamic banners, sliders, and even more complex widgets on the website – so you won’t need a developer to build that from scratch.
Steve Hutt (SH): What are the biggest misconceptions you think folks have about personalization technology?
The first thing many people have wrong is that personalization can only influence known users. Meaning, people who are already in a database or logged in. The assumption is that only if you know who they are can you dynamically change their experiences. But that’s fundamentally wrong because personalization can also be achieved for a completely new user. There are so many different things you can do, even within the very first session, without identifying the user through contextual data like the local weather, for example. So, if you’re selling coats and you have different types, exposing products that are for warmer conditions versus colder.
Learn more about what you can do to personalize experiences for anonymous visitors.
I do also see a lot of confusion in the market about the relationship between A/B testing, CRO, and personalization. Most believe these to be three distinct practices, but the truth is that they’re all interrelated and teams can use each according to what they need. They work hand-in-hand to achieve that one goal of improving revenue for business KPIs.
For example, you can run a personalization campaign that also includes A/B testing components within it. Think about a very simple campaign on your homepage – you have one banner that is geared towards men, and another towards women. Within each of these experiences, you have different A/B testing variations to see which banner for men would work better. And then maybe five to twenty variations which you’ll let machine learning algorithms choose as the top-performing ones. In other words, A/B testing is one of the engines behind the personalization campaign or program. So those three practices, A/B testing, CRO, and personalization work together, and if done correctly, will help you scale up your efforts.
Watch our recent webinar on the different interpretations of personalization, where A/B testing and CRO fits in, and how to take both a macro and micro approach to your CX strategy.
SH: As a follow up to that, what about overstock or skews that are very high, gross margin? Does Dynamic Yield allow for the promotion of these items through merchandising rules either manually or through an AI component?
YN: It’s a fantastic question because it’s so important, and absolutely. We don’t believe machine learning-based decisions should 100% replace human beings. Especially not in retail, where we’re talking about a merchandiser in most cases. This is their livelihood and expertise – they know which products to promote, when, and to who. Machines would not always be able to pick up on all of those nuances. We think automated algorithms can improve the efficiency and scale of the decisions humans initially make. If you think about the application of product recommendations, it relies heavily on different types of algorithms. Dynamic Yield has deep learning-based algorithms, and you can pick between sets of strategies that you want to use for specific purposes.
But then on top of that, the merchandiser can also introduce another layer of merchandising rules to control things like pinning products, changing the strategy per slot, and creating their own sets of recommendations strategies. They can also apply rules to show only products that are on sale or that are above a certain threshold, have limited inventory levels, and so on. Basically, we allow you to control all of that and create your own custom rules on top of the automated machines that are working behind the scenes.
SH: And what about dynamically reordering the collection to, for instance, push an out of stock item lower down the page? Does Dynamic Yield fit into that equation as well?
YN: Yes. We essentially take your data, which in this case would be a Shopify product feed, with all of its different attributes including stock availability, color, and price and so on, and give you the option to use them to make better smarter data-driven decisions on the website. How you want to use it is completely up to you – if you want to use it to personalize your product listing pages (PLP), you can do that with our deep learning algorithms. If you want to use our data activation to surface stock availability messages on product detail pages (PDP), you can also do that.
SH: This show has a diverse range of Shopify Entrepreneurs that listen each week… If you were to give some advice to those in the early stage of their business journey, what would it be?
What I’m going to say might sound obvious, but invest in data. Data is core to everything you will do, and you should get your data straight before it’s too late. What I mean by that is you should learn how to identify customers across different channels and then focus on the data that aligns with your business goals and values. It doesn’t make sense to measure everything and collect the entire world of data – just focus on the data that’s important to you.
For example, see if you can uncover valuable customer information, such as purchase history, affinity, preferred channel, and different KPIs like average order value (AOV) and more. Data will make the most positive impact if there is a specific reason why a retailer is choosing to use it. So use the data-driven approach to transform your organizational process around customer-centricity. This is one of the things that will allow you to better understand customer wants, needs, behaviors, and so on, and will eventually help you build a more accurate and relevant personalization strategy. Data is the foundation of every personalization program, so get it right first.
SH: For these early-stage brands, how do you identify a customer other than the one that’s actually on your website?
YN: Ideal resolution across channels and devices is very complicated. You should start with known users and make sure there is a clear incentive for them to log into the website because it’s not enough to say you can create an account and then use it. If you want to buy something, give them an incentive to also log in. Think about Amazon – you go there and you’re always logged in. Why? Because you know that by logging in the experience will be personalized. Your recommendations will be more relevant and you will be able to see your order history, making it easier to order again. So there is a clear incentive for you to stay logged in, even if you connect on a different device or computer. I would suggest that you focus on that before solving more complicated problems that would require the right tech.
Read our book about Amazon, which details how the eCommerce giant tailors its customer experience across channels, including dozens of use cases to leverage for inspiration.
SH: What should a mid-market brand be working on today?
YN: It really depends on the team’s level of personalization maturity. We run into large teams that are less advanced in their efforts and small teams that are super sophisticated. No matter the size of the brand we work with, we encourage teams to develop a clear understanding of their key audience segments. Again, going back to the data point and then identifying what each one of your audiences represents. This allows them to continuously learn from their behavior and also institutionalize insights, which brings brands closer to their customers. Essentially, it all comes back to the data piece – understanding your audience and identifying the opportunities you can influence through this deeper knowledge of the customer means teams will be able to generate a more accurate personalization strategy, improve different KPIs, and achieve greater incrementality at scale.
Want to discover your company’s personalization maturity level? Take this brief, 8-question survey.
I’ll give you an example. I’m a power user of Dynamic Yield, myself. I’ve been using it on our own website and drinking our own brew, so to speak. I’ve been running different A/B tests, personalization, and optimization campaigns, and I’ve been doing most of that work almost entirely by myself. I have a web developer who is helping me every once in a while with some small things, but really, 95% of it is done by myself, alone. I just looked this morning and I have around 39 running A/B tests that are currently active on the website. Just think about the number of tests and optimization campaigns that are currently running on our website, which is a relatively small SaaS website – dozens of tests are running and just one person is creating all of these things. With the right state of mind, the right approach, and the right tool, obviously you can make a difference. You can do more things with fewer resources and still make an impact.
SH: So what does the future look like for Dynamic Yield?
YN: We all know 2020 ended up much differently than anticipated, but I’m grateful to say the company has been doing amazingly, even in these difficult times. And while nobody knows what the future will hold, we feel one thing has become very clear. That is the ability to tailor digital experiences and quickly adapt to shifts in consumer behavior, whether you’re a large, medium, or small business, has been an invaluable tool for many. COVID, unfortunately, has tested the agility of teams and organizations, and I suspect that this will remain true even as we continue to navigate through their recovery phase, into 2021, and beyond.
From a solutions perspective, I’m excited to share that Dynamic Yield is moving towards different applications of deep learning-based algorithms, which I’ve mentioned a bit throughout our conversation. And we’re going to slowly, gradually release more and more of these applications to allow for the scaling of our customer’s personalization programs. We’re planning some really interesting and innovative product features and new capabilities for 2021 – stay tuned!
Are you a Shopify customer? Learn more about the Dynamic Yield Shopify App and how to bring personalization to your store.