Comments for BuzzLogic: Interview with Co-Founder Mitch Ratcliffe

BuzzLogic: Interview with Co-Founder Mitch Ratcliffe

2006-10-09T07:42:03Z

BuzzLogic was one of the hits of the recent DEMOfall 06 event. As Alex Iskold noted in his original post about the company, its software helps marketers track social influence among blogs and other web sites. We were as intrigued by BuzzLogic as everyone else, so Alex Iskold followed up with Mitch Ratcliffe to find out more. In this interview we discover that Mitch co-founded the company in early 2004 "in order to find automated ways of analyzing influence." So it's been a long journey to launch - read on to find out how it got to this point and what we can expect in the future from BuzzLogic.

R/WW: What is your company about?

We're seeking influencers in conversational markets, giving them more power and marketers better insight into how to deploy their limited resources to deal with the explosion of incredibly specific feedback that companies, organizations and political campaigns are receiving through social media. The tools we're building will be applied to many of the challenges we talk about with social media, such as valuing individual contributions to the conversation, rather than just traffic - and how to interact through a dynamic personal form of communication in order to get the results you aim for.

R/WW: Why did you start this company?

It actually began with a question about how to deal with a blogger, from a friend who is the CEO of a public company. I found myself drawing maps to explain not just the connectedness of a blogger to others in the market, but the way that a blog's influence varies from topic to topic. It was a short step from that to deciding, with Todd Parsons, to start the company in early 2004 - in order to find automated ways of analyzing influence.

]]> Basically, I was doing something with those maps that is analogous to building a spreadsheet with pen and paper. There had to be a technological solution to gathering the data and a methodology for processing it into meaningful insight. We started by drawing maps with a project called MyDensity, which let bloggers display a fairly rudimentary social map around their blogs. It got some traction, but we ran through our crawling/hosting capacity without coming up with a business model to support it. That quickly evolved into a business built on providing much more detailed data to paying customers, since the backend processing was awfully expensive. We've always been focused on actually building a business, which is why we brought on a team and CEO Rob Crumpler, all who worked for next to nothing while we made progress toward the launch.

Now, the company has built an infrastructure that is both more powerful and much more efficient - so we are able to launch our first product, which is a hosted influence monitoring and tracking service that combines our analysis with tools for interacting with influencers, so that marketers can measure the results.

R/WW: What are you major accomplishments thus far?

Well, we launched. The backend is built on a solid index and patent-pending analytics that produce really useful information. The initial feedback from our launch at DEMO was very positive, especially with regard to how we display information and the UI that customers use to manage their interactions in conversational markets. That was particularly satisfying for everyone at BuzzLogic, because it would be easy to ship something that only managed to add noise rather than increasing signal in these incredibly active social networks.

R/WW: What are your major challenges?

Handling the growing amounts of data flying back and forth over blogs, mainstream media and all the others sites people talk about, particularly the false information generated by splogs. Our engineers have developed some really clever approaches to identifying spam, so our index is very clean. But, there's a paradoxical problem, too, in that sometimes what appears to be noise can be orchestrated to have an impact on how people find and come to perceive information, so we have to keep the spam in context to provide customers some warning if splogs begin to reshape sentiment about their product or brand.

Then there is the problem of scaling the index for performance and growing customer demand, the usual challenges for a Web service. The beauty of these challenges is that every time we overcome one, a new market for our data appears.

R/WW: What are you going to build in the next 12 months?

Lots more cool stuff, especially some tools for bloggers to use to understand and act on the market around them - and to tap into the value they are creating. Our marketing and PR tools will be substantially enhanced, drawing in more data from specialized sources that shape conversations, as well.

R/WW: What is the most important thing for a start up to be successful?

Hydration. You need to keep the fluids coming to everyone on the team. After that, it's all about trusting people to do their best and tolerating the cost of people learning, because the sum of those experiences is greater insight into the thing you want to build.

R/WW: What web sites / blogs do you use / read often?

Blogs:
Fred Wilson's A VC
Alex Barnett
Another World is Here
Bubblegeneration
Confused of Calcutta
Doc Searls
Ethan Zuckerman
Brad Feld
if:book
Greater Democracy
Battelle's SearchBlog
John Robb
Mashable
Read/Write Web
ResearchBuzz
Reverse Cowgirl's Blog
Seth Goldstein's Transparent Bundles
Nick Carr
Scoble
ZeFrank

Sites:
Baseball Prospectus
CNET
Economist
Nature
New York Times
Washington Post
Tacoma News Tribune
Techmeme
ZD Net

R/WW: Which web 2.0 things are noise and which are signals?

I'm discinclined to celebrate the Ajaxification or Rubification of existing functionality, so the announcement of a drag-and-drop feature in an Ajax office app is just noise - because that feature isn't going to differentiate the product from its desktop predecessors.

Signal is taking something new and turning it into something practical and surprising. I don't think a lot of people agree with the one I was surprised by at DEMO, but Cuts - which lets people create edit paths through media, including DVDs and television shows, and share those edit paths through a community - was a big signal for me, because it shows how people can route around dumb industries by passing instructions for using products created by intransigent industries, from person to person.

