As Marc Andresseen so famously said, software is eating the world, and as a VC, I’m always searching for industries that haven’t yet been disrupted by software, and companies that are poised to benefit from being the first software disruptors in those industries. Hence my excitement about Phrasetech.
The process for writing a product description hasn’t changed since the beginning of e-commerce. Someone, quite often in India or another third-world country, is hired to write the description and the other verbiage about the product (product name, product bullets,and product specifications).
It’s a major budget line item for large retailers that can easily have 500,000–2,000,000+ products that need appropriate verbiage. Depending on the quality, and the amount of verbiage desired, it generally costs about $15-$50 a product. It’s such a big cost that most e-commerce companies only update the product descriptions every 2–3 years, even though the descriptions are often out of date after 2–3 months.
Turns Out There’s A Better Way
Enter Phrasetech, leveraging NLG (Natural Language Processing) to automate the writing of all the product verbiage. This is not a trivial task.
Shmulik Fronman, Phrasetech’s Co-Founder and CTO, is an alumnus of the IDF’s elite Talpiot program where he served in Unit 8200 as a researcher, developing complex algorithms for solving unique mathematical problems. He later led a team that tackled challenges in the fields of data mining and statistics, managing a group of 25 algorithm researchers. Shmulik’s algorithmic expertise is complimentary to Co-Founder and CEO Eran Utmazgin’s 15-years in e-commerce and digital advertising. Most notably, Eran spent 7 years with Shopping.com (an eBay company), in both Israel and in California.
Phrasetech Saves, But The Far Bigger Benefit Is Increased Revenue
The cost savings are obvious having a machine do the work as opposed to a human. The savings is the most compelling feature of the technology, until a potential customer really understands the revenue impact of the technology.
In advertising, we all see personalized ads, tailored specifically to us based on a myriad of information known about us (age, location, gender, what food we like, what bands we listen to, our purchasing history, etc). That same technology is used in e-commerce, to decide what products to show us. But the verbiage around that product is generally the same, regardless of who we are, because, until Phrasetech, there was no way to dynamically generate verbiage optimized specifically for us.
In the simple example above, Phrasetech’s technology creates different product bullet points based on the gender of the shopper and how often they’ve viewed the product. It’s early to know the impact of product verbiage optimization, but research by Experian indicated that personalized email messages increased conversion by as much as 6X. The factors that can be considered and optimized for are endless, ranging from location, to season, to prior purchasing history, and so on, and so on, ad infinitum.
In addition, Phrasetech can adapt the verbiage to optimize for SEO. If an e-commerce site simply uses the text provided by the product company, Google knows its not original content, and the page gets a very low SEO ranking. By using Phrasetech, e-commerce companies can generate original content, in the voice of their choosing, resulting in higher SEO rankings. In addition, Google is continuously evolving its algorithms, and Phrasetech is uniquely positioned to help e-retailers optimize the verbiage on their site to adapt to the latest changes in Google’s SEO algorithms.
Early Traction Is Very Encouraging
Phrasetech already has it’s first paying customer, GetPrice, a comparison shopping network in Australia. More importantly, Phrasetech is running trials with several of the largest e-commerce companies in the world, across hundreds of thousands of products. The fact that some of the world’s largest e-tailers are testing with Phrasetech is a testament to the technology developed by this 10 person team sitting in a tiny office in Tel Aviv.
In addition, I’ve partnered with Iterate Studio, the R&D arm for 20 of the top 200 e-commerce sites in the world, who look to Iterate to scour the globe and find the best e-commerce technology for them to deploy. Iterate appreciates the world class technology being developed in Israel, so we’ve partnered to bring that technology to their clients. Iterate actually A/B tests different technology solutions on their partners sites and guages the impact of the technology. Iterate has never seen any technology solution like Phrasetech. So the A/B testing by Iterate is underway. Another testament to Phrasetech’s unique technology.
E-Commerce Could Be Just The First Vertical Disrupted By Phrasetech
I have a long history in semantic technology, including being one of the first customers of Oingo in early 2000 (when I was CEO of .tv). Oingo was later re-named Applied Semantics and purchased by Google for over $100 million in pre-IPO Google shares in 2003.
I’m also an investor in Grapeshot, which leverages semantic technology developed at Cambridge University to provide a blend of contextual and personal targeting for advertisers to optimize their online advertising campaigns. It’s the #1 app on AppNexus.
Any industry that has a large number of listings like restaurants, travel, or real estate, all present great opportunities to leverage Phrasetech’s technology.
So I’ve got great appreciation for the tremendous opportunity for Phrasetech to disrupt other industries with it’s world class technology. But first we’ll start with e-commerce.