A question that has come up is how do you ethnography for distributed populations (e.g., online communities) and massive data sets… and for that, I’ll talk about mine and my colleague Nic Ducheneaut‘s (around here, we’re referred to as the “Nics”, and no, we didn’t name ourselves that!) work in virtual worlds.

Virtual worlds: the opportunity

Virtual Worlds (VWs) provide a number of unique affordances and challenges for researchers:

• Unlike the physical world, the VW comes inherently instrumented with high-precision movement sensors, perfectly transcribed conversations, and instantaneous teleportation.
• On the other hand, VWs bring interesting challenges in terms of ambiguous multiple identities (i.e., players having multiple characters) and, more importantly, the sheer scale of data that can be collected at the click of a button. For example, in our original “PlayOn” study, we only collected 7 character variables from the online game World of Warcraft, but it took us about a year and a half (and about half a dozen publications, see some of them here and  here) before we completed our analyses.

The hidden variable for us? Time. 7 variables from 150,000 characters over 1 year at 15 minute intervals provides many ways of looking at changes over time.

Virtual worlds: the methods and what’s needed

Over the years, we’ve used a broad and customized mix of methods to study VWs. These include participant observations, structured in-world interviews, lab experiments, online surveys, server-side data mining, and more. More importantly, we’ve learned that a certain interdisciplinary agility is needed to make the most of research in this area, and that perhaps these collaborations point to a larger intersection of social science methods facilitated by these novel research spaces.

Consider for example the need to bounce back and forth between different levels of analytic scale. In our study of player guilds (i.e., player organizations in online games), we leveraged the PlayOn data to map out the range of guild sizes and types. This allowed us to then target representative members of representative guilds for structured interviews. Not only did the interview data help us understand the larger trends, but the high-level data also allowed us to provide good estimations of how generalizable our qualitative findings were — a perennial concern for sophisticated ethnographers.

Another example to consider is the ability to leverage real-time data monitoring to identify low incident events, such as leadership changes in guilds or guild fracturing. By leveraging the PlayOn architecture, one possibility is to set up alerts for these low incident events and alert ethnographic researchers. In this way, VWs make it possible to gather many data points on events that would otherwise be difficult to gather data for.

Our examples highlight the ways in which qualitative and quantitative research teams can complement each other, but the examples also hint at the need for technical expertise — in terms of extracting and processing large data sets from VWs — which both psychologists and ethnographers don’t often have. Thus, we believe that VWs and similar Web 2.0 spaces hint at an emerging mixed or “hybrid” methodology that depends on agile collaborations between quantitative researchers, qualitative researchers, and software engineers.

Why bother?

This is not just an academic enterprise. The ability to glean this data has many implications for designing and scaffolding online communities, learning new aspects of personality and social behavior in online worlds, and mapping digital personas to physical needs.

The ability to leverage this architecture for more tailored marketing is one commercial opportunity. In addition to inferring basic demographics, personality inferences may lead to more nuanced methods of targeted advertising. And the ability to infer demographics based on online interaction metrics helps fill in the gaps left from zip code segmentation alone — after all, not everyone who lives in your neighborhood (or in your home!) is exactly like you…

Be sure to check out our current PlayOn blog for our findings and more. You can also see entries from our earlier PlayOn work here. And finally, do check out this video where we describe some of the above…

As reported in my previous post, we collected a random sample of public tweets from Twitter’s Spritzer feed over a 7-week period, yielding about 74 million tweets. From these tweets, we identified 8.24 million of them as retweets. That is, 11.1% of the 74 million tweets are retweets.

Next, we searched for those tweets and retweets that contain at least one URL. We found that 21.1% of tweets and 28.4% of retweets include URLs, suggesting that a tweet with URLs is more likely to get retweeted.

We further investigated whether the retweetability of a tweet has anything to do with the type of website it refers to. Since most of the URLs included in tweets are shortened URLs, we first expanded the abbreviated URLs into their original URLs, and then extracted the domain names from the original URLs. [For example, given an abbreviated URL http://bit.ly/c1htE cited by a tweet, we first unshortened it to http://en.wikipedia.org/wiki/URL_shortening, and then extracted the domain name of en.wikipedia.org.] The URL domains are indicative of the type of content sources visited and shared by Twitter users.

Analyzing the 74 million tweets, we identified the 20 most popular URL domains referred to in our tweets and the number of tweets containing each URL domain:

On the other hand, the following table shows the 20 most popular URL domains cited in our 8.24 million retweets and the number of retweets containing each URL domain:

As can be seen, these two lists of URL domains do not match each other exactly. For example, formspring.me appears only in the first list, while mashable.com appears only in the second list. That is, the fact that a website is frequently cited in the tweets does not guarantee that it is also frequently referred to in the reweets, and vice versa.

For each URL domain, we computed a retweet rate by dividing the number of retweets containing the domain, by the number of tweets containing the domain. We then normalized the rate so that a value of 1.0 represents the average retweet rate of 11.1%. [For example, for twitpic.com, the retweet rate of 1.47 was calculated as (129,692/793,680)*(74/8.24).] A URL domain with a retweet rate higher than 1.0 indicates that, compared to the average case, the tweets containing this domain have a higher chance of getting retweeted.

The following table shows the retweet rates for the 10 most popular URL domains cited in our tweets:

As can be seen from the above table, the retweet rates vary greatly depending on the URL domains. For example, formspring.me, which is the 5th most popular domain, has a retweet rate of 0.05, suggesting that tweets containing that domain are very unlikely to be retweeted. On the other hand, the retweet rate of twitlonger.com is 6.07, suggesting that tweets containing that domain have high retweetability.

In the following plot, we show the retweet rates of the 50 most popular URL domains. The X-axis is the popularity rank of URL domains based on how many tweets contain each domain. The Y-axis represents the retweet rates of domains as computed above.

Overall, we see that not all popular URL domains in tweets are popular in retweets. The domain of the URLs also matters.

You can learn more from our paper about this work.

Traditional design rules don’t apply

What is now scarce in our busy world is reader attention, not “column inches” of news print. So digital media is breaking some of the “old” rules and assumptions for presenting news, such as:

• Only main subjects (topical channels) need be presented to satisfy readers.
• Only a few stories need to be presented.
• It’s too difficult, expensive, and manually cumbersome to organize thousands of stories.

