Bill, it’s my pleasure. Social CV is one of the few products in recruitment that has an immediate ‘wow’ factor. Certainly that was the vibe we all got at TruLondon 3 when you guys demo’ed the product. For the benefit of the readers of Wise Man Say’s blog, put Social CV into a nutshell for us.
We’ve talked about this as a ‘Social Jigsaw’. You can discuss it as ‘Google for People’. What we are doing is applying semantic tools to create an index for individuals. This is something that a traditional search engine doesn’t do. Here’s how we thought about it. There are about 1 billion profiles out there on the social web. That is the opportunity. The challenge is that the data is unstructured. People say different things, about themselves, on different platforms. Some people will say on Twitter that they are the father of 3 kids, but not necessarily mention that they are the CTO of the Fortune 500 company. They may say on Facebook that they really like surfing, and maybe mention that the work in a technical role. But not all of that together in the one place.

What our company can do is figure out which profiles and which updates belong to the same person and tie them all together in a single place. That’s not easy to do. This is where semantic search comes in.

….right…..
(Wise Man Say’s policy is to agree with what’s being said when things get complicated)

Try looking at it this way. A very famous challenge in search is the ‘New York Times Challenge’. There are millions of profiles out there that mention the ‘New York Times’. But there are only a couple thousand people that work for the New York Times. Our challenge is to come up with the semantic technology that can interpret when a person is talking about New York Times the employer, rather than ‘New York Times’ as something he read. You can’t do this on key word search alone, you have to search keywords…..

…in context?
Exactly. We have to teach the computer to search on keywords in their proper context – in effect, to teach the computer to read. We have an index hundreds of thousands of words that we think people write in social profiles and use algorithyms to determine whether they are talking about ‘New’, ‘York’, ‘New York’ the city, the ‘Times’ or whether they are actually talking about working for the ‘New York Times’

Clever. What about that bit about pulling it all together?
Right. So another thing we do is have the technology which tracks inbound and outbound urls. For example Twitter, no one mentions University. On MySpace, almost everyone does mention University. So we have technology that spiders off from one profile to the next, following the urls that are embedded into the bio, hunting for information about that one person. We go through 30 different social media sites in order to build a full social profile of the individual we are gathering information on. And we are building 400,000 profiles per day.

4,000? Did I get that right?
400,000

400,000 records. Every day. That’s massive – what is the tech behind this Bill – all this sounds like some serious Kung Fu
Yes. We have gone through 5 or 6 database architectures throughout the lifecycle of the product…..

(there follows a highly interesting 15 minute diversion on Mongo DB, Cloud Computing and the merits of Key-Value vs Relational databases. It’s the judgment of the editor that it’s way over the heads of the readers of Wise Man Say’s blog)

Any more amazing things to tell us?
Yes (he means it). Something we are now looking at is….you know people have Avatars on their social profiles?

..yes..
Well, we are now experimenting with facial recognition software, so that that can help us confirm the profile is who we think he is. We can’t rely on a name because firstly there are many duplicates which might generate false positives, and secondly, many people use pseudonyms rather than first name second name. ‘Alphadog29’ on one site might be DarkAngel100 on another. We can tie profiles together that would be impossible to tie together if you use Google or something similar, but we are always working on doing it better – hence the face recognition software we are experimenting with.

Hang on Bill, I’ve got to rewind this back. It must have taken years to get to this stage. What was the genesis of Social CV?
We are a semantic search company that started off with a partner product with Google that wasn’t commercial successful. I mean, it wasn’t a disaster, but it never made the sort of profit that we were hoping for. But we put a lot of R & D into it. What came out of that was the technology that came to form the basis of the things were are doing with Social CV, and some of the other products we have out there, like TwitJobSearch. So, we’ve made use of the technology we built, re-purposed it to focus on the recruitment market and we’re really excited with what we’ve done so far. We’re not finished at this stage yet….

…actually, where are you guys at the moment? When I go onto the site, I see it’s a BETA…
We describe it as a private BETA at the moment. We’ve got 100 recruiters, all over the world, testing the product and giving us feedback for us. Some of the feedback we’re getting is that people who want to get it integrated into ATS systems, so we’re working with some of the biggest ATS/CRM providers out there on how to integrate the product.

OK. Let’s have a look

(Bill proceeds to demo social CV. We run searches. We see distributed social profile data pulled into a single page profile. We see the semantic extraction of profile data from across 30 leading social media sites. We see geo-coded profiles, with maps on where they have checked in. We see a Talent Ranking system ranks profiles according to data quality, and quality of the data* We see watch lists and talent pools. If there was any doubt before, there is none now. This is a potential game changer)

Impressive Bill, impressive. Now, how are you going to make your money?
(laughs)
It’s unclear. We haven’t launched yet. I think we’ll end up working with resellers. We are in talks with a number of companies – ATS providers, major CRM providers, Job Boards. We’ve built this thing and we’re figuring out what to do with it. No one has really built anything like this before, so we’ll probably end up partnership with organisations that want to solve the problems that we can solve, and leverage their marketplace and expertise to get our product out as quickly as possible.

How long until world conquest?
Oh I don’t know. I think we are having the right conversations to take the product forward. We have built a product that can scale and have partners that can scale and can scale. We have to make a couple of tweaks, but we’re comfortable where we are.

And how long before formal launch?
We don’t know. We’re working on a few things now. But we’re talking weeks, not months. Though that could be 8 weeks. (laughs)

Bill, I think that covers it.
No problem Wise Man Say, thanks for coming in.

Thanks to Bill Fischer, Co-Founder & Director at Social CV. Social CV, is Google for People. Find out more on www.thesocialcv.com and follow Bill on Twitter @williamfischer