Interview with Antoine Azar from Thirdshelf: Enhancing customer retention through data

thirdshelf

As one of our previous speakers, Antoine Azar from Thirdshelf accepted to give us, two years after his conference, an update about his work, his business, and ecommerce.

1.    What does your everyday life at work look like?

Startups are awesome for the particular reason that there’s not really an “everyday”. Every day tends to be different, because things come at you from every angle, you have the freedom to steer in any direction. You’re dealing with organized chaos, and for some people (like me), it’s a thrill. In my role as CTO, my expected tasks will be composed of syncing with our dev team regarding our progress, review our roadmap and its priorities, meet with my cofounders to review progress on our most important initiatives, work on PR and business development (which means meeting a lot of new and fascinating people), and of course, actual software development.

 

2.    What are Thirdshelf’s objectives?

Thirdshelf is a marketing automation platform with A.I. at its core. We use artificial intelligence to crunch through a retailer’s POS and ecommerce data to extract patterns and insights. These insights are then fed to our marketing automation engine, to intelligently re-engage the customers and grow sales. When a retailer turns Thirdshelf on, they basically get a world-class marketing department in a box, that is able to tailor-make a loyalty program and ongoing marketing campaigns for their business.

 

3.    What kind of data does Thirdshelf need in order to analyse data and optimize campaigns?

We connect to transactional systems such as point of sale systems and eCommerce platforms. From there, we’re able to extract transactional data down to the SKU-level. So we don’t just know that a customer made a purchase at a certain date for a certain amount, but we actually know what it is they bought. Based on this data, we can derive a lot of interesting insights.

 

4.    Concretely, how does it process data to enhance customer retention?

We have some very interesting technology working behind the scenes to understand customer behavior and optimize retention. The most basic level is called RFM analysis. We’ll build a profile on each customer based on the recency (R) of their last purchase, the frequency (F) of their purchases and the monetary (M) amount they’ve paid to date. Where it gets more interesting is in the sku-level machine learning algorithms of the platform. These algorithms’ role is to predict what a customer will buy in the future, at what date, and why, based on a timeline of past purchases. This is a very challenging problem in machine learning that we’re tackling, and we’ve received a R&D grant from the National Research Council of Canada to develop this technology.

 

5.    Could you tell more about the principles of machine learning?

Machine learning differs from traditional programming based on the fact that the developer is not hardcoding conditions and outcomes in the code, because he doesn’t know what they are, and there’s just too many of them. Instead, we ask the machine to look at data, and based on a mathematical model, to try and make sense of this data – finding correlations, deducing causations, posing a certain action, and then measuring the impact of that action to assess the quality of its understanding. Because it’s a closed loop, the machine is able to learn from its actions and keep improving as it collects more data points. It’s truly fascinating technology.

 

6.    Focusing on the customers, how does Thirdshelf build a high-value relationship with them?

The big buzzword in the industry is “omnichannel retail” (offering a smooth experience regardless of the channel customers use to interact with a retailer). Our vision for retail goes beyond the nuts and bolts of channels. We call it “relational retail”. We believe high-value relationships are built when a retailer offers exceptional customer service, regardless of the number of channels. Let’s stop obsessing about multiplying the channels, and let’s make sure we focus on a customer’s individuality and personalize our offering, even if it’s only through one channel. This resonates highly with independent retailers, for who 90-95% of their business is in-store, and investing online only has a tiny impact on their bottom-line.

This relational retail is based on values of trust, relevancy and respect.

Trust: Customers tend to have a lot of trust towards independent retailers, and for that reason we’re completely white-label. We don’t come between retailers and their customers (like many of our competitors do) because we want to preserve that relationship of trust.

Relevancy: We live in an incredibly noisy world, and what separates signal from noise is relevancy. Marketing is too full of irrelevant messages, promotions of products we don’t want, or sent at the wrong moment. We focus on personalized offers for the right products, sent at the right time to the right person.

Respect: The reason of an artificially intelligent marketing platform is to avoid blasting entire customer lists with generic messages. At its core, this comes down to respecting the customer’s individual needs, respecting their time and their attention. Customers appreciate this and reward those merchants accordingly.

 

7.    How does personalization take shape?

Personalization takes very different shapes. At the core, we personalize the content of the message and its timing. If you buy a mountain bike today, you might be interested in an offer to upgrade. But you don’t want to get it right away – you’ll want it in (say) 12 months. So both message and timing are critical. Secondly, with the multiplication of communication methods, the medium becomes very important. Some customers want to communicate via email, others via SMS, and others via a social network like Facebook Messenger. We adapt to each customer’s preference.

 

8.    Is it simple for a small business to incorporate this strategy?

It used to be impossible – technology was simply not available to small retailers, and the big retailers spend millions with big-corp suppliers like SAP, IBM and Oracle. With a platform like Thirdshelf, a retailer can turn it on in a few minutes, and start seeing the revenue boost in a matter of days. Because we’re connected to the transactional platforms, we’re able to measure in real-time the hard dollars we bring in. Retailers get very excited to watch their Thirdshelf dashboard and see the revenue grow based on each campaign we automate!

 

9.    Now that we understand better the process of customers’ reactivation, what about maintaining current customers loyalty?

Our platform’s first step is to segment the customer base according to their stage in the “shopper journey” – from first-time customer, to churned. Of course, our goal is to bring customers to the middle, most profitable segment of loyal repeat customers. A customer in this segment gets highly rewarded for their loyalty, and our platform works hard to keep them there.

 

10.  What’s next for Thirdshelf?

We’re working very hard on our cutting-edge machine learning technology alongside AI researchers (there’s some great ones in Montreal!). In parallel, we’re very excited about some of the joint initiatives we’re working on with our POS and eCommerce partners to bring this technology to many more retailers. As our customer base grows, the intelligence of the collective network of Thirdshelf retailers grows with it, and we’re able to keep increasing the value we provide to our retailers. Our goal is to become the reference for customer behavior in the SMB retail industry.

Leave a Reply

  • (will not be published)

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>