x
READY TO BECOME A MEMBER?
Stay up to date on the digital shelf.
x
THANK YOU!
READY TO BECOME A MEMBER?
THANK YOU!
“Shoppers come to the site, they've got their own goals, they've got their own process. They don't care how you organize the site. It was all a waste. They immediately would come and be like, ‘This is what I want’ and just jam that into the search box.
At that point, you're off into search land. One of the insights we had was that facets were a way you could carry the information architecture.
You could take your categorization — the way you'd organize your products into groups that were meaningful — and carry that into the search experience.”
— Adam Ferrari, VP of Engineering at Salsify, Former CTO of Endeca
From the beginnings of faceted search to the powerful role that product data now plays in increasing brand manufacturers' search visibility, ecommerce tech has come a long way in the last 20 years.
Dr. Adam Ferrari has been a part of all of it.
As chief technology officer at Endeca, an information retrieval and analytics company that powered the search and navigation for Walmart, Target, The Home Depot, and hundreds of other companies in the early 2000s, Ferrari has been at the forefront of retail tech innovation for two decades.
On a recent episode of the Unpacking the Digital Shelf podcast, "Ecommerce Tech Stack History From the 90s to Now," Ferrari indulged in a bit of nostalgia and looked back at the origins of ecommerce tech, how the landscape has evolved, and what this means for today's brand manufacturers.
In the early 2000s, the ecommerce tech landscape wasn't as expansive or as sophisticated as it is today. Ferrari said the tech stack was more primitive — and search was no exception.
Companies used document-level search engines like Verity. But without a lot of product content, this led to inefficient searches that didn't surface relevant information consumers needed.
"At that point in the late 1990s, you had really bad first-generation ecommerce experiences that were stood up, and you could do rudimentary shopping, but it certainly wasn't convenient," Ferrari said.
"It wasn't easy. If you landed on a site in those days — boy, like navigating it — you'd go in there, you'd try to search for something, and you'd get a zero results page," he said.
"At Endeca, the core thing that we came up with in those early days was this idea of faceted search. The idea that it wasn't just about keywords, but you also had structured attributes in product data that you could navigate by," he added.
Thanks to the foundation laid by companies like Endeca, modern ecommerce search has evolved from a keyword-centric process to incorporating other key factors that are important to search relevance.
These factors include whether an item is on sale or among the most popular in a brand manufacturer's catalog. In the process, this has accelerated consumers' path to purchase and given brand manufacturers another lever they can pull to increase conversions.
One of the reasons ecommerce search was so friction-filled in the early days was because companies had lots of unstructured data that could be tagged in multiple ways. Business intelligence tools also weren't as advanced as they are today, so making sense of all this data and categorizing it in a meaningful way for consumers was a monumental task.
Before Endeca, no system actually could scale and produce search results in real-time in a way consumers would expect. Ferrari's company made meaningful progress in this space because it took a different tack.
"We started with what the user experience wanted to look like," Ferrari said of Endeca's approach to search.
Ferrari said wrangling "messy, fast-changing data" was a crucial part of Endeca's magic. The company was able to preserve the investment brand manufacturers had made in thinking about the organization of their catalog. Also, they made their catalogs more interactive, dynamic, and responsive to search.
"It [the data] was volatile. It was sparse, and it was tough stuff, but we wanted the experience to be this almost like a concierge of 'Here's all the options that you could select!' and have that be super fast since it's shopping, and then work backward from that reality of the data and desires of the experience," Ferrari said.
"We actually got to a really pleasant underlying technology insight, which was the structure of a traditional search index could be ever so slightly morphed to also capture structured data and deliver a really, really high-speed way of aggregating that [data] into those facets."
"It was kind of cool and one of these nice stories of user experience leading the technology," Ferrari said, adding that this led to an epiphany that's still true for ecommerce today: "Search is the site."
As ecommerce tech has evolved, cleaning and standardizing retailers' data has become even more critical, especially when deploying big sites for leading brands with huge catalogs.
In the beginning, Ferrari said his team realized that they had to help retailers operationalize data clean-up, but as time went on, it became increasingly clear this wasn't scalable.
"The lesson we learned through those years was that the retailers were going to need to source the data better from the vendors. It was just too complicated and messy for them to do it centrally," Ferrari said.
Unfortunately, it took more than a decade from the emergence of ecommerce tech to come to this realization.
By about 2010, ecommerce tech faced a whole new set of challenges: a fragmented ecosystem where retailers cobbled together different technologies to build the infrastructure necessary to launch their ecommerce sites.
Mid-sized and smaller retailers opted for commercial off-the-shelf solutions while big retailers built their own custom ecommerce platforms.
"The early days of the tech stack were changing very quickly. Replatforming was the constant thing, just by the necessity of web technologies," Ferrari said.
But as ecommerce grew, so did retailers' digital maturity, and they began making more platform investments.
"That's where you saw those tier-one retailers really recognize that, in effect, they were tech companies," Ferrari said.
Eventually, with better data attributes came a better search experience. And, in turn, a better shopping experience.
Application programming interfaces (APIs) have now made ecommerce tech extensible, giving retailers a low code, low-touch way to add additional features and functionality that enhance the online shopping experience.
"Things are a lot more nimble, and you don't have to actually do code integration for every last bit of your site," Ferrari said. "So, we're in a moment in time where tier ones are these completely custom, gigantic worlds of their own."
"But the broader ecosystem and the ability for brands to have a really compelling experience is now possible — if we think about facets and relevant attribution that gives you a great shopping experience, who knows that better than the maker of the product? Suddenly, the technology is there, and it's nimble enough that they can deliver a really beautiful shopping experience," he said.
A thriving API ecosystem has led to the emergence of headless commerce. This IT architecture enables developer teams to separate the front and backend of retailer sites and streamline content management.
Ferrari said mobile shopping has been a key driver of headless commerce.
"We shop on mobile now. It's the primary shopping location. You want to be able to deliver consistent but customized experiences across channels. That's where the idea of headless commerce emerges from, is being able to decouple the glass from the core merchandising, catalog, and the underlying backend concerns of commerce," Ferrari said.
"You see all the commerce platforms being influenced by the headless trend and providing better API capabilities because there's just a greater desire to deliver innovative shopping experiences," he said.
Though Ferrari is happy to reflect on where ecommerce has been, with a robust API ecosystem, he's excited about where ecommerce tech is headed.
"I tend to think of these things as evolutionarily versus revolutionarily. We're going to continue to see this trend towards service orientation. API tech will make it a lot easier and more powerful to compose things," Ferrari said.
"The innovation happening now is much more in the plane of the consumer experience, which is exciting. It means great things for us as shoppers," he said.
Listen to the full episode to hear more about the evolution of the ecommerce tech stack from the 1990s until now.
