Chris Reddick (President and CEO of Clarity Ventures) and Ron Halversen (Vice-President of Sales and Marketing at Clarity) discuss the many ways that product catalogs can be adapted to your industry.

Part 3 of an 8-part series (Return to Part 2)

RON: Now let's move on to—and again, you mentioned the bulk upload of the products, but that's a big one—let's talk about the products. Because theoretically, the online marketplace is a massive product catalog, right? And all the logistics of both sellers, buyers, and then you as the aggregate owner...there are different contexts and different things that everybody has to deal with when it comes to SKUs and products. So I'm going to throw a couple of the things out there and then I want you to dive into them. 
 
Obviously, category and category selection, [that’s a] big one. Search, filters, store-within-a-store. I find something that Chris sells that I like, he's a personal trainer. I want to follow him, but he's got 2700 products within eight million products on the Amazon store. Do I have the ability to favorite Chris' store? Is a favorite seller so that I can come back and just click on Chris, and boom, I'm looking at Chris, his store within the store, right? They call that store within a store.  

Can I watch specific sellers? Can I sign up when they launch new products? Can I get automatically notified? What are my notification rules and how do I deal with products? Are there out-of-stock alerts? Obviously all the wish lists and favorite lists and different things that we use on all the different eCommerce marketplaces we shop. 
 

what is hipaa

RON: So those are just some of the things that I want you to dive into when someone wants to create a marketplace. And obviously, please expound on anything else you've got, especially when it comes to bulk import and how do we deal with UPC codes or ASINs, and how do we deal with multiple competing lookup numbers? So I've got a competitive SKU, but if somebody is looking for that product, my products just to know OEM version of that product or my products, the real brand version of that product. So how do I go in and specify multiple custom SKUs in different numbers that that are the same product, right? 
 
CHRIS: That's a great jumping off point specifically, because I've been in meetings with clients who have a very large marketplace with hundreds of thousands of different products to millions, multiple millions of different products, and it's really interesting at that scale because a user can go and type “hammer,” and you would think, “Oh yeah, it's going to pull up a hammer.” And things like a “drill hammer” come up because it has hammer in the title. And you could have like a sledgehammer and you can have different types of hammers just like you described with the eCommerce SKU versus an OEM and some of the different options.  

The system itself needs to be intelligent and as you scale, it has to be exponentially more intelligent. And so it needs to be able to think and get information based on what the users want and effectively read their minds. And this, at least as a minimum, as a baseline, has to be as good or better than everyone else out there. If it isn't, then you lose. That's it. And that's the nature of C2C, B2B, or B2C marketplace eCommerce. 
 

what is hipaa

CHRIS: Keep in mind that you have 10,000 pound gorillas like Amazon, where a lot of times people will just buy from them because they know the Amazon process, and governance is extremely solid, and they have the resources to back that up with the logistics and the accountability for the sellers and the process that's really robust. And so foundationally, it's getting the products in, having a process for allowing those products to go through some audit that ideally at scale needs to be more and more automated and intelligent. 
 
But in addition to that, the marketplace platform also needs to be intelligent with how it's presented to the user. And this is really, really challenging, but it's a solved problem. It's not something that is literally rocket science, for example. It's something that's been solved. And so fundamentally what it boils down to is having a really solid feedback loop and then being able to isolate the different niches and categories so that you can have a very specialized set of search results for every category and for the different visitor types that are on the site. 
 
With the technology today, with deep learning and machine learning, which are subsets of AI in eCommerce, it's even more possible and even more important that you consider those types of technologies within your search results and your overall experience with your site. A lot of the users that are going to be using your marketplace site will have experienced the benefits of these types of capabilities, but they won’t necessarily know that it's AI and machine learning and deep learning that's going on. They just see the benefit of it.  

As you scale up, you're going to want to implement the ability to leverage your reporting and analytics data on your user behavior and allow the users to have a very personalized experience over time, and form an opinion quickly, and then read the user's mind. That's the bottom line here. 
 

what is hipaa

CHRIS: This applies across a lot of nuances like you were talking about, Ron, with, “Do they have the ability as users within the marketplace to give feedback to the system that's explicit, that's very direct, like selecting favorite stores? Or is the system just going to learn based on their behavior? If there a lot of feedback that this marketplace can get just based on how long people are looking at things and what their intent appears to be based on their behavior within that particular product detail page, or within a particular search result. 
 
We're responsible, within the marketplace software side of things—as the operators of the marketplace, as the developers of this technology—we're responsible for making sure that we're at or above this baseline that you're seeing in the overall marketplace market. I’m interested to hear if you have any thoughts on that. It is a very dynamic space, and one of the biggest foundational pieces is being able to deal with the fact that there is going to be potential to have a lot of confusion. 

As an eCommerce marketplace owner and operator, you have to work with software that's going to be proven to be able to deal with that and elegantly present it to the user in a way that they don't see or deal with the inner workings. They just get the result that they're expecting. And then you have to have a constant feedback loop around that, one they can directly influenced by favoring things, by selecting stores, by changing the filtering based on the category they're in, and it being very discrete and specific to that category.  

But also intuitively learning based on their dwell time and based on their actual searches. “Okay. This person is really interested in custom wheels for their car.” So we need to learn from that and present accordingly. 
 
RON: Yeah, I agree. Some call that machine learning language, some call it AI, some call it predictability and personalization. But I know that we've been doing for the last few years—and it's funny because we've had a machine learning module in our platform for more than five years. I don't know, I've been selling it for a long time, but it's only been about the last year that people are really starting to understand it.  

You know, it's almost like they call me, they ask me, “Can you guys design and build a high-end website?" And then when I start asking questions about, “How do you want to market it, what's more important to you, SEO or CRO? We need to do both” And they're like, “We don't need to do that.” And it's just like they understand one concept and myopically think about the one thing, but they don't think about the full aggregate view of everything that they need to deal with.  

Continue to Part 4 to find out how AI works in marketplaces.