Archive for Analytics

Pandora Radio and Predictive Analytics

The fall lecture series on disruptive innovation wrapped up at Linda Hall last night with Nolan Gasser, the chief architect and musicologist behind Pandora’s Music Genome Project.

The talk was good, as all of them have been, though maybe a bit less organized around a central theme. There was quite a bit of dabbling in advanced music theory (that left many attendees scratching their head) as well as some very accessible music playing, from Coltrane to Zeppelin to Pachelbel. Clearly, Glasser possesses a unique genius.*[1]

Amidst the offshoot thoughts spurred by the talk were some audience questions about predictive analytics. Pandora is, at root, a big pop culture analytics experiment. Each track is scored on a scale for 100s of attributes. There are millions (billions?) of tracks. Those are a lot of data.[2]

Three thoughts on these data and how they are analyzed:

1. The human element has not been automated.

Each Pandora track is coded by a human being. I found this really surprising. And sort of gratifying. I remain suspicious of (though optimistic about) highly automated analytics, especially social media monitoring and natural language processing programs. That the Pandora model, especially given its mathematical and theoretical basis, has not completely handed the data inputs over to an algorithm made me feel better about my suspicions.

2. There are critical differences between things that can’t be predicted…

Someone asked if the model could be adapted to predict future radio hits. Gasser was doubtful. It’s a great idea, and maybe it is possible, but it’s important to be realistic about the limits of your data, your algorithm, and your knowledge of the human framework that underlies all our decision-making.

3. …and those which can be predicted.

The beauty of Pandora radio, of course, is that it predicts what songs you might like and adapts its model based on your inputs. Pandora works not simply because it has a massive, static database but because that database is modified with individual inputs.

Another place where Pandora succeeds is in measuring its audience analytics. Gasser mentioned briefly at the end that Pandora has really redefined online marketing with its ability to deliver targeted advertising. I would love to hear more on this topic.

(Aside from analytics, this talk opened up a whole bunch of questions for me on aesthetics. Gasser works from the premise that we are biologically hard-wired to like music, and that there are mathematical properties that make us like some music more. Yet he is unwilling to reduce taste to an equation. I asked him a question about this, not very well, but there was one piece of his response that I particularly liked. Basically: “We’re all hard-wired to laugh, too, but each person’s laugh is different.” I thought this was an interesting take on the matter of taste. But I’d love to ask some follow-up questions on this topic as well.)

[1] Maybe all genius is unique…a topic for another day. Anyway, he played lots of music on a keyboard, along with a few clips from his laptop. But the versatility and the music were both enjoyable.

[2] Sorry, I take small pleasure in data being a plural.

Comments (1)

Video Analytics Are the New Future

In a post a couple weeks ago, I referenced the SKC Tech Summit as an example of a conference outside my field by trade but inside my field of interest. Though it wasn’t an explicit theme, the idea of a “video revolution” certainly loomed large. Topics like telemedicine and virtual education were highly visible throughout the event.

Market research and customer insight was not—although companies like TASKE emphasized an analytical approach to call center software that included rudimentary service satisfaction surveys. Not market research, per se, but a good indication of how integrated insights are adopted by non-traditional market research departments.

Video has been a staple of focus groups for years, of course, and focus group facilities have mostly used advancing technology to help present the work to those who can not be there in person rather than change the way the work is done. QualVu does a great job innovating around video—exploring new techniques and making it more accessible, more diverse, and more entertaining on the report-out side. But as the name suggests, it continues to be primarily qualitative.

What struck me at the SKC conference was the potential for qual-quant hybrid techniques through the usage of video analytics. There are some facial recognition and eye tracking software programs out there that edge towards this idea, but I’m talking about teaching computers to analyze visual motion data, rather than just text/voice or still images.

Considering that text analytics and still image analysis are still immature, I expect video analytics to be at least two technology leaps away—but it’s not too early to get in position to lead in this area.

Take large scale in-home ethnography as an example. QualVu does in-home “ethnography” on a small scale. I put ethnography in quotes, because the respondent is pretty interactive with the camera—it’s more about them showing than you observing. And small scale because the analysis is intensive enough (and qualitative enough) that traditional qualitative sample sizes are both effective and practical.

But imagine setting up three video cameras in 1,000 kitchens for a week. Or two days a month in different households for a tracking study. Quant sample size. Multiple cameras to capture different angles. The client could be a CPG company, a grocery store, a cookware company—you name it. Sophisticated visual analytics could potentially tell you:

  • What’s your morning coffee routine?
  • How much time do the kids spend in the kitchen?
  • How many of those fancy knives do you actually use?
  • Are you drinking wine while cooking dinner?
  • Are Ziploc bags a replacement for Saran Wrap or Tupperware?
  • How many trips to the pantry are required for each meal?

Yes, you can ask things like “Do you clean as you go or wait until the meal is finished?” on a survey, but that sort of misses the point of ethnography. You don’t always know the right question to ask. It’s less about testing hypotheses than uncovering latent needs and motivations. Pervasive video makes observation of people in their natural environments possible in a way that it has never been before.

“Pervasive video” immediately suggests Big Brother (uh, maybe not this one) and privacy concerns. Big Brother or no, privacy has changed. I won’t even get into the respondent confidentiality aspect that consumes much of the traditional market research world. The fact is, individuals have shown an increasing willingness to broadcast their lives that shows no sign of abating, regardless of what Chuck D thinks.

The technology, on the other hand, in not there yet. Talking to people at Cisco, video analytics is on the radar, but not yet a high priority. The more pressing analytic need is to search for text in video, search-engine style. This technology—think automated transcription—is available but still a bit clunky. And of course, creating the broadband and hardware infrastructure to easily facilitate web-based video transmission is a project that is also just underway.

But as we know, technology moves quickly, and I have no doubt video analytics are coming. A five- to ten-year time horizon would not surprise me.

And once we have sophisticated visual analytics, the Big Data explosion will reach another order of magnitude. Potential applications, that may be easier to imagine than voluntary in-home surveillance, include:

In-store video cameras that can provide comprehensive insight into the shopping process, and combined with web analytics, give great insight into online/offline integration for retailers

  • Evaluating B2B performance and client relationships as videoconferencing skyrockets
  • Mobile streaming/life casting—which is already happening but will become more comprehensive as video becomes easier and cheaper to upload

And for those outside the Kansas City area, know that Google is launching a Fiber network here in 2012 that promises to make uploading (and downloading) video much, much faster.

What other applications do you see for insights and analytics as video becomes more prevalent? How far off do you think we are from video analytics?

Leave a Comment