1. "pay per action" is paying for a qualified lead, as opposed to a raw lead (which could be anything, e.g. even a wrong number).
2. Doing it with phone calls (as opposed to clicks on online ads, or emails etc), because that is how many businesses and their customers contact each other. e.g. gyms, vets, car repairs.
3. Doing it for local businesses.
The technical achievement of speech recognition is significant. I like his point that because the speech is interpreted relative to a specific business type, the domain is highly specialized, and so the task of recognizing the particular terms of that domain is much easier. I've been impressed by automatic speech recognition for telephone directory lookup - for me, it works 90% of the time; and that's out of the whole phone book, not restricted to a domain. I guess this is an example of "speech recognition" is still not good enough in general, but it's good enough in specific roles.
The next step after "pay per action" would be "pay per sale", which Google can do when you use both their adwords and their payment facility (I don't know how well that's working).
Anyone care to correct my misunderstandings, or elaborate on these points?
I was at TC50 and it's impossible to overstate how dominant this guy's presence was during his demo. No one could believe their eyes or ears that here was this amazing, large, profitable company doing something everyone wants and no one had ever heard of them. I would have offered them money on the spot.
I don't watch TC50 videos, so maybe it's just that I am unaware of the typical presentation quality there, or maybe it doesn't come through on video for me. His presentation was confident and well-rehearsed, and his startup was pretty cool, but I wasn't incredibly impressed by his presence.
I doubt the video would be impressive. It was the feeling of everyone in the room going through this process simultaneously:
I don't get it - neat - holy crap - is this real? - Holy Crap!! This is real!?!?! - Wait, did he just say $20 million in revenues? - Why have I never heard of this before? - $20M?!?!
There's a feeling in the air when 500 people all have the same amazed reaction that doesn't come through video.
Particularly fun was watching Marissa Mayer's reaction when he started talking about Pay Per Useful Call.
Interesting article on a veterinary site in which one person accuses Yext of ripping her off while a few others tout pretty impressive ROI from using it.
"Dr. Randy Wiltshire reported in a VIN discussion that he’d used YextVets for six months, and during that time, was directed 67 new clients from the site, for which he paid Yext a total $2,345 in advertising fees. Those clients, in turn, spent $9,100 in his practice."
So if he paid $2,345 for 67 calls, Yext was charging him $35/call. That's a pretty nifty digit.
Well, they got in on the merits. It certainly wasn't like HP demoing a new product (like we saw at DEMO); it was a company coming out for the first time. It certainly qualifies as a startup in my mind. It may be a $20M company, but still, no one had ever heard of them before. The look of astonishment on the panelists' faces was awesome.
It may very well have been the greatest gambit to create a competitive bidding process for VC investment in history. Seriously. They presented at TC50, and in less than a month, had a VC firm who they had absolutely no prior relationship with wire $25 million to their account. (Sure, they had previously raised $3.5 million, which marked the IP as safe and the revenues as real, but still.)
I get that they are still a startup, but they are a funded startup. It was a great demo, but still it seems like TC50 should be restricted to companies who have yet to do a series A round.
Lots of companies had already raised funds - Clasemovil raised half a million, Clicker has 8 million, lots were self-funded by people with exits under their belt, etc. The only requirement was you had to do your first public release at TC50. The judging criteria were useful demo, attractive, professional, polished design, and useful business, and the first three favor those with funding.
i dont get it ... how do they help the businesses with the junk calls with transcribing and categorizing them? it seems to me that those junk-callers call anyway.
The purpose is to get customers for the businesses. The startup supplies phone calls to the local business (i.e. potential customers). The local business pays for the calls, but they don't have to pay for junk calls.
The purpose of recognizing the nature of the call is to determine whether they need to be paid for; not to protect the business from junk phone calls.
1. "pay per action" is paying for a qualified lead, as opposed to a raw lead (which could be anything, e.g. even a wrong number).
2. Doing it with phone calls (as opposed to clicks on online ads, or emails etc), because that is how many businesses and their customers contact each other. e.g. gyms, vets, car repairs.
3. Doing it for local businesses.
The technical achievement of speech recognition is significant. I like his point that because the speech is interpreted relative to a specific business type, the domain is highly specialized, and so the task of recognizing the particular terms of that domain is much easier. I've been impressed by automatic speech recognition for telephone directory lookup - for me, it works 90% of the time; and that's out of the whole phone book, not restricted to a domain. I guess this is an example of "speech recognition" is still not good enough in general, but it's good enough in specific roles.
The next step after "pay per action" would be "pay per sale", which Google can do when you use both their adwords and their payment facility (I don't know how well that's working).
Anyone care to correct my misunderstandings, or elaborate on these points?