Posts
Last week, I attended my first Social Media Breakfast. Social Media Breakfast - Minneapolis/St Paul is a group that meets for breakfast (bacon!) periodically to discuss social media.
At the most recent SMBmsp (the 19th!), the event was divided into two sections - a presentation and demo by mixmobi and an interactive discussion about the state of social media use within corporations.
Mixmobi is a quick and easy way to deploy announcements/coupons/offers via social mediums and the ability to report on those campaigns in near-real-time. The tool, which they demoed live at SMBmsp, allows nearly anyone to deploy campaigns - no coding or technical background required. The resulting campaigns can be deployed via SMS or Twitter, I believe, although they were looking to expand that list.
The feature that I found most fascinating (as you'd expect, given my background!) was the near-real-time reporting capabilities. During the presentation, they deployed a postcard via Twitter - and near the end of the presentation, went into the reporting feature to show it off. Already, we could see some of the stats emerging. Quite a departure from my experience with reporting!
(If you're interested in more details on mixmobi, check out an article here, including a demo video.)
The second half of the event was an interactive discussion about the use of social media. People shared their frustrations about being social media evangelists within their company, and techniques for helping others realize the power of social media. There were many interesting points and discussions, but one stood out for me. If someone doubts the power of social media, or feels they can ignore it until they're ready to tackle it - take them to Twitter or Facebook and search on their company name. I imagine the results would be eye-opening to many.
Overall, an enjoyable, interesting, and thought-provoking morning. I look forward to attending more of these breakfasts in the future.
Showing posts with label professional. Show all posts
Showing posts with label professional. Show all posts
Tuesday, September 29, 2009
Monday, September 14, 2009
Good to Great - the book
I recently finished reading Good to Great: Why Some Companies Make the Leap...And Others Don't by