The new design principles

With information abundance, news consumers want systems that help them to manage their reading attention. It’s not practical to turn too many pages or see them all at once. So what are the new design principles for publishers to compete in the Age of Abundance — when a digital newspaper can contain thousands of pages on hundreds of topics?

1. Prioritize channels to focus information. Each of us can follow only a few subject areas, yet these areas are not the same across every person. Prioritization helps information consumers find channels in their primary areas of interest — mainstream and specialized.

2. Bring in the experts; organize. It is easier to find, catch up, keep up, discover, navigate across, and learn information if it is well organized. Expert curators provide the best organizations, particularly if tunable with feedback.

3. Help users forage efficiently. Based on the principles of information foraging, readers shift their attention based on whether they have a minute or two to browse/skim the news, or the mindspace to read in detail. Efficient scanning can be enabled by providing “top” or “hottest” stories they can skim across their channels and within each level of topics in a channel. Efficient navigation can be provided by enabling readers to drill down on interesting topics to get the next level of topics and more articles.

4. Provide the sideways perspective. Like a tree with its branches of stories, each channel has its topics and subtopics, and collectively all the trees form a forest of organized information. If we are browsing topics in one channel, how can we discover related information in another one?  Exploring related topics is more than just seeing “related stories”; it is about jumping to different points of view and different specializations. Ideally, connections between topics of interest should be discovered automatically and presented to readers.

Thriving with information abundance

Abundance is here to stay. No matter how the economics sort out for publishers’ subscription models, paywalls, content farms, aggregation, and advertising, consumers will not be satisfied with artificial scarcity. As Steve Rosenbaum argues in his analysis of media consumption, publishers who limit readers to their own narrow reporting are losing audiences.

There’s an opportunity to create a new generation of news/information delivery systems — optimized for information abundance – on all kinds of digital devices.

Try out the Kiffets system, where we have been experimenting with these principles. The tool is a prototype, intended to be refined with publishers, but you’ll get the idea; this video also demonstrates how a personalized news system can help news consumers manage their attention over abundant news.

Don’t get me wrong, I think aggregators like Flipboard offer a great way to read content from your social information streams. But is it the best way to get your news?

Or let me put it this way: are you doing yourself a disservice when you only read news that comes to your attention through your friends?

Who needs friends when…

On one hand, “friends” serve as a convenient filter for interesting content. I certainly enjoy seeing the links and likes shared via my follows and friends on Facebook and Twitter.

But I didn’t become friends with people because they were experts in my interest areas. Frankly my friends’ interests don’t necessarily overlap with my own,  and the cumulative interests of my friends doesn’t exactly cover all of my interests.

The ideal news reader:

• would allow me to specify all my interests,
• and then intelligently present me quality information in my interest areas,
• with just enough transparency for me to choose what I wanted to read, and why.

It would also organize the information along multiple dimensions to give me paths to explore further, to discover new things, and to help me answer questions without forcing me to search in Google or Wikipedia.

I can’t get that kind of content or organization from just my friends and the people I follow.

At PARC, we’ve been thinking about the question, “What news should we read in a world where we are overloaded or can’t get what we need?” (okay, that’s not really the question form but you get my point) — and we’re working on some research projects related to this. One project that we’ve only recently made available in a web demo is Kiffets, which you can read more about here.

I’m not claiming it’s the perfect news reader (yet), but its focus on long tail news curation complements the news I get through my social networks.  Read more about how Kiffets helps you tune your news priorities to get what you need.

What are your requests for the perfect news reader?

Louis Gray points out that “With tools like Twitter and other networks making it ever easier to hit the publish button, our ability to screen, filter and decide what information is good for us is going to be increasingly tested.” So curator-bloggers like Robert Scoble are trying to “find the best interpretations of the news from [their] favorite sources.” Rishad Tobaccowala notes:”It is not content that is rare. It is not compelling content that is rare. It is time that is rare” — and then asks, “Who can curate, combine and help us discover this content so we can make the most of our time? Who can get us things at the right time not just in real time?”

Enter curation

Curation is one way to deal with this problem. In many ways, it already happens everywhere… Individuals curate by sharing links in Twitter and Facebook.  Organizations curate for their members through newsletters. Crowds curate if you aggregate individual tagging and attention data.

But sharing is not necessarily curating.

Curation at its best produces a structured collection that gives readers more value than the sum of its parts. It requires domain knowledge and strategic thinking to organize topics with a purpose and point of view for the curated collection.

The future of news

News companies already do this for readers with their own content. Newspapers and magazines rely on editors to classify, link, and lay out stories. These editor-curators provide consistency in tone, perspective, and quality.

What’s missing: an effective, scalable way to do this across the Web. As the Recovering Journalist asserts, media publishers are “trapped in their walled gardens, putting together their daily reports only from the sources they pay for: their own reporters, maybe some wire and syndicated copy and photos, and that’s about it.”

This is a huge opportunity. Applying their curation strengths to the Web, editors can package syndicated and user-generated content to add value to their original content. Wouldn’t it be great to have alongside every news story an index with links to analytical articles from different perspectives, the backstory, reference articles, reviews, and opinions?

But it’s too expensive for news companies to do this manually at scale; they need an automatic approach with the precision of human editors.

Enter social indexing

PARC has been working on a research project for the past few years with this concept in mind. You can read about the backstory of this work here. We just launched an open beta of the demonstration site for personal curation and news aggregation powered by the Kiffets Social Indexing Engine. You can try out the beta here or watch this video overview.

A Social Index is a topically organized information collection created by human curators. The curators specify good topics, sources, and examples to create a point of view for the collection (see my Netbooks index or Mark Stefik’s Future of Journalism index as examples.) Kiffets learns the curator’s intent from the examples, creating models for classification that are more sophisticated than keyword- or entity-based patterns. Kiffets then uses the models to classify new content from the sources as well as recommend new sources for the curator.

Kiffets’ understanding of topics evolves over time as curators give feedback. Furthermore, the system automatically finds relationships between topics, articles, and indexes. By making these connections visible, Kiffets encourages readers to discover and explore related content when they are reading articles. This increased engagement creates value both for the reader and the publisher.

