New CEO comments on CompetitionXxxxxx, you're right, we don't talk about competition nearly enough. However, rest assured that I have as good a beat on the competition as anyone at Slyce. With our consumer apps team, we've rigorously tested every player in the market. We just did a spike last month (dev speak for a testing and research) and built out demo apps using IBM Watson's image service, Google Cloud Vision, Visenze, Wolfram ImageInstance, Meta Mind, and Cloudsight. Suffice it to say, there is real competition, all going after the same $30bn market. And there are some we're more nervous about than others. However, our approach is completely differentiated -- no one else has the human tagging and machine-learning one-two punch that powers Slyce.
Our human data center is basically a facility where we custom curate and tag reference image data for all our clients (and much more). Other companies, including all of the above, do not have this capability. In order to use their services, you need to BYOD -- Bring Your Own Data. That is a huge problem, as virtually no retailer has the kind of tagged reference image data that we have curated in house. And if they do have any such data, it's not nearly enough of it to power automated systems. Our solution, by contrast, can get a retailer up and running on Day 1 without any historical data and, in the extreme, a fully manual tagging flow. But even in such a scenario (where none of our automated systems can be utilized), Slyce still produces an incredible experience — identifying images quickly and extremely accurately. And very quickly, we've built up enough reference data that feeds into a neural network and begins firing off automated responses.
Put another way, try any other competitor's service and you'll quickly see how inferior it is to what Slyce offers. Cortexica, which has managed to get a couple retailers we didn't land (Macy's) is a clunky experience with more manual steps and worse results. Camfind's solution (white labeled as Cloudsight) is in the Kohl's app, and it's embarrassingly bad. Amazon Flow still fails to identify most products. And even the so-called unicorns in the space -- Blippar has a $1bn valuation, and their image recognition capabilities are laughably bad -- can't match Slyce accuracy. In terms of Visenze, I'm curious if you've tried any of the retailers they're currently deployed on?
Next go try JC Penney, Home Depot, Neiman Marcus ... It's no contest.
Last point: the human tagging approach does scale; the problem is that Slyce's prior management chose to scale it prematurely, to create excessive capacity before we had the customer load to support it. Now that we have corrected much of that, I'm confident the system and staffing we have is in a good place. And again, I'm confident the system will scale. We parsed over 250 million images at SnipSnap, and the tagging expenses were fairly insignificant in our overall cost structure.
I wouldn't have accepted the position to lead Slyce -- or sell SnipSnap to Slyce -- if I wasn't 100% sold that we had the best solution in the market. Human-trained computer vision is why we've seen so much traction and emerged as the market leader. And it'll be a big part of the way we talk about and position Slyce going forward. It's how we win.
Thanks for joining the call, asking the follow-up questions, and reaching out.
Best,
-Ted
--
Ted Mann
President, Slyce