Comment from Nikola Markovic on 2006-10-09

2006-10-09T10:16:19Z

Finally, I found someone NOT reading the 'Glorious Techcrunch'! ;-)

BuzzLogic is indeed a novel concept, too bad that end-users can't have a taste of it without filling the detailed form, even a 'tour' would be good.

Comment from Emre Sokullu on 2006-10-09

2006-10-09T12:19:26Z

Interesting idea, a service that makes data-mining in blogosphere to allow companies/organizations/individuals to keep an eye on blogosphere.

By the way, another interesting company of DemoFall was LeapTag, I think. I was with their founder last week, it seems really promising, I describe them as "if stumbleupon is digg, LeapTag is reddit for link recommendation systems".

Comment from Mitch Ratcliffe on 2006-10-09

2006-10-09T16:37:55Z

Nikola‚ÄîWe're doing something for the blogger, too, but we have to start where the money is with a compute-intensive application like this and, so we've got to begin the sales process with access to the beta. Sorry about that--we'll have a "tour" up after the beta period, as well.

Emre‚ÄîThe system tracks influence *outside* the blogosphere, as well. We think it is important to see where stories are starting and where they are accelerating, which happens in a variety of media. And, if you can address blogs and media side-by-side in that context, you create a meaningful way to value advertising based on influence that could, we think, be tremendously beneficial to the blogosphere.

Comment from Ferodynamics on 2006-10-09

2006-10-09T17:56:13Z

In this case I think the word "influence" needs to be used carefully. Popularity and influence are not the same thing.

Who would use this? Politicians? Reading the blogs yourself seems smarter and faster to me. You can't use an algorithm to identify a person wise or foolish. And if your message is so important, why silently observe when you can spend your time influencing instead?

Comment from Mitch Ratcliffe on 2006-10-09

2006-10-10T01:05:45Z

Our influence algorithms examine a variety of factors, popularity being a lightly weighted one. Traffic doesn't necessarily mean influence, though our users will be able to adjust the tools to weight popularity more heavily if they like. Pols are certainly one group that might use our tools, but so will PR, marketing and product people who want to find the most efficient path to influence in a conversation.

With regard to it being "smarter and faster" to read blogs yourself (though we look beyond blogs), that may be a viable option if you have one topic you follow. But most people and all companies have to follow a lot of different topics with limited resources. They are already overwhelmed and seeking tools like this‚Äîthey've told us so. As I explained in the interview, you could do a spreadsheet by hand or do it using a computer faster so that you can apply the information. You could read blogs and attempt to intuit which places, whether blogs or MSM, are getting the most traction in the conversation, but a tool that like ours lets you apply your limited time most effectively.

The second aspect of our product is the ability to track your participation and the results of that participation, so we certainly are not suggesting anyone "silently observe." We're also hoping to make it possible to justify blogging in the corporation by showing the impact on the market statistically, as well as the value of interacting with bloggers compared to just aiming at the established media.

Finally, the current product is only our first application of the tools we've built. If you can measure influence on a specific topic for your blog, you may be able to get a higher fee for a click than a site that generates clicks based on brute force traffic. We think there are a lot of ways to use this data once you have collected and analyzed it....

Comment from Ana on 2006-10-10

2006-10-10T16:12:21Z

I'm wondering if your "proprietary algorithm" is some centrality measure or something else? I am doing my disseration on social media and am curios to compare different methods.

Comment from Mitch Ratcliffe on 2006-10-10

2006-10-10T16:28:36Z

I don't recall using the word "proprietary" in the interview, but I'm sure our marketing team does.... The way we will implement our service in an upcoming release, the algorithm is partially exposed to the user so they can adjust it to arrive at their own influence algorithm if they do not agree with our defintion‚Äîthis field is so developmental that we wanted to avoid locking down just one "influence" and being stuck with it.

The way we calculate influence uses a number of factors, though "centrality" as it is defined in network analysis is not exposed to the customer, because we have tried to stay in the realm of concepts a customer will be able to understand and manipulate if they want to. Instead, we apply the concept of centrality in distinguishing nodes in the network as influencers and amplifiers as we look at the spread of a particular keyword(s).

The customer is not studying network theory, rather they have some task they need to get done.

If you'd like to talk about using the system in your research, drop me a line at godsdog (@t> ratcliffe dot com.

Comment from Ferodynamics on 2006-10-10

2006-10-10T21:33:06Z

Sounds good to me. I'm very curious.

What do you all think of Google Alerts and Nielsen BuzzMetrics?

Comment from Mitch Ratcliffe on 2006-10-11

2006-10-11T07:07:51Z

Google Alerts and BuzzMetrics are different than BuzzLogic in several ways.

Google Alerts is basically a clipping service, which does a good job of finding stuff on active sites in a timely fashion. The problem with Google's index is also its weakness, it is all-encompassing and largely retrospective, often it takes weeks to get a new document in. It also shows a lot of redundant information, so you don't get any context about the information you are looking at, rather you have to find the meaning‚Äîand it provides no social analytics whatsoever. A great tool if you want to drown in information, something I use for keeping an eye out for mentions of topics but that I find I use very little because of the undifferentiated volume of data.

BuzzMetrics is a competitor, so I'd rather someone else contribute their opinion than be perceived as taking a shot at them. We're different.