Kiffets is not unique in combining human input with algorithms. What is unique, however, is Kiffets’ ability to infer the perspective of human editors so the system can intelligently classify new content according to that point of view. Curators retain control to remove articles that do not fit, but Kiffets does all the heavy lifting.

Technology powers the future of news

The future of news is still being defined. But technology can help news companies adapt to changing times.

Technology already affects how news is collected, how it is curated, and how it is delivered. While the future of news is certainly about its commercial evolution, it is also about the powerful and important role that news organizations play in how society understands itself and the world.

Our Social Indexing Engine is a first step and we invite you to join us in the journey. Contact me, and read more about this technology, our views on curation, and the future of news at the Kiffets Voice blog.

Behind the research scenes

This project (as with many things at PARC!) has roots in a trajectory of evolving expertise — spanning early collaborative filtering and later information visualization and sensemaking systems, to social computing today.

After 9/11, PARC developed and deployed information systems for U.S. government intelligence analysts — whom we dubbed “the jet pilots of sensemaking” because their information-seeking and collaboration challenges were so intense. One output of this research is my colleague Peter Pirolli‘s book on Information Foraging theory, which has been referred to as “the ultimate source“. In our research, we developed a set of “cognitive amplifiers” for analysts based on artificial intelligence (AI) and collaboration technology.

Intelligence analysts’ situations then are not that different from people’s information needs now: too much too fast or too little too late. So insights from our analyst research have guided development of the Kiffets system, which personalizes news along people’s content needs and passions. As in our systems for intelligence analysts,  the AI and collaboration technology serves as a cognitive amplifier that enables scaling the sheer amount of  information that needs to be collected, filtered, and organized.

Within the intelligence community, some analysts act as curators or guides in sharing their expertise. These individuals help others to catch up and keep up in their areas of expertise. Now with open information on the Web, curation is emerging as a growth opportunity for information consumers and publishers.

The nuances of curation

In a recent graphical analysis of online media consumption, the conclusion was that winning media sites “all show a mix of large collections of content, mixing high quality created content with contributed, and curated content”.

The success of some curating news sites has generated buzz — although different people mean different things by “curation”. For example, the Nieman Lab pointed out important differences between AGGREGATORS (collectors), INDEXERS (auto-clusterers), and CURATORS (manual selection, organization, and commentary).

But, but… these distinctions, though useful, represent trade-offs: at one extreme, minimal human labor; at the other, time-consuming, unscalable human effort. Guess which approach is most expensive for publishers. Now guess which approach is most convenient and cost-effective. They won’t be the same.

So why not design a system that combines these approaches — recognizing the mundane, repetitive steps that can be amplified, as well as the opportunities that can benefit from large scale? Additionally, news companies already know they can compete against the pack by: (1) deepening coverage (investigative reporting or well-written opinion pieces) and (2) extending coverage into long tail areas. What they can’t do is scale these approaches because there are too many niche areas with too small audiences. Kiffets addresses all of the above by drawing on three sources of power for filtering and organizing information: the hard work of the few (curators), the light work of the many (crowd wisdom), and the tireless work of the machines (AI).

Open beta

To deliver the best of this information organized for its users, the Kiffets Social Indexing Engine combines the expertise of human curators with artificial intelligence technology and social input. Kiffets users subscribe to specialized channels on subjects they care about, curated by people that they trust. The curators (users, media companies’ in-house editors, others) tell Kiffets how to select and organize information on a subject. Then Kiffets collects, classifies, and delivers the information automatically to its users according to their interests.

Kiffets is now in open beta release on the Web. Try it out. Also check out and subscribe to the Kiffets Voice blog, which covers some of our thinking on the future of journalism, personalized news, and curation.

Jumping on the all-electric powertrain

Arguably, the all-electric powertrain will be the most significant disruption in automotive technology since Henry Ford introduced the first assembly line.

So far, the gasoline- (or diesel-) powered internal combustion engine has been a tremendous success despite its limited efficiency, particulate matter and CO2 emissions, complexity, and maintenance demands.

An all-electric powertrain, however, would eliminate many common failure modes associated with internal combustion engines such as oil leaks, broken fan belts, clogged fuel injectors, bad spark plugs, worn out piston rings, and more.

Owners of electric vehicles (EVs) would also enjoy benefits such as zero emissions (from the vehicle itself), instant torque, rapid acceleration, whisper-quiet engines, and significant reductions in their commute costs.

But don’t bid goodbye to the internal combustion engine yet

The recent introduction of plug-in hybrids such as the GM Chevrolet Volt and all-electric vehicles such as the Tesla Model S and Nissan Leaf (with its marketing-play acronym of Leading Environmentally friendly Affordable Family car!) have sparked a new wave of public interest and heightened expectations:

…Is it finally time for electric vehicles to displace the internal combustion engine for good?

While I believe the future of personal transportation (ground, water, air) is in all-electric propulsion, I have my doubts that it’s time.

We need a Moore’s law for batteries

There are still a few significant gaps in battery technology:

1. The energy density of gasoline is more than an order of magnitude better than the best Lithium ion batteries in the market.
2. Batteries deliver less energy (and age faster) during sustained high-power operations.
3. Batteries have limited life spans measured as total charge-discharge cycles [I've also written about this problem before, in connection to V2G schemes.] These deficiencies result in limited range and load capacity compared to fossil fuel vehicles.
4. Further, these technology gaps render current batteries impractical as energy sources for vehicles that require continuous high-power operations (such as personal aircraft).

Consider the infrastructure

In addition to battery technology gaps, there are significant infrastructure issues in widespread EV adoption.

For starters, aging local power distribution grids may not be able to support large numbers of electric vehicles charging at the same time without capacity upgrades. Houses, garages, and carports need to be equipped with high-current charging stations. A pervasive public charging infrastructure needs to be developed to alleviate the “range anxiety” experienced by early adopters.

Finally, if the power for electric vehicles has to come from coal-burning power plants, electric vehicles would not be as “environmentally-friendly” as they are glorified to be — see for example Scientific American’s recent coverage about the The Dirty Truth about Plug-In Hybrids (see the brief infographic here or get the full article).

Achieving pump parity

I expect the new generation of electric vehicles will be popular with enthusiasts and early adopters. Personally, I am considering buying one. However, I fear that this new generation of EVs will require too many compromises to appeal to the mass public.

