Insights

Creating Personalized experiences in Real-Time: Part I

A conversation between Dynamic Yield and Gorilla Group on the current and future state of personalization.

Colton Perry
VP of Channel at Dynamic Yield

Colton Perry, VP of Channel at Dynamic Yield recently sat down to have a discussion with Gorilla’s Digital Marketing Director, Randy Kohl. Dynamic Yield uses an advanced machine learning engine to build to actionable customer segments in real time for the activation of truly personalized experiences. Colton and Randy discussed how personalization can inform ecommerce experiences over using data, Amazon, Google, and more. This is the first part of the engaging and informative Q&A, check back soon to read the rest.

What are the primary applications of personalization within ecommerce, and how does Dynamic Yield fit into the overall landscape?

Colton Perry: The primary application of personalization, first and foremost, is leveraging and activating data to create and deliver optimized and personalized customer experiences. The application of this can take many forms, and you might actually leverage data or the types of experiences that you deliver to an individual use. That can take the form of messaging and notifications in real time. It can be dynamic or personalized content and creative. It can be product or content recommendations. And it can happen across all the different channels.

So really we look at personalization more as a framework, a combination of different tactics and strategies that you can enable or implement that are aimed at presenting relevant content and experiences to a customer. It really helps us to make progress towards online business goals and objectives. So things like “how can I affect average order value,” and “how can I impact conversion rates,” or “how can I reward my most loyal customers”, and so on.

At Dynamic Yield, where we fit into this is that we take a fundamentally broader approach to the enablement of rationalization to our clients, so we don’t look at personalization as a thing. We are seeing a lot of continued fragmentation in marketing technology and personalization is not just, for example, product recommendations. It’s not just segmented delivery of content; it has to be a sum of many parts. So for us, sitting in this landscape, is that we have created a platform that is purely stat and cloud based that covers all of the different digital touchpoints, data and audiences in a centralized location.

Then we are able to inform or deliver any aspect of a personalized experience from transforming pages and templates to delivering dynamic and personalized creative, targeted promotional offers, real time messaging notifications, content and product recommendations, and so on all in real time. We very much take a machine learning approach to what we do, so we are continuously crunching data to understand individual user context. Anything we deliver from a variational perspective is contextually relevant on a 1-to-1 basis, and we are optimizing the desired business outcomes that our clients can then designate.

How should retailers begin to incorporate personalization into their ecommerce strategy?

CP: I was meeting with a top 10 retailer a few weeks ago in Boston, and when they started to wrap their heads around the options and capabilities that were available to them, they literally said “I see that with a platform like Dynamic Yield, I can do a million different things. So where do I start?” I think from that standpoint you can definitely take more of a crawl, walk, run approach. So you can start with a number of best practices that are rather simple. Things like very basic audience segmentation and targeting around things like demonstrated affinity based on category of product, for example. Starting to understand somebody as a higher, lower, or mid-tier average order buy (AOB) shopper. You can start connecting your inbound marketing channels, so the emails for ad groups that people respond to drive traffic to the site use that intelligence to then have a seamless experience from media marketing channel to the site, and then optimize that to get something like conversion rate.

Randy Kohl: I’ll add from our side that a lot of things, at a very high level, dovetail with yours. Part of that is knowing who your customer is. Really starting off at a broader level of what kind of different customer journeys there are, and then getting increasingly granular as we identify exactly where people’s journeys diverge and using those as opportunities to incorporate personalization. Another thing from that crawl, walk, run approach is that there are a million data sources that our clients work with. Making sure data is updated, cleaned up, and just ready to go. I think we’ve all had cases where a firm is trying to do personalization and has greeted us by our last name or a wrong name or just something else that is really just an unforced error. I think that is something that really needs to be thought about.

The overall strategy as you guys talked about, you can do a million things. What you want to do first should have the highest impact. We would also say to make sure as a retailer or brand that your internal teams and processes are configured to support the personalization program. Make sure everyone is on board, make sure there is a clear set of goals, make sure you have the KPIs that you want to measure. And make sure that it all incorporates with the technology, because the technology alone is not going to be the silver bullet. It needs a little bit of help to be successful.

How can online retailers use personalization to better balance the need to deliver high-impact email campaigns with positive customer experiences?

RK: Personalization is a balancing act. It is a requirement, but we want to provide relevance without crossing the line of intrusiveness. Making sure we give people what they need, when they need it; but not going too far and getting into that creepy area that we have all experienced. We like balancing promotions with value added content as something that we see as key because we do find that customers can get promotion blind if all they receive is promotion after promotion from retailers. No matter how personalized and targeted it is, it won’t be enough to give it variance. So adding some kind of value added content at intervals, but also having personalized content can really help engagement over the long term.

And then just learning users’ preferences. Letting them participate in the process in terms of email: when and how often do they want to be contacted. Asking them this is not necessarily a bad thing, and people will definitely let you know that they like to be contacted only so often. Less is more in some cases, being at the top of mind without bombarding someone’s inbox. So those are the different things we look at.

A great example of that is when someone buys from you, it is good to give some post purchase breathing room. We’ve all had the experience where we bought something, and right after you get the order confirmation you get another promotional offer from that same brand, giving you a better deal on what you just bought, leaving a bad taste in your mouth. So making sure that all works together to deliver the right experience at the right time for each customer.

CP: I completely agree, I subscribe to the mandate of don’t be creepy. I think there is definitely a line that should not be crossed and that as we think about application of different strategies around personalization, it should be more around relevance and that we know or understand your affinities or your preferences and things like that but we should not cross that line to where we are intrusive.

Especially today in the heightened state of data privacy around GDPR and issues going on with Facebook, we want to be responsible with the data that we hold and use it in a way that is responsible. I look much more around in how often we interact with people. For us it’s more important to consider, if we are going to send an email, how do we make it as relevant and personal as we can from a content and product recommendations perspective. For example, not looking at it from a quantitative perspective, it should be much more around relevance and quality.

Checkout the full interview on SoundCloud.