Technology and Organizations

I can’t say there was standing room only at Marissa Mayer’s (VP of Search Products & User Experience) PARC Forum talk “Innovation at Google: The Physics of Data” — because the fire code for the auditorium doesn’t allow standing. They did bring in extra chairs and let the rest watch via TV in the lobby.

So many reasons for the talk to be popular: Google, Innovation at Google, Big Data issues, Marissa’s presentation style (engaging, you feel like your getting a secret tour). I’d gone because of the big data and innovation focus and came away with a better understanding of their process, and how their management style is as empirical as their products.

She opened by explaining the title of the talk. Besides the fact that it sounds cool, Google is on a physics kick with products like: Google Wave, Google Fusion, and their focus on acceleration.

She quickly moved to how data helps them build better products. She gave a clear example in terms of the testing that went into their choice of the shade of blue used in their links (see mention on Gigaom). Another example was how Google Trends can help anyone know if a product is trending up or down to help them think about stocking. Below is a Trend search on Crocs.

The highlight for me was how a culture of data/evidence supports their management. I’d seen Marissa talk about this in a video (below, section starts at 16:28) as “Data is Apolitical.” Evidence-based management (here and here) at a company that builds the tools that let us all organize our evidence.

From the video:

I think that the internal politics inside of Google have remained minimal compared to other corporations of its size because we rely so much on the data and we do so much measurement that you don’t have to worry, will your idea get picked because you’re the favorite, or will someone else’s idea get picked because they’re the favorite or because they have a better relationship with the person who’s the decision maker. The decisions get made based on data, and that really frees people from a lot of those types of concerns.

How can we use big data to make better management decisions? In earlier posts (here and here) I’ve listed a few tools and asked similar questions. Marissa described the power of other Google Tools to help us “understand problems in new ways”:

• Trendalyzer -- animations of trends over time
• Fusion Tables -- largely for researchers -- mashups and embedding of datasets.
• Google Squared -- googling your Google search results (thus squared) or the square that results. Not perfect -- but it is editable (e.g., you can add your own columns).
• And of course their (and everyone else’s) move to manage the real-time web.

What problems can we understand in new ways? How can the exploding amount of data (for example from the US Government) help us manage better?

Some starters: Facilities planning, recruiting, flexible work based on better predictions of business cycles, data mining of employees’ social media content…

Tweet This Post

Yesterday was Reid-Hillview Airport Community Day. One of the activities was a tour of the Control Tower. Great experience. Thank you to Vincent and Spencer for taking the time to explain the process that keeps hundreds of flights going in and out safely. Thank you to the rest of the team for letting us observe you at work.

I was surprised by how physical the process is, versus my high tech expectations. Yes, they have access to radar and a huge portion of the work involves radio communication with the pilots going in and out of the airport. But they also make heavy use of those big windows and a unique physical tracking system.

They track planes by type, tail number, and request for inbound or outbound route — by writing the information on plastic “pucks” with a grease pencil, and then physically sorting that puck onto the taxi and runway slots. We weren’t allowed to take pictures, so I’m showing a similar process below using wooden blocks.

When I asked about the process, using the plastic pucks versus keeping track on a computer, I was told that sometimes “elegant is best.” Great point! The solution is elegant in that the physical blocks trigger sensemaking (in my words) more than a screen version might. They can push a puck slightly out of its track to highlight that more action is necessary. All the members of the team can immediately step in to provide relief given their common understanding of the system. Elegant, green (no need for power or paper), easily visible to all in the room — good for team visualization.

Beautiful approach to a complex problem. Sometimes systems savvy means using elegant, but less high tech systems. Comments appreciated describing other examples.

Tweet This Post

Last Friday I had the opportunity to attend Web 2.0 Expo in San Francisco.  San Francisco transit connections being what they are, I missed some of the opening keynotes, but luckily I didn’t miss Jeff Veen’s Designing for Big Data.

Jeff’s presentation echoed some of my thoughts — we have access to much more data than ever in the past — how can we gain value from it?  As he noted, we don’t want to just “decorate the data.”  We want to gain value from it.  His new start up, Small Batch (apparently stealthy given this website), created Wikirank for visualizing and comparing what people are reading on Wikipedia.  The tool provides interesting data in a clean interface with some user control in that they allow you to decide what data to compare:

Compare Wikipedia topics.
At the top of the front page, you’ll see a comparison of two or more topics that we’ve chosen. From that section, you can view more detail on the featured topics, share the chart with others, or choose your own terms to compare.

They also ask that you share particularly interesting comparisons with them (and provide an email link).

I think this last, user control of data comparisons, is going to be key in how we “design for big data,” and eventually in how we gain value from big data.  We have the data.  We are gaining the tools and skills.  Next we need to understand how to extract value from this combination.

Quoting from a prior post:
Bellamy et al. draw from Card, Mackinlay, & Shneiderman’s book, “Readings in Information Visualization: Using Vision to Think” to suggest that visualizations should increase people’s visual capabilities and amplify cognition. They note that Card et al. outline six visualization benefits:

1. Increasing the users’ memory and processing resources
2. Reducing the search for information
3. Enhancing pattern detection
4. Enabling perceptual inference
5. Using perceptual attention mechanisms for monitoring
6. Encoding information in a malleable medium

I followed that post with one thinking about how teams might gain value from visualization of team data.  Veen’s presentation, and his wonderful links back to data mavens like Edward Tufte, push me to think about how big data can provide more general value.  Tom Davenport’s ideas around Competing on Analytics are a great background, but we have even better access to data inside and outside of our organizations since his 2007 book.

Here are some views into big data.  I’m not saying they all have value, but provide them as a start to some brainstorming.

Wikirank (described above)

Yahoo Pipes allows normal people to “aggregate, manipulate, and mashup content from around the web”

Twitter visualizations

ProgrammableWeb.  If you want to go deep, go here.

What else should we add?

Tweet This Post

One of the benefits of living in the Silicon Valley is the ability to interact with the world’s great innovation experts.  Early in the week it was Vint Cerf at Stanford, and today, at my own Santa Clara University, Sarah Miller Caldicott.

Ms. Caldicott is founder of Power Patterns of Innovation, co-author of Innovate Like Edison: The Success System of Americas Greatest Inventor — and a great grand niece of Thomas Edison – holder of 1093 patents.

We need some perspective to understand Edison’s innovation contributions. Edison created the first R&D laboratory (Menlo Park) with a “systematic approach to innovation.”  Evidence that his methods work: A 1919 New York Times article reported:

It has been estimated that there are a thousand million dollars invested in the industries which he has either created or for which he has laid part of the foundation; and a million employees are in these industries.

He established 6 industries (in 40 years):

• Document duplication (1873)
• Telecommunications (1876)
• Recorded sound (1877)
• Electrical power (1879)
• Motion pictures (1893)
• Portable power (1905)

Working with the Thomas A. Edison Papers Project at Rutgers University, Sarah Miller Caldicott has identified the organizational processes Edison developed to support his innovation.  Key is how organizations today (e.g., Google, Apple, Target, Starbucks, Pixar) use these same processes with great success.  She did a wonderful job of bringing the issues down to a personal level: The Five Competencies of Innovation.

• Solution-centered Mindset: (she says most important) Imagine the solution, then get ahead of it, and use experiments to nail it down.  Edison conducted over 10,000 experiments in the first year of working on the storage battery.
• Kleidoscopic thinking: went beyond brainstorming – whole brain thinking. “To have a great idea, have a lot of them.”  Once he invented the light bulb, he had to invent the system to run them.  He looked to telegraphy and electricity (multiple streams of knowledge).  He drew six pictures and one became the first electric circuit.  Use fantastical storytelling (dropping self censorship by getting your brain to think in fiction form).
• Full-spectrum Engagement: Important to work in solitude, and also important to work in teams.  Opposites — important to be intense, but also to relax.  Sharing and protection.
• Master-mind collaboration:  Cross functional teams, organizational design (org and physical structure) that enabled ad hoc interaction across expertise areas — in the late 1800s!
• Super-value creation:  Ethnographic research to understand the market. “A believer in going and looking.”  Creation of a brand.

Again, my professional interest is in how he was able to systematize the innovation process.  The answer is hard work and innovating around process as well as products. The above five competencies are known, if not applied widely, today. But in the 1800s these were unique approaches.

In the question session, one of the participants asked about Edison’s friendship with Henry Ford. Ford was 20 years younger, but worked at the Edison Illuminating Company and eventually became friends with Edison.  Both industrialists were organizational design innovators, as well as great innovators of products.  It also came out that Edison hated people calling him the Wizard of Menlo Park as it made it seem like what he did was easy — rather than 99% perspiration and 1% inspiration. We can all keep this last point in mind: Innovation doesn’t fall from the sky.  Even process innovation must be carefully managed — but the results are well worth it — even in 1919 dollars!

Tweet This Post

It seems that another train has left the station. Location aware computing is available, and its ubiquitous integration with our cell phones, laptops, and person is only a matter of time.

In January, Mathew Honan described (Wired) his experiments with a location aware lifestyle. Yet, even as recently as 2007, locational privacy was on the table in discussions of Google Earth (Randall Stross, in Planet Google, pp. 142-151). (Google Privacy page)

Clearly times are changing. I wish I’d kept the peer reviews from 1989 calling me a fascist given that my dissertation was on the effects of computer performance monitoring. Never mind that in the later field research the chip inspectors found electronic monitoring less invasive than physical (watch over your shoulder) monitoring. My argument has always been that monitoring is a tool and that the outcomes are a combination of the tool’s capabilities with the people and organizational practice.

These issues identify that a key skill we need in our roles of accidental systems designers is the ability to understand and manage privacy and information access. Motahari, Manikopoulos, Hiltz, & Jones (2007) describe these Seven Privacy Worries in Ubiquitous Social Computing in their paper of the same name (pdf):

1. Inappropriate use by Administrators: e.g. The system admin sells personal data without permission.
2. Legal Obligations: The system admin is forced by an organization such as the police to reveal personal data.
3. Inadequate Security.
4. Designed Invasion (Poor Features): e.g. a cell phone application that reveals location to friends, but does this without informing the user or providing control of this feature.
5. Social Inference through lack of Entropy: See CampusWiki example…
6. Social Inference through Persistent User Observation: e.g. Bob is so often in Alice’s office. Their relationship must be romance.
7. Social Leveraging of Privileged Data: e.g. David can’t access my location, but Jane can. David asks Jane my location.

The Economist published Every Move You Make in 2008 noting the following about Waber and Pentland’s study of locational monitoring (via identity badges that could track location and the timbre and inflection the wearer’s voice!) in a US high-tech firm and a German bank:

An interesting experiment, then. But how widely this approach can work in practice is unclear. Many people may object to having their behaviour scrutinised so closely and Mr Waber and Dr Pentland are, indeed, sensitive to privacy. They believe that the risk of rejection can be minimised by using the badges only for short periods of time, so that they do not become part of a routine monitoring system. It will also help, they believe, if everyone is treated equally, so that the boss’s actions, foibles and shortcomings are as transparent as those of his minions. Now that really would be a revolution in management science.

Contrast the above with the sentiments expressed by this Carnegie Mellon graduate student (using Lococcino, 2:43min video) and the popularity of iStanford.

I continue, without success, to search for serious sociological research on our changing views regarding locational and other monitoring. Pointers appreciated.

Tweet This Post