It will take another generation of battery technology improvements for electric vehicles to reach pump parity with their internal combustion counterparts.

We rely on folders and keyword searches to sift through thousands of emails to locate buried messages and documents… but the problem goes beyond the inbox.

Today’s business processes are more dynamic, more human-centric, ad hoc, unscripted, and loosely orchestrated – they represent the framework for our interactions with team members, business partners, and customers. The information that fuels these interactions is digital: emails, documents, web site links, database records, IMs, tweets, and so on.

Keeping track of all this information in the context of a person, a partner or customer, or a particular activity is a TIME CONSUMING, MANUAL, CUMBERSOME process. And it’s only getting tougher.

Search is not enough

Sure, search can help – but it’s limited to finding information based on a matching, often arbitrary and ill-defined, keyword, and only then in the information format supported by the search tool. But…

• What if you can’t remember the right keyword?
• What if you want to find a document based on its meaning? (e.g., something related to a question in an email message)
• What if you want to find something that looks like something else (e.g., a sales chart used in a PowerPoint deck)?
• What if you want to find all the email messages that included a certain slide deck or contract?

Search fails, or at best, requires repeated, manual efforts to narrow down the results. This is time consuming, prone to error, and so, so frustrating.

Connecting information by context

What we really need is a way to connect, or relate, information across different domains by context – connecting information by relevance. Remember when you last gathered all the relevant financial papers to file with your annual tax return? Even for the better organized of us, this is painful and time-consuming.

Meshstro, a Xerox-funded project incubated at PARC, is addressing this “‘information connection” challenge (note I didn’t just say “overload”!) by using a specific combination of PARC technologies and competencies straight out of its research labs to create an information net – or what the team calls an “iMesh” – of the relationships between information elements such as the people, companies, and topics within email messages, documents, and other data sources.

Knowledge workers, unite!

Simply put, Meshstro is a software product that connects information by context to improve the way knowledge workers find, sort, and interact with information. We want to make people smarter and faster when managing their information.

Right now, Meshstro is available as a contextual Outlook sidebar – focused on the problem of business email first. Not only is email our #1 business application (57% of us are using email every hour, and typically receive 50-200 work-related emails per day), but it also contains the context of most of what we do. Encased in the messages we send and receive daily are the lifeblood of our tasks and responsibilities, our business relationships, and interests.

Knowing this context isn’t just powerful, it’s mission critical.

So, for now, Meshstro helps users find information buried deep in their email – accurately and rapidly – and very soon the product will link information systems together to provide a “contextual intelligence system” connecting email with file storage, business systems, and the web.  Then you can find everything you want smartly and quickly.

Sign up for the beta here.

In this second post in our series on ethnography [you can read the first one here], I thought it might be useful to provide an overview of data COLLECTION methods (and methodologies) that ethnographers use to understand a particular population or situation of interest; while specific needs vary, for our clients the general goal is to help them address a murky problem or innovate differentiated products.

Note the emphasis on data “collection” as opposed to data ANALYSIS. The latter is an important — though far less visible — other half of the equation, and is often where an ethnographer’s true expertise shines… But this is a different topic, so we’ll share more in an upcoming post.

Observation, observation, observation

Before joining PARC, I used popular observation methods to understand public health interventions. Since joining PARC, I have greatly expanded my definition, scope, and toolkit for “observation” because I have been faced with conducting ethnographic studies to satisfy the diverse goals of varied clients.

My colleagues and I still do what PARC has been credited with pioneering in the technology space, and that is human-centered observation: particularly through the use of video ethnography. We observe, record, and, when feasible, actually participate in the activity at hand. Video doesn’t lie — it adds a great deal of power to human behavior analysis — and we have an entire lab space devoted to video data sessions where our ethnographers join together to analyze human behavior recorded during ethnographic observation. Again, this observation methodology is an area of expertise that deserves more detail in a future post.

But the insight doesn’t stop with us. By showing client stakeholders examples of phenomena observed through ethnography – whether in the mobile domain, leisure settings, workplace, cityscapes, or elsewhere – we try to bring a “real world” view they might not otherwise see. In a recent workshop where we presented the end-user’s vantage point, our client commented, “Wow, we never really considered that user of our product — we have been so focused on our competitors’ product features that we didn’t consider this group of users.”

While the insights we show to clients are private to them, here’s a video of projector use, just one example of how we use video to show human behavior.

Asking, watching, and listening

Observation, though extremely powerful, is usually not enough. Ethnographers therefore can use any of the below methods depending on the situation or need to gain different slices of understanding a target group or situation of interest.

Semistructured or in-depth interviews. Ethnographers create a set of concepts or research questions on behalf of stakeholders. For us, the exploration plays out much like a conversation — answers tend to come out naturally, we may vary the order of the questions, and we can allow tangents if worthwhile. This flexibility (as opposed to sticking to a templated script makes more probable unexpected discoveries and really getting what people care about.

Show-and-tell. To avoid relying on imagination in unfamiliar situations, we often ask study informants to demonstrate the very things they are describing.

Think-aloud protocols. We ask study informants to fully explain what they are doing and thinking as they do it, so that we can better understand their objectives, thought processes, and decision-making processes.

In situ interviews. Often used in combination with more unobtrusive direct observation, we ask questions as people go about their usual activities so we can understand context that may not be obvious.  While we do this sparingly to avoid burdening study informants, we are always balancing the need to observe undisrupted behaviors with the need to understand what we are observing.

Shadowing. In some cases we are interested in the behavior of one individual over a length of time so we will shadow virtually every second of that person’s life for a set period of time (e.g., 1-8 hours a day).  We mike the subject and use wide-angle HD video cameras to capture as much as possible of his/her life within the given time period.

Focus groups. Used sparingly at PARC (at least in the way that market researchers use it!), this technique is more often a problem-solving vehicle or design endeavor. So we may use these in an ethnographic consultation where local informants are given the opportunity to participate in outcomes. Groups can range from the classic 8-12 participants from a particular demographic to subjects gathered at a study site for a particular set of questions or reactions.

