Start-Up Looks to Power E-Commerce Sites with Search and Discovery
This week, Twiggle, a start-up developing e-commerce search technologies, secured $12.5 million in funding from South African Internet firm Naspers. Yahoo Japan and State of Mind Ventures also participated in the funding. Twiggle was founded in 2013 by Dr Amir Konigsberg and Dr Adi Avidor, two former Google employees.
Twiggle claims to use artificial intelligence and machine learning to generate more-relevant product search results for e-commerce sites. According to Dr Konigsberg, “Twiggle is taking search to the next level. Our engine is able to actually understand queries, understand what people mean and intend when then they type a query.”
The start-up plans to shake up an environment dominated by Amazon by “building the future technologies of e-commerce.” There’s certainly some accuracy to Twiggle’s claims that there’s room for improvement. In many instances, searches on e-commerce sites fail to deliver useful information and often visitors are forced to shape their queries to be computer-friendly, rather than the site adjusting to natural human interactions.
Twiggle says it uses machine learning and “deep product knowledge” to understand search queries. This enables its search engine to understand intentions. One example the company provides is a search for “smartphone with best camera”. A query worded in such a manner should not return a listing consisting of both phones and cameras, but rather, suggest the smartphones with top-rated image sensors. Currently, searching with such natural language gives mixed results and shows no indication that the site understands that one term was meant as an attribute of another.
Twiggle says that it will enable e-commerce companies to redefine their shopper’s search experience which in turn will result in improved retention, revenue and profitability. Twiggle says it will roll out the technology to a number of “large e-commerce players” during the coming months and challenge Amazon’s dominance of product search engines.
Twiggle does not plan on becoming an Internet search engine and taking on Google, but rather is focused on improving online shopping experiences. Amazon’s own search, developed more than a decade ago, is used across Amazon’s sites. Twiggle is challenging a very dominant company, but no industry is immune to fresh approaches.
Even with this narrowed ambition, Twiggle faces a hugely challenging task. The history of the Internet is littered with search start-ups — including Clusty, Kartoo, eeggi and blinkx — that had a better idea or technology but failed to convert it into a business large enough to benefit from the economies of scale that Google and Amazon enjoy. Even Facebook has not realised the full ambition for its Graph search, announced in March 2013, which — like Twiggle — was based on understanding human intention.