Menu
RSS

How Retailers Can Set Themselves Up For Successful Personalization At Scale

  • Written by  Q&A with Colton Perry, Dynamic Yield, and Reid Bryant, Brooks Bell
How Retailers Can Set Themselves Up For Successful Personalization At Scale

Q: How should a retailer structure its team internally to drive the best results with personalization? When is it time to think about bringing in outside help?   

COLTON PERRY: Personalization needs to be an interdisciplinary and broad initiative.  When done correctly and strategically, it touches multiple departments, teams, and constituents.  These can include email and CRM, marketing, customer support and call center, digital and customer experience, merchandising, analytics and more. That being said, it is crucial to have someone on the team who is able to work with internal teams, translate their goals and customer profiles, and create and implement an integrated personalization program.

If that resource does not exist internally, partnering with an agency or consultancy with expertise in personalization and optimization is a great way to go.

REID BRYANT: We believe that personalization programs can be a natural evolution from existing optimization programs. It’s best to bring in outside help when you no longer see desirable results from your optimization program, or you simply want to stay ahead of the competition.  With mature optimization programs, you begin to see diminishing returns when trying to optimize for all users.

ADVERTISEMENT
The opportunity to reset the ‘result for effort’ curve comes from contextualized experiences implemented via your personalization program.  Personalization is inherently more complex than optimization so it can accelerate your path to ROI when partnering with an experienced agency.

Q: Let's talk about data. What are the most important data sources to inform your personalization efforts?

BRYANT: From experience, the most important data in structuring personalization programs is appropriately granular and organized behavioral clickstream data. A user’s past site interactions give you the best clue as to how they interact with your brand and what they will likely be seeking in the future. By offering a value proposition to facilitate authentication, you can tie that clickstream data to a customer profile and then bring in any additional data like in-store purchases.

PERRY: It’s less about the sources of data and more about how you leverage the data to create optimal and relevant customer experiences.  For example, it means nothing that you can onboard CRM data; the key is that you use that data to understand where an individual is in their relationship with a brand, whether it’s lifetime value or purchase history, and craft customer experiences that meet and address their needs proactively. That’s where the magic happens.

Q: How should the volume of traffic shifting to mobile web and apps affect your personalization strategy? Do the same best principles from desktop apply?

BRYANT: With the shift of traffic to mobile and native apps, it is becoming more important to gain a unified view of the customer, in an authenticated state, across all devices. If the user is not authenticated, it limits the inferences that you can draw and the context that you can provide for each interaction.

Mobile experiences are unique and mobile offers the possibility of layering in locational cues to augment an experience. For retailers, you could envision a scenario in which a customer has purchased online and is presented with a coupon for a relevant upsell when arriving to pick up their item in store.

PERRY: While the overall goal of presenting the most relevant content endures, there are fundamental differences to building mobile-first experiences. For example, on a mobile device, bringing users back to the context of their previous shopping sessions is essential to driving conversions. A user may have two minutes to complete a purchase before her train darts into the tunnel. As much as you may love your new home page banner, showing previously viewed products is better use of that real estate.

The ambiance around desktop shopping is often far more pleasant, so encouraging serendipitous experiences aimed at getting a user to add more items to a cart or increase basket sizes is often a more fruitful approach.

Q: With all the buzz around AI, will AI tangibly transform personalization in the coming years? If so, how?   

BRYANT: AI will certainly change optimization and personalization programs. As marketers gain additional proficiency in data science and technology facilitates its adoption, marketers will be able to hand off targeting work to algorithms. In the short term, offline studies will help identify the most common and profitable customer groups, creative teams will build those different experiences, and AI will match the visitor to the most appropriate experience via machine learning rather than manually-set targeting rules.

PERRY: Marketers are going to start to demand tools that not only ingest large amounts of data but can actually begin to proactively identify opportunities. Over time, machines will be able to study data patterns and determine the content that performs best for different segments and proactively present this intelligence to the marketer.

Q: What is one piece of ’conventional wisdom’ in personalization/optimization that you believe is incorrect?

PERRY: The idea that personalization is somehow still a ‘nice to have’ and not mandatory.  Customer experience and brand relationships are the new battlefield; competing on pricing is a dying proposition and a race to the bottom.  We know that personalization will drive an $800 billion revenue shift to the 5%-15% of brands that get it right and can lead to hundreds of millions of dollars in sales uplift for enterprise clients.  The hard truth is that serving static experiences through a CMS simply isn’t enough to compete anymore.

BRYANT: I believe that true 1:1 personalization is not the north star that all programs should be initially shooting for. Maybe it’s semantics, but to me, 1:1 means that every customer will have a unique experience — unique creative, personalized site navigation and pathing, prominent recommendation modules, etc. I think that is overkill and could reduce a user’s ability to serendipitously explore your brand.

Q: How do you measure the ROI of personalization? What metrics should e-Commerce marketers pay most attention to?

PERRY: We love looking at uplifts in average revenue per user — it’s the easiest way to prove impact. If each visitor is worth even $2 more, think about what that will mean for your bottom line!

When evaluating the ROI of implementing personalization, it’s also important to evaluate several soft factors in addition to gross revenue uplifts. Serving better experiences will often boost brand affinity, customer loyalty and word-of mouth from happy customers, soft benefits that should not be undervalued.

BRYANT: I agree that the best way to measure ROI is directly through challenger/control testing. However, for personalization, it doesn’t have to manifest itself in a traditional A/B test with a defined traffic allocation and a defined stopping point. It is, however, beneficial to retain some baseline for measurement over the same period for causal inference. The importance of measuring ROI of your tools and partnerships should not be diminished, it’s the ammunition that will allow you to advocate for future program funding!


 

Colton Perry is VP Channel at Dynamic Yield, and Reid Bryant is VP Analytics and Data Science at Brooks Bell.

back to top