Tracking

It may not be enough — or even feasible — to watch and ask. So we’ve adapted or developed many methods to assist in tracking and ultimately understanding human behavior, including:

• surveys — one-off or repeated (e.g., daily)
• diaries — paper or electronic
• mobile- or desk-based experience sampling – which may be randomly timed or triggered by contextual data from logging, sensing, or instrumenting interactions in information, social, and virtual environments.

These methods can generate qualitative data (e.g., diaries) or quantitative data that might be analyzed using descriptive or inferential techniques and include machine-learning approaches.

Enter customization. We often have to build custom software to collect this data, which often requires advanced computational methods to analyze it. Since PARC has the advantage of social scientists working closely with other scientists, we regularly collaborate with our computer scientists to conduct “hybrid” studies that combine some of the above. And yes, we’ll have much more to say on this topic.

The science and art of ethnography is not in a preset formula for these individual methods. It’s in the selection, unique combination, customizations, and analysis — which together can yield the “deep” understanding that in turn inspires innovation, or fosters change.

There IS a method to the madness.

Currently, our Slideshare channel features some of the most highly attended or popular talks from the past couple years — as well as other PARC videos and presentations.

However, you can watch or listen to all of the available video and audio anytime at www.parc.com/forum, the multimedia archive of PARC Forum talks — organized by category (including PARC Forum talks by PARC speakers), speaker series (whenever we host special themed ones such as Ethnography in Industry, Cleantech, Going Beyond Web 2.0, Innovation, and more), and year.

Be sure to subscribe to email announcements or feeds for regular updates about upcoming talks and more.

What is ethnography?

First: there are some helpful definitions that can be found through a simple search.

In case you’re in a hurry, I’ll also summarize it (albeit inadequately, no doubt) for you: a holistic, in-person, and qualitative approach to the study of human behavior and interaction in natural settings.

But rather than expound on the semantic aspects of ethnography in my very first blog post here, I’d really rather respond to the obvious and eminently reasonable question I often hear in my work as a researcher in the field of user-centered technology innovation:

“What’s it good for, in my business?”

Ethnography adapted for industry

In today’s hard-nosed and often economically trying times, ethnography can be seen as a tactical weapon enabling companies to gather new insights and thus gain advantage over their competition.

Traditional ethnographic studies were conducted at a relatively leisurely pace. They had, at least as far as I can tell, no particular useful or focused objectives other than to uncover as much as possible about a culture or practice of interest in an unfettered manner. (Indeed, having an explicit agenda was considered to be rather bad form and was liable to get you kicked out of polite ethnographic circles…wherever those might have been.)

Out of the academic Garden of Eden, modern ethnographers have been driven to move and produce compelling results faster, while operating within a number of budgetary constraints and oft-conflicting business demands.

Ethnographers’ data collection and analysis methods have therefore been condensed, recombined, adapted – both systematically and as-needed – to meet these business demands. We’ll describe the methods to this madness in our next post, but in this post (below) I categorized some of the commercial objectives for which these methods are applied.

Objectives for ethnography in industry

There are many practical applications of ethnographic methods in commercial contexts, particularly those that involve technology and workscape innovation (which is what my colleagues and I are most approached about).

Because there is considerable confusion about these applications, I want to propose a simplified classification of purposes to which ethnography can be applied.

Engage in opportunity discovery.

• For innovating products and solutions, the objective might involve understanding the culture, practices, and needs for support within a particular workscape – for example, employees of a bank or a manufacturing plant.

• For innovating or improving processes, the objective might involve identifying problems or undesirable phenomena in some setting that could be fixed by some organizational or procedural change – such as misunderstandings that lead to mistakes in satisfying customer requests in contact centers.

Sometimes, the goal is to understand the “big picture” (e.g., leisure outings with clients as part of the sales process). While these goals may seem open-ended on the surface, in industry settings one or more stakeholders always have an objective in mind (e.g., to understand the subtle returns on investing in leisure activities, since they can’t be captured in other ways).

• For settings, the open-ended objective could be to understand the rich culture and variety of tacit and explicit practices within a particular workplace – those of bank or manufacturing plant employees, for example.

• For activities, the open-ended objective might be to understand all tacit and explicit practices related to a particular role, practice, or function (such as contact center operations).

Gather information for a targeted purpose.

• Diagnosing the cause and/or nature of one or more specific problems (such as employee dissatisfaction and high turnover in a company) or product failures (for example, why a certain hardware device is proving to be unpopular in its target market).

• Identifying requirements that will inform the design of a specific product or enterprise solution – for example, what’s the best form factor for an e-reader targeted at executives and knowledge workers reading work-related content?

Test assumptions and/or evaluate performance.

• Oriented towards practices, the test objective usually involves understanding the real vs. ideal execution of some role or process of interest – for example, if and why contact center operators aren’t following their script? An evaluation objective might be to assess performance in terms of some explicit values – for example, how satisfied is the customer, or, how often does a store assistant get the customer’s order wrong?

• Oriented towards usage, the test objective may be to understand the real uses of some artifact or product “in the wild”, as opposed to how it was designed to be used – how are people really using photocopiers, for example? (This was famously studied by Lucy Suchman at PARC in the 1970’s, which is when the notion of corporate ethnography started to take off). An example of an assessment-oriented objective here might be: how quickly or smoothly can someone perform a task with a specific device?

Transfer or create ethnographic methods/ competencies.

This is an area PARC uniquely specializes in, and gets numerous requests for (particularly from foreign corporations who seek to differentiate themselves from their upstart competitors).

• The didactic objective might involve some sort of learning-by-doing (where we scaffold analysts or employees in practicing the methods themselves), or some demonstration (where we illustrate methods and their value to those who want to learn their value).

• Finally, there is also reflexive methods development, where one develops and tries out new ethnographic methods in practice by trying them out and refining them based on ongoing experiences and findings. This is something we constantly customize to meet specific study objectives and continually evolve the field.

Mapping objectives – often in the plural

I would caution that there are definitely some nuances in how you customize, combine, and apply ethnographic and other related methods to suit specific needs that I didn’t have time to go into here — this list is not complete and represents a simplified classification. My intention was to explicate and delineate the diverse objectives behind ethnographic studies that I’ve experienced at PARC.

But don’t assume that there can’t be multiple objectives for any given ethnographic study. The opportunity to glean field insights is rare, so it might as well serve multiple objectives at once – as long as one doesn’t dilute focus in attempting to answer every possible question.

And of course, there are always high-level objectives such as “give Apple a run for their money” or “make our company look more intelligent”. The art of an experienced ethnographer is knowing how to map the general classification above on to business leaders’ particular desires.

THE WHY & THE WHAT: A refresher on how to define what you want

In part 1 of this post, I shared some advice for ways to define what you’re looking for, and adding the crucial first step of defining why you’re scouting at all.

Just a refresher: I believe that technology scouting mandates should be defined as part of a company’s technical and product roadmaps. At that level, leaders can concretely say what they will invest in creating internally, what they want to get from partners, and where they want to place multiple bets as a risk-mitigation strategy.

THE WHO & THE WHERE: How do you find what you’re looking for?

The amount of available technology has exploded. At conferences like TechConnect and World’s Best Technologies (WBT), scouts can see hundreds to thousands of technologies in just a few days. Pharmaceutical companies’ technology scouting offices are used to seeing several thousand submissions per year. One veteran technology scout from the aerospace and defense industry told me he wish he’d known, when he started years ago, “just how much technology is out there”.

Finding “potentially interesting” technologies is easy. Finding great business opportunities is not.

There is an important difference between the role of a corporate venture group and that of the scouting organization:

• A venture group is looking for a complete story — a relatively proven technology and a tested, validated market proposition of clear relevance to the strategic investor.
• The technology scout that needs to see such a tidy bundle is missing the point. A technology scout can add more value to the company by seeing nascent opportunities than by only serving as a filter.

Here’s why.  Like many technology resources, as PARC develops its own technologies, we identify specific market opportunities that can be served by the new approaches. For example, we’ve invested in new forms of networking. We’ve identified distinct market opportunities, such as alleviating congestion on the telecom backhaul and improving in-home connectivity for IPTV. Once we identify these pain points and a way for our technology to address them, we will be talking to every major company in that space, as well as evaluating the spinout potential. If you’re a telecom vendor hearing from us about a telecom opportunity, most likely your competitors are hearing from us too.

Compare this to a technology scout who “wants to hunt where my competitors don’t,” a phrase I heard recently from one of my potential clients. The opportunities and application spaces we can come up with together are going to be much more robust and differentiating than what either of us would have come up with on our own.  More importantly, the scout will have discovered and honed a unique proposition that his/her competitors aren’t even thinking about.

Fertile hunting grounds for finding technologies

Below is a list of places to find technologies that came up when I posed this question to a panel of scouts last year; I’ve added my suggestions for each.

Technical staff. Look for the internal people that are connected to applied research in adjacent industries. If working with universities, seek the technical staff’s help in understanding which professors and institutions are more “commercial friendly” than others.

Retirees. The fact that they know your company internally and your market externally can make this group a surprise asset. Particularly valuable are those that have gone on to consulting and/or sit on advisory boards that enable them to keep their fingers on the pulse of the landscape.

Start-ups. Your marketing team is watching the competitive landscape, and will likely be able to point to smaller emerging companies that may be interested in mutually beneficial technology exchanges. When working with start-ups, though, remember they need to stay focused, particularly if they are venture-backed and looking for an exit.

Search. Some people have come to us through patent searches and technical publications. Some scouts tell me Google is their best resource for such research.

Specialists. At PARC, we have had good success with Visiting Technologists, individuals steeped in a specific industry who are able to help precisely map technologies with application spaces in a domain (they have what I call the “Industry Instinct”). They also see opportunities to pair our technology with other technologies that we might not be aware of.

Existing suppliers/customers/partners. This works best when you are making incremental advances within a market segment, but is less useful for forward- or backward-integration, or if you want to enter new markets.

Intermediaries/matchmakers. These are good resources for what I described as “horizon 1″ needs (usually < 12 months) that can be succinctly and specifically described. Since such intermediaries tend to have an industry or discipline focus, be sure their network of solvers is aligned with the type of problem you want solved.

Industry trades, newsletters, meetings. Like the intermediaries, these help set you up for serendipity. Efficiency is key when using resources like these due to the sheer volume.

Independent centers & national labs. Ideally, look for organizations that you might do business with more than once. In the best scenario, these resources will do more than license their technology. Whereas a start-up or other company may make their technology accessible, it is harder for a start-up to make people available. The independent labs (like PARC) should be able to help in customizing a technology to an industry application, supporting tech transfer, and providing post-license consultation. You’ll also usually have someone like me to help make the right connections and navigate both the available opportunities and processes for securing them.

THE WHAT NEXT: How do you get it, and manage it once you’ve got it?

This should be the easy part, because it’s just about cutting a contract and acting on it, right? (I can hear the collective groan from all of you who’ve been through this!) Okay, here are some tips.

You’ve found it. Now what?

Centralize mentorship for the acquisition process. One of the common failure points for scouting is having responsibility for finding technology, yet no role in acquiring it. Although the scouting function doesn’t own the business outcome, I strongly recommend centrally coordinated mentorship for the acquisition process so that you are not re-inventing the process from scratch every time.  Note that I’ve chosen the word mentorship rather than management.  Integrating outside technology is hard enough in most companies, without having a function that is seen as a bureaucratic barrier. A centralized team needs to be facilitators and experts in what to expect in such deals, to guide a variety of stakeholders throughout the acquisition process.

Involve stakeholders early. Gene Slowinski points out in Reinventing Corporate Growth that every alliance is really 3 separate alliances: the visible one between the companies, and the invisible ones that have to be built within each company between the internal divisions. Those internal alliances should be built at the same time or earlier than the alliance between the companies.

Avoid cookie-cutter contracts. The people putting together the agreement need to understand the business as well as technical and legal fundamentals, and will have to make trade-offs between those. Handing them a template and saying “you can do whatever you want as long as you don’t change it” ties their hands and forces suboptimal deals. It also makes the company appear difficult to deal with. See next tip.

Seek to be the Partner of Choice. This applies throughout your interactions with a potential partner. In an open ecosystem, the company that genuinely seeks out win-win becomes the preferred partner for innovators. One of P&G’s competitors shared with me that his primary metric for success would be getting to see a deal before P&G; clearly, within that segment, P&G is seen as the Partner of Choice.

You’ve got it. Now what?

Once you’ve acquired rights to a technology, the organization needs a practice for managing to the business results.

Create an integration team. One of the best practices I’ve seen is creating a cross-disciplinary team early in the process that creates the integration roadmap — a document that outlines what are you going to do once you have it.  Usually involved in the technology evaluation, this team is setup to be the team that runs with the commercialization and integration processes after the contract closes.

Set intermediate milestones. These are especially important when a relationship requires working together. PARC contracts frequently have interim milestones that allow a company to evaluate their continued advancement of the technology, and the relationship, because both matter towards a successful end result.

Create an internal audit process so you know how the licensed IP is being used. Since in-licensing agreements frequently have field-of-use limitations, it’s important to track what the boundaries are and ensure you are abiding by them. Similarly, if royalties are due on products that use the other party’s IP, you need a mechanism for tracking not only the initial products that come from the relationship, but derivative products that may also require royalty payments.

Establish a review/check process. This is as important as the intermediary milestones for both parties and, if conducted as an honest health check, will help identify potential issues before they become serious problems. It can also help identify further opportunities.

It’s more than what you see here.

Any treatment of this topic is going to be incomplete, and that’s the beauty of this medium. I’d love for readers to add comments about other best practices you’ve seen and would recommend…

While I’ve written this talking largely about the formal technology scouting function, in reality, the principles apply for many people that help keep their companies connected to the rest of the world. As companies are developing more dependencies on external partnering, the models and standards are still very much emerging, which makes this a wonderful time for learning from each other.

A telling trend

One of the most telling trends I’ve seen is revealed through a National Renewable Energy Laboratory chart [at right] showing solar cell efficiency versus time using silicon, amorphous silicon, CdTe, copper indium gallium selenide (CIGS), organic, dye-sensitized. The trend lines have a nearly universal slope, with efficiency increasing in the lab at the rate of only about 4 percent per decade.

While there may not be a universal law behind this trend, the sobering fact is that materials-related research takes a very long time. Progress comes in fits and starts, often with serendipitous occurrences. One striking example is the way that CIGS efficiency improved around 1993 when the material was grown on soda-lime glass instead of borosilicate glass. It was later found that sodium from the soda-lime glass was diffusing into the solar cell, improving its performance.

Why is this trend telling? The consequence of such slow materials development is that a new or complex materials like nanostructured photovoltaics do not have a real shot at mitigating global warming. Caltech’s Nate Lewis has argued that we will need more than 10 terawatts of carbon-free energy by 2050 ” the equivalent of building 1 large nuclear power plant every other day — for the next 40 years. If we assume that it will take 15 years for a new material to reach the commercial market at 1 megawatt scale, and that the annual production capacity thereafter grows with a 50 percent CAGR (which is greater than solar), then the technology’s installed base would still only amount to 0.075 TW by 2050.

Given the math, it is clear that global warming needs to be addressed by rapidly scaling the technologies that are already proven at gigawatt scales (wind, PV, nuclear), and by implementing energy savings technologies that are not based on radically new materials. So, while it might be possible for a technology or company to be a “winner” in terms of VC investment, this is not the same as winning against global warming.

Promising technologies

With those realities in mind, I highlight some promising technologies from the ARPA-Energy event:

• Transonic Combustion’s supercritical fuel injection system looks like it could save 50 percent of fuel usage across a wide range of engine types. The attributes that make it attractive are that it doesn’t entail any fundamentally new materials, and could be scaled very rapidly to yield large savings in energy usage.
• While admittedly biased, I include PARC’s thermoacoustic cooling technology [pdf]. By using high-intensity sound waves to move heat near room temperature, the technology could potentially achieve twice the efficiency of current air conditioners. Again, what makes this approach attractive is its rapid scalability through engineering existing components.

Another interesting technology is based on a simple strategy we often use for rapidly scaling new technologies: leverage existing manufacturing approaches.

For example, working with BP Solar and other companies, PARC has been able to exploit its printing expertise developed for Xerox to print high-efficiency contacts onto solar cells. Since the focus is on working within existing manufacturing constraints, our method uses a drop-in printing approach that can convert a 100 MW fab into a 106 MW fab — at roughly the same cost as existing processes.

In general, printing is one of the lowest cost means to fabricate complex patterns on a substrate. We’re now extending this approach to print complex 3-D structures such as battery electrodes.

Recurring themes

The recurring theme here is scalability, which is essential in addressing the true scope of global warming. This attribute alone separates the clean tech winners and losers.

Clearly, it’s a different innovation landscape today. No longer can companies work in silos with their captive research departments to achieve organizational goals, or simply pick up a company to acquire expertise they still won’t know how to scale. To generate high-value clean technologies, researchers, VCs, large corporations, government agencies, universities and policymakers have no choice but to work together.

Editor’s note: The Economist invited us to contribute an abridged version of this post, “How do you define innovation?“, which you can read at their first Innovation Summit: Fresh Thinking for the Ideas Economy blog. A separate but related Q&A interview, “Creating Innovation in the PARC“, is also available at Blogging Innovation.

Innovation is a sorely overused word. Yet we are constantly asked to define it. A number of theorists and practitioners have offered up their variations: product innovation, business model innovation, technology innovation, design innovation, radical innovation, incremental innovation, disruptive innovation, open innovation…and so the list goes on. All are useful; none are complete.

I don’t have a pat answer, catchy definition, or compelling metaphor for this. But here’s what I do know: however it is defined, innovation is a valuable change, unconstrained by the way things are.

Must. Add. Value.

I think I can safely claim that we’re speaking from experience… PARC was extremely fortunate to have been given the space and freedom to create “the office of the future” (Xerox’s founding charter for us), even though no one knew at the time what the outcome would be. We were even more fortunate to have delivered on that mission – transforming the way people worked through distributed computing technologies such as the GUI, Ethernet, mail & file servers, WYSIWYG displays, laser printing, collaborative workspaces, and more. The impact on the world was indisputably valuable. Xerox adopted the results of the innovations they understood best (such as laser printing, which became a multi-billion dollar industry for them), and other companies helped commercialize their piece of the full vision (see partial list of companies here).

Today, PARC focuses on “valuable change” for all of our clients, which include large multinational companies from different industries to startups incubated on site from our own work or select entrepreneurs. The change almost always involves some kind of new business creation – from helping our clients envision what the innovation in their industry will look like, to collaboratively defining the short- and long-term goals of the innovation and co-developing the solution. So for us, the focus on innovation-as-valuable-change hasn’t changed one bit. What’s changed is who drives that change, the range of industries we now innovate in, and how we manage the process for value creation.

A Changing Model

First, let me say this: innovation today is different than it was 30 years ago. Ditch the image of a lone inventor, discard even Silicon Valley’s beloved image of two inventors in a garage. In fact, forget about the notion of captive R&D as the sole source of competitive advantage.

The fact is that innovation today explicitly involves an expansive ecology of large enterprises, VCs, government agencies, university researchers, and others. This not only engages the multiple players necessary to transform a given innovation into a valuable change, but it also provides the breadth of perspectives that we use to address a specific problem (such as flexible electronics, for example).

Government agencies can provide the leeway to pursue platform- and market-agnostic approaches, while large enterprises provide the manufacturing and value chain constraints that will make the solution viable for wide adoption. It’s not a difference of kind (public vs. private sector funding?), but degree (how much? what’s the right mix of sources) and timing (when? what’s the best stage of development to engage a particular driver).

Key Ingredients

Other key ingredients for cultivating valuable change include:

Openness. This doesn’t have to mean transparency – innovation can happen under confidentiality agreements! – but an organization’s genuine interest, flexibility, and adaptability in absorbing solutions from the “outside”.

Breadth and depth of disciplines. This is not as simple as “stick a bunch of smart people in a room” and brainstorm with sticky notes and design thinking. Valuable, transformative innovation requires deep domain expertise and scientific insights – including those of social scientists who have been immersed in diverse field settings – that enables them to see possibilities where others cannot.

Range of problems and opportunities. The problem can be one presented by a client, it can be an implicit user need made explicit through ethnographic observation, or it can be motivated by scientists’ passions to address a fundamental human need (such as clean water). Note, though, that the way the problem is defined might be different from what the problem really is. Savvy inventors need to be able to address and articulate this gap.

Market expertise. While market insight can come from many sources, for us it is provided by the client (“here’s what I need/ where I want to be”), embedded business development (“what are the valuable applications for this technology”), and/or EIRs (“what’s next?”). Xerox PARC cultivated three decades of deep expertise that were applied to a very specific set of markets, but today PARC, as a commercial innovation center, is taking that expertise into different industries. For example, the same science behind manipulating toner particles in copiers has led to a potentially valuable hydrodynamic separation technology for the water and industrial filtration industries – but to make this a marketplace reality, we need partners who can qualify the best potential application values.

None of these ingredients is a secret. The hard part is managing the mix so that we – and thus our clients – stay on the positive side of the value equation.

But first: What’s an ARPA-E?

Organized by the the U.S. Department of Energy’s Advanced Research Projects Agency – Energy (“ARPA-E”), the inaugural Energy Innovation Summit brought together key players to spur the networks that will, according to the organizers, “bring about the next Industrial Revolution in clean energy technologies, in the way the U.S. has led previous revolutions in life sciences and information technology“.

• LA Times — “A federal effort to lead ‘green’ technology”: An energy research agency applies the same theory that led to the invention of the Internet: that government funding churns out radical innovations. Director Arun Majumdar talks about the challenges.
• Grist.org — “Obama’s ‘director of game changers’ talks energy breakthroughs”

My take: The technology showcase

The Transformational Energy Technology Innovation Showcase (I didn’t call it this, they did) was fantastic — everything from zaney new wind turbines to quantum tunneling thermal energy converters. [FYI, here's a list -- with links to downloads -- of what PARC shared in the showcase.]

Overall, the showcase displayed a diverse creativity that I found truly inspiring. Here’s a quick rundown of what I found noteworthy:

Photovoltaics (PV). Surprisingly (or not, given how saturated the space is) there were only a few things I saw on PV devices. We showed one idea on using laser ablation to make point contacts [PDF] — but we didn’t really emphasize it, given that the work is already funded by the Department of Energy, and honestly it’s more of an incremental improvement to silicon PV technology. While our technology is important in driving down manufacturing costs and increasing efficiencies, the mission of ARPA-E is to focus on the truly bold, new ideas. Meanwhile: there wasn’t much novel thin-film PV work visible at the meeting. That seems right to me, given how many companies are already awash in VC funding ($100M+) with the goal of making CdTe and CIGS a reality.

And Thermo-photovoltaics (TPV). Then there is the quantum-tunneling approach to TPV. Now this is a really cool idea: couple evanescent waves from the heat of a blackbody by placing a TPV receiver only a gnat’s eyelash away (at 100 nm). This way, you can extract more energy than the blackbody radiation law would seem to allow. It should work because within the evanescent field, you’re taking advantage of coupling to NON-radiative modes. But the big issue is how to maintain a very, very small gap over macroscopic areas. IF that can be done, the approach looks very interesting.

Wind power. I’m actually really impressed by the scale achieved by wind turbines made by companies like General Electric. So I have to wonder if some of the seemingly wacky ideas around wind capture that I saw at the showcase will, er, fly (pun not intended). Especially because I believe wind technologies should increasingly lower cost through manufacturing improvements and economies of scale. It’s not just the cool factor but the technology adoption that matters.

CO2 conversions. There were a number of ideas around converting CO2 to liquid fuels (including our own). In fact, there were specific ideas that would require high-pressure, pure, CO2 — which is exactly what the PARC air capture process produces [PDF]. From a partnering perspective, I think this is a solid opportunity where the sum of the parts might be much more interesting as a whole.

Thermoacoustic cooling. Wrapping up the list, the showcase displayed PARC technology on thermoacoustic cooling [PDF]. Another very interesting idea, although I’m admittedly biased. [The technology was also mentioned or covered by Technology Review, Greentech Media, and Scientific American. Nikkei Electronics too, though I can't read Japanese.] How it works: it uses thermoacoustics (high intensity sound waves to drive heat), but we have added some proprietary ideas to make it more efficient than what has been possible before at near-room temperatures. The result would be an air conditioner that is 2X more efficient (a COP as high as 9) than conventional compressors. Huge market.

If you were at ARPA-E, or simply interested in these technologies, we’d love to hear your thoughts…