Taking the Guesswork out of Marketing – How Guess Uses Predictive Analytics

 

Guess Director of Marketing – CRM, Victoria Grahan, discusses how Custora is helping transform the ways Guess communicates with their customers.

Imagine you’re Victoria Grahan, Director of Marketing – CRM at Guess. You have the resources and support of global fashion brand and you work with a talented team of data-savvy marketers.

Surely you have nailed down marketing analytics and customer segmentation, your email campaigns are micro-segmented and ultra-personalized, and of course, your target CPA (cost per customer acquisition) is as chiseled as the muscles of those Guess models.
Maybe not.
We recently hosted a workshop with Victoria at Fashion Digital New York, where we discussed how a large data-driven organization with a dedicated, sophisticated in-house CRM and analytics team is leveraging external tools like Custora to extend and automate some of their analytics and marketing programs.

Here are a few highlights from Victoria. You can watch the entire 45-minute recording below.

 

 

The CRM team was very data-driven. We were starting to know and understand our customers, but all of our analysis was done manually and in-house.

 

The Guess marketing and CRM team identified two opportunities to extend the scope and impact of their analysis: Expanding the customer database beyond their loyalty program, and layering on predictive analytics.

 

Opportunity 1: Combining multiple customer and transaction data sources

 

 

The CRM team was focused specifically on customers enrolled in our loyalty program. If you work for an omnichannel retailer, you know that if someone walks into your store you have no clue who they are, unless they’re enrolled in your loyalty program. If our loyalty shoppers shopped online, we were capturing that too, but we were missing the fact that there was a whole group of online shoppers, that we could have been doing the same level of analysis on. That was a big opportunity for us.

 

Opportunity 2: Layering on predictive insights

 

All of our analysis was based on past purchases behavior. We’re not data scientists, we don’t have PhDs in our office, so any predictive analysis — like who might buy in the future, and what a customer might buy in the future — was not something that we were able to capture.

 

“Learning & Earning”
At Custora, we sometimes refer to the process of leveraging our platform as “Learning and Earning.” (Okay, a little cheesy. Watch the full workshop video for an extra cheesy pun involving Custora and a classical painter. And here’s a delicious cheese shop near our office in Manhattan – the Pleasant Ridge Extra Reserve is highly recommended. Anyway.)

For Guess, the “learning” part focused on identifying and understanding their customers across two key dimensions: Purchase-based customer personas, and high-value customers. The “earning” part applied these customer insights to Guess’ customer acquisition and retention campaigns.
Saying goodbye to “Batch & Blast”

 

Have you bought accessories in the past? OK, we’ll send you an accessories email or direct mailer.

“We partnered with Custora to identify various Customer Personas: Groups of customers who stand out as being different from the rest of the customer database in the way they shop. The key persona that we identified is the Accessories Persona.

“We’ve actually known that we have an Accessories Persona, but its definition was always limited to loyalty members and to past purchases. The definition was rudimentary and rule-based: Have you bought accessories in the past? OK, we’ll send you an accessories email or direct mailer.

“Now, we’re able to supplement this segment with customers coming from the Custora database, who are likely to buy accessories. We’re even able to add customers who have never bought accessories in the past, but something about their past purchase behavior, or their demographics, suggests that they might buy accessories in the future.”

 

 

If we’re going to email everyone every day, let’s talk to them in a way that’s meaningful and relevant to them.

“Guess has gone through a change in our email strategy. Three years ago we were emailing customers three, possibly four times a week. Today we email every single customer every single day, sometimes even twice or three times a day if it’s Black Friday, or if there’s an after hours sale.
“We were very much “batch and blast”, and our email calendar was driven by our merchant team: If there’s a product launch, or a big promotion, like 40% off all sweaters, that was driving the email calendar. We’re now in the process of changing that.

“We’ve been cognizant of the fact that it is quite likely that we were irritating our customers with constantly talking to them. We had two options when we talked about getting personalized with our emails: One option was to cut back on emails. If today’s email is about denim, and you like accessories, you just don’t get today’s email. But the thought of cutting down the number of emails we send out was scary. The other option was, if we’re going to email everyone every day, let’s talk to them in a way that’s meaningful and relevant to them.

“We did a test two weeks ago, and just got the results. We isolated the accessories persona and the non-accessories persona. We mixed them up and sent one version of an email that was already pre-planned, and happened to be about night out looks, and then took the other group and sent them an email that was accessories-focused.

The results look great. Two areas are especially exciting: clickthrough rate (CTR) and conversion rate.

We compared the accessories customers who received the accessories-focused email (group A) to accessories customers who received the regular email (group B). The CTR and conversion rate of group A far exceeded those of group B. There was a huge lift in CTR, and conversion was much higher. In fact, there was no conversion for group B. Group A — the accessories customers who got the accessories email — converted, and we made some money there. That was super exciting, and it’s helping us to build a case to roll this out in a bigger way.”

Identifying high-value customers to optimize acquisition

 

 

Even if the first purchase isn’t necessarily that $500 mega-sale for us, it could be that this is a customer who is going to repeat very frequently, and over the course of two years is going to be worth much more than the person who’s had one biggie sale and never came back.

“At Guess, we were always concerned with Return on Ad Spend (ROAS) and ROI. We have tried a myriad of acquisition programs — everything from renting lists to traditional display advertising. Everything has had very mixed results, which caused us to be very conservative.

 

 

A lot of what we did was based on assumptions.

 

“We assumed that we had a very metropolitan customer who likely lived in big cities, we assumed that she or he loved denim, but also loved accessories, and these assumptions were powering a lot of our acquisition decisions.

“Once we started using Custora, the first thing we looked at was who are our top customers and what makes them different. We discovered many interesting things. It’s true that our top customers live in large cities, and they do buy accessories. But when we looked at where they really over-index and differentiate themselves from the rest of our customer database, we found that they were more likely to live in suburban areas. Arizona popped as a big state for high lifetime value customers. Their first purchase tended to be a knit or a sweater or denim.

“Interestingly enough, when we did the opposite and looked at some of our lower lifetime value customers, they over-indexed in some of the more metropolitan areas – Brooklyn was one that popped for us – and often come into the brand through an accessories purchase.

“So for us in thinking about acquisition, while we know that our customers are metropolitan and like accessories, we now want to build lifetime value over time, and we’re switching some of our acquisition strategy to focus more on the high value customers.”

“Together with the e-commerce channel managers, we’re now looking at optimizing acquisition through lifetime value. Let’s not look just at that first initial purchase, but let’s think about this customer over their lifetime – over the course of a year or even two years.
“Even if the first purchase isn’t necessarily that $500 mega-sale for us, it could be that this is a customer who going to repeat very frequently, and over the course of two years is going to be worth much more than the person who’s had one biggie sale and never came back.”

 

 

The shift in persona mix is not something we’ll be able to do on our own; that level of analysis in the past would take us weeks to do, so the fact that it’s happening by magic behind the scenes for us is really helpful.


Watch the complete workshop with Guess Director of Marketing – CRM
The recording includes the full workshop with Victoria and Custora’s CEO and Co-Founder, Corey Pierson. It covers many topics beyond this blog post, including optimizing winback email campaigns with predictive churn detection, more examples of email segmentation, and aligning the organization around predictive analytics.

Fill in the form below to watch the recording.

Leave a Reply

Your email address will not be published. Required fields are marked *