Monday, June 4, 2012
Interview with Chris Treadaway, Polygraph
A couple of weeks ago, Austin-based Polygraph (www.polygraph.com) launched its Facebook analytics product into the market. Chris Treadaway--a technology industry veteran who last worked as a executive at Microsoft and who is the company's CEO-- told us how the company came into being, and spoke to us about how the firm is looking to help mine the vast amounts of data now being added to Facebook and other social sites to help companies understand their customers.
Can you explain how Polygraph came about?
Chris Treadaway: At a high level, it was about a year or year and a half ago we were approached by Demand Media, which asked us to clean the profanity off of YouTube. They have one of the largest channels on YouTube, called Expert Village, and they had this problem where people were trashing each other on their YouTube videos. We went in, and agreed to help them clean out that profability. In the course of doing that, we realized how big of a problem it was. We had to go through 185,000 videos to find that profanity, and delete it. What we thought was a trivial problem, ended up being a very significant, big data problem. In the course of doing that work for them, we developed a data mining engine for social media, which would look at the comments and then act upon them.
We realized that we were looking at a business problem. We realized that, gosh, there are now these large, self-contributed database from consumers in the world, and you can look at Facebook comments, Twitter comments, LinkedIn, and all that, where people are willingly saying all sort of things on their pages, profiles, adding their observations on politics, news, purchase intent, or whatever. We looked that this and realized that there was a big business opportunity here, where you could mine the largest consumer database the world has ever seen. Polygraph was born out of that, with the core technology we'd developed solving that profanity problem for Demand Media. We used that core technology and repackaged it for agencies and brands to understand what consumers were saying about them, and allowing them to see what their customers and doing and how they are reacting on popular social networks, with Facebook being first.
What's your background and how did you get into this?
Chris Treadaway: I was at Microsoft for three and a half years, and the last year there, I was responsible there for the web and social media strategy across the company, giving the Windows, Office, and Bing team, and the advertising folks something coherent and consistent. That was kind of an impossible task, actually. A year and a half ago I moved back down to Austin, after spending three and a half years in Seattle.
What types of customers would fine your product the most useful?
Chris Treadaway: We've been working a number of agencies around both the East and West Coast, as well as in Texas. We initially started talking at SXSW, and we got lots of good, and unexpected feedback from agencies. We thought the market would be more well defined and crowded than it actually is. When i started talking with agencies at South By, we got the universal feedback that nobody had quantified Facebook as well as we had or had made sense of what was happening there. Most of the companies working on analytics in social have been working on Twitter, and very few do Facebook really well. So, we held off on our launch to go deeper with agencies. We spent April and par tof May sending reports to agencies to bring to major companies, such as Fortune 500, household names, and brands, and so on. We got lots of good feedback and tons of great responses, from movies, major airlines, major retailers, and others interested in that consumer data. We're now in the process of getting those people converted over to paying customers.
You mention there were many companies you found working on this kind of information for Twitter, but not for Facebook. Why do you think companies hadn't been looking at this area?
Chris Treadaway: People aren't looking at it from a data mining perspective. Definitely there are folks using and trying to do some sort of Facebook analytics, but all they're doing is taking Facebook's Insights and repackaging it, the vast majority of the time. I don't know how familiar you are with Facebook Insights, but the biggest complain we heard is that Facebook provides metrics that people can't understand. If you look at the information that gets the most complaints, and metrics people talk about, is that people don't really know what they mean. What we did, is we do data mining out of Facebook, and collect information on every individual social interaction that takes place, and deconstruct that into our own set of metrics, which can help community managers, executives, consultants, and agencies who really care about their performance on Facebook, which they can understand. Data mining out of Facebook is actually really, really challenging. Twitter data is pretty easily available, the APIs are well built, and are soft of designed for people to be able to pull out data and manipulate it, and do things with that. But, Facebook is more challenging. There are more hoops to jump through, more difficulties. Now that we've figured that out, there are lots of interesting things you can do with that data, once you're able to reliably grab and manipulate it. There's a big technical barrier to entry for people to do Facebook well.
What drove your decision to move back to Austin after your time at Microsoft?
Chris Treadaway: My wife and I both consider it home. My wife is ex-Trilogy, and moved here in 1998. I moved here in 1997, and was one of the co-founders of Stratfor. We both really love it here in Austin, and are Southern by nature, and prefer the warmth to the cold, and all that. It was a no brainer. When I moved to Seattle, I always planned on coming back, but didn't want to just do a nine month tour up there and have to explain that to other people like investors, or if I were going to look for a job. I wanted to spend meaningful time in Seattle, but also wanted to expand my personal network of people I knew and had worked with, and fought battles with. But, the entire time I always intended to come back.
What's been the most unexpected thing you've learned from that Facebook data?
Chris Treadaway: In the course of our testing and initial work, we've done a few thousands reports on Facebook on different pages. I think there are a few takeaways. One, is companies are dumping a ton of money into Facebook ads and social software. There are companies here in town like Spreadfast and other competitors like that. But, I think people have been doing that because it's been conventional wisdom, that companies should be doing things in social. But, when you look at our report, and the things we've seen across different properties, we see that it doesn't make as much sense for all of them to dump lots of money into social marketing as they are. The question is, if they're dumping massive amounts into Facebook ads, are they really getting their money's worth? To me, it's actually been a little surprising, because the answer has come up as no as often as it has been coming up yes. I think we're going to see lots of people respond to that with an increased attention towards metrics and analysis, and important key performance indicators. As things mature, it's a natural evolution.
One of the other things we've seen that has been a surprise, is that brands don't post anything on weekends on Facebook. The weekends are less crowded, and I think it's been one of the big surprises to me that how much conventional wisdom has led people to think they need to post at 10am Monday through Friday only. It's left a massive opportunity to have less competition for people's attention on the weekend. That's been one of the consistent things across thousands of reports we've run, which is brands are ignoring the weekend.