Digitally-native eCommerce businesses are used to working with their customer data in order to write copy for marketing campaigns, run PPC ads, calculate customer lifetime value, and make decisions based on core metrics within CRM dashboards.
They understand the power of data to earn revenue and win market profit share perhaps more than even larger, older industries such as banking and insurance.
Some of the largest online retailers have been in business long enough to have enough data that one might accurately call it “big,” and these eCommerce businesses are eager to leverage their big data for artificial intelligence and machine learning projects.
That said, we’ve found over the course of many interviews with retail, eCommerce, and marketing experts, as well our research into AI vendors offering products to eCommerce companies, that AI applications in eCommerce are underdeveloped.
There are many AI vendors purporting to offer AI products to eCommerce companies. Oftentimes these products involve predictive analytics and natural language processing. These vendors claim their software can predict the success of potential marketing campaigns, find new customer segments, and even place optimal bids on Google Ads automatically, without input from a marketer.
Unfortunately, more often than not, these eCommerce-focused AI vendors are actually being deceptive about the capabilities of their AI products.
They tend to hire fresh college graduates and call them “data scientists” on LinkedIn to pad their ranks and appear as though their team consists of cutting-edge data science and AI talent, when in reality, not even their leadership teams and C-suite have the academic or business background in artificial intelligence necessary for developing a robust AI product ready for sale in the enterprise.
We won’t name names, but even some of the highest profile “AI” vendors for marketing and sales are fudging their expertise. As a rule of thumb, business leaders should air on the side of caution when searching for an AI solution to their business problem, and they should never look to integrate AI into their business for its own sake. They should always assess whether AI is appropriate for solving their business problem or not.
At this point in time, eCommerce leaders should focus on preparing their systems for when machine learning solutions are farther along in their sector. To do this, they have the option of contracting with vendors that sell data management systems built to allow for artificial intelligence applications when needed.
Most of these vendors claim to offer data organization for the eventual use of predictive analytics, and some of them offer platforms with built-in predictive analytics capabilities, however nominal.
In this article, we’ll run through three big data platform and/or big data analytics vendors that offer software for managing customer data and predicting successful campaigns and sales, and we’ll begin with AbsolutData.
AbsolutData
AbsolutData offers a big data platform with built-in predictive analytics capabilities that clients can use to gain insight into their customers’ shopping behavior, make product recommendations, determine optimal product pricing, and create cart abandonment flows.
The company claims its platform also has a recommendation engine aspect to it that can recommend products to customers based on their location. It’s likely that these products are recommended based on what customers in a nearby location have purchased in the past.
The platform can also purportedly determine the impact that returns and exchanges are having on the bottom line, which could inform decisions on how they are handled and offered to customers.
The predictive analytics algorithm behind the platform also seems able to predict demand for certain products and certain categories of product. As a result, eCommerce brands could prevent out-of-stock scenarios by ordering new stock well in advance.
In addition, the company claims its platform can optimize how much money is spent on promotional campaigns in a variety of channels. That said, it’s unclear how it might do this, especially considering what we discussed above. This is a difficult and nascent use-case for machine learning at present.
Overall, AbsolutData claims its software is capable of a large variety of big data-based analytical processes, but the company doesn’t seem to have a demonstration for how parts of its software work. The company lists Adidas, Dole, Epson, Etrade, Hershey’s, Hyundai, and Levi’s as just a few of its past clients. It lists many others on its website. AbsolutData has raised $20 million in funding from Eight Roads Ventures.
LK Sharma is the Head of Technology at AbsolutData, and he holds a Master’s in Computer Science. Sudeep Haldar is the SVP for Growth Analytics and AI Solutions at AbsolutData. He holds a PhD in Marketing from Cornell University and a BS in Electronics and Electrical Engineering, focusing on AI from the Indian Institute of Technology. Previously, Haldar served as Senior Director for Strategic Insights at Kraft Foods, Director for Strategy and Analytics at Staples, Senior Marketing Consultant at McKinsey and Company, and Manager for Advanced Analytics at ACNielsen.
Amazon Web Services
Amazon Web Services offers software as a service for eCommerce brands to build data lacks on which they can run AI-based big data analytics processes, including predictive analytics. Amazon claims eCommerce companies can import their data using AWX Direct Connect, which purportedly allows for a secure and reliable connection between the system from which the client is uploading their data and the newly built data lake.
The process of uploading the kind of volumes of data most eCommerce brands have can be lengthy, and any disruptions in the process could result in missing data on transfer. In addition, it can be costly. Amazon claims its services can help make this process go smoother.
AWS’ big data analytics include a variety of different applications for several use cases, including real-time analytics, data warehousing, and big data processing. That said, the company is quite vague about the benefits that eCommerce brands can garner from using their platform.
It’s very clear that AWS’ platform is intended for a company’s data scientists, which may mean it’s at present inaccessible to most eCommerce companies. The vast majority of them don’t employ in-house data scientists, but AWS’ software is a framework on which data scientists can work with data to build machine learning algorithms that can run on that data.
However, below is a short 5-minute video explaining how AWS’s big data offering supports marketing initiatives by collecting data from a variety of devices such as mobile, tablet, and personal computers. The video explains that the AWS big data platform is able to process 15 to 17 terabytes of data daily:
AWS claims to have helped ride-hailing company Grab use big data computations and data streams to serve 1.5 million bookings in Southeast Asia. Grab needed to scale to serve the huge demand for rides from customers. At the same time, the company needed to be stable to ensure its drivers are able to serve the riders efficiently. Grab turned to AWS to build its big data infrastructure that could address current and future big data needs.
Once deployed, the infrastructure was able to maintain a steady data stream and serve the company’s engineering, marketing, data, and other teams. The company was also able to build predictive analytics models which allowed it to alert drivers which areas would have high demand for rides at a certain time of day.
The big data infrastructure being on the cloud took off maintenance and operations off Grab’s hands and saved the company from 30 to 40% in resourcing and manpower costs.
AWS also lists NASDAQ, Unilever, Yelp and Novartis as some of their past clients.
Fractal Analytics
Fractal Analytics is rather vague about what their platform can do and the benefits eCommerce clients might garner from using it, but their team members have the academic and business experience necessary to build machine learning algorithms. As such, we felt they were worth discussing in this article.
Fractal Analytics claims to have worked with an unnamed specialty retailer to improve its “redemption revenue,” likely revenue generated by customers that had previously abandoned cart or viewed a product page but didn’t purchase.
The company claims their software came up with 50 customer “markers” for over 60 million households, allowing the client to determine which retargeting campaigns to send to which customers. As a result, the retailer purportedly increased their redemption revenue by 230%
Fractal Analytics also claims to have helped an unnamed retailer implement an AI-driven application that enabled the client to track the online behavior of its shoppers. The data gathered from the site visitors’ clicks uncovered issues that discouraged the visitors from buying products.
According to the study, the patterns discovered from the data allowed the retailer to make changes to its website. As a result, this boosted sales by 25%.
The company lists Aimia, Philips, and Franklin Templeton as some of its past clients, and has raised $125 million in funding from Khazanah Nasional, Aimia and TA Associates.
Prashant Warier is the Chief Data Scientist at Fractal Analytics. He holds a PhD in industrial and systems engineering from the Georgia Institute of Technology. Previously, Warier served as Vice President for Retail at Fractal Analytics and Senior Research Scientist at SAP.
Header Image Credit: Blu Constellations
