Lucy: Spotify. I listen to a lot of music.
Lynn: Liberty Mutual. Last year, during the hurricane in New York, they sent a relevant message. It hit me in knowing the brand was listening to what was happening where I was living. Keeping that in mind is part of my own strategy.
Narasu: Tell us about a business problem that personalization has helped you overcome.
Lucy: It’s pretty much a black box why a customer disappeared. Take that lead information and personalize it. Ask them, did you actually purchase this vehicle?
Scott: We have an online eye exam. Your eyes can’t be too messed up. Segmentation on data side, we do it through segmention and email the folks that qualify. Just launched a predictive model based on activity on website though social channels, email. We are in week two of that testing, predictive model is holding true. I’m pretty damn excited about it.
Lynn: Personalization started with Avon Ladies. All the email we send out in on behalf of the representative to their actual customers. This has been going on for years. This spring, we took years of data and asked what can we message in terms of creative. We built customers off a RFM model. Layers on top of that were sales, frequency, understanding when they were purchases, within which category and then developing emails based on that. Lots of testing. Resources and time taking to load it, lot of pain points as well. We would create very big spot of creative and complement it with a makeup offer, fashion offer. No one was touching those bottom pieces at all. They wanted their offer because they were a skin care buyer. Reduced to make it simple for team to focus on A spot, send email out. Every week, I changed the creative to see if they would buy something else. But they were true to skin care routine. I started to send only skin care offers.
Wayne: One of the areas we’re able to address using personalization is product expansion. Lot people who use Uber use for one specific reason. Some will use to go out for night. We’re looking for opportunity for other uses of products. An example, use personalization to identify the target, make message relevant is for business travelers from airport to hotel. A lot of them don’t use Uber Eats. Idenitfy users, can offer message, ”Hey one option is you could get room service or you’ve got these eight restaurants nearby. You can use Uber Eats, get better food that what they have to offer at the hotel. We have a lot of insights about our users. We can extrapolate what they might be interested in. What’s local, right here, relevant.
Narasu: How are you thinking about customer journeys and personalizations?
Wayne: A lot of our conversations are one on one. We know a tremendous amount about what a user wants. We’re able to detect patterns, behaviors that we expect people to take to get to the “happy path.” We’re mining for insights to determine how to personalize to get folks to that next step, whether that’s presenting things they’ve completed or what’s next. Understanding those next steps that take a customer from a new to a real power user, advocate of your brand.
Lynn: Collecting offers. Remerchandise the website. Things we can do that are relevant, understanding automated series going out. I have 45 automated triggers, first time user, pushing over to Facebook, complementing it with a relevant message, creative. Great results in conversion.
Scott: On the road map, we are unique in that contact lens are a single-skewed business. The prescription is to a specific product. You get a prescription so we can’t cross sell. In the pipeline, we are working on churn models. Can’t get to that until we get to people who buy form us three times. That’s where we’re starting, locating consistent cusotmers.
Lucy: We send out confirmation emails for action taken on app or on site. Send a lot of alerts. Doing well there. But there is a huge opportunity in building on a full customer cycle. For someone looking to buy a car, that can be as long as a 17-week process. One of those challenges is understanidng where customer is when they come to us, whether they are high-, mid- or low-funnel. I’m trying to look at the signal on website in terms of their behavior. They go through quiz, set up first search. We can determine where they are in the buying cycle, how far out are you from buying a car.
Narasu: What do you do to measure your programs? Is personalization worth it?
Lucy: We measure our email engagements in click through rates. Pretty early on.
Scott: We’re all revenue per click. Orders per email, per click. The email itself doesn’t have much personalizeation but website is personalized a bit. Your cart repopulated what your last order was. The idea is To make it as easy as possible to click through and make order as quickly as possible. Personalizaiton for me is asking yourself with every email what you are expecting recipient to do. Express exam. Testing from 6 am to 1 pm, RPE went up. Testing 1 pm vs 6 pm because buying contacts takes 2 minutes. Exam tasted 30 minutes. When are people going to have more time? We’re also testing not sending on early part of week but on weekends. Revenue might not be great but if we get more people taking the exam, we’re going to get more customers that way.
Wayne: Most measurement is in downstream. We don’t do a ton of focus on channel specific measures. Most campaigns intended to streamline. But it’s really important to focus on insights, strategy forming personalization. Determine what to measure. Should be an insight, strategy that says this is something we can address. This is an opportunity to add value by personalizing. As long as it’s driven by insight, then in terms of right level of personalization, the goal is to challenge self. What do i know now that I didn’t know before.
Lynn: We do outreach to local and national journalists. From that we send a lot of personalized to journalists with studie where city ranks in terms of mobility. Details about city, ranking. Monetary details around that. Lately not getting as much local content as we used to. Content team think about what other types of content we can offer.
Scott: if the email is not different enough form what thy normally receive doens’t matter. We can mail openers and not use revenue and I was wrong. If they’re not opening, they’re not interested. Push my data guy, take last week’s mailings in any channel and split it into operas and non operners. value in just being ther ein the inbox. We have a brand clearly has equity. Personalization test the living crap out of it. We’re testing the yes people vs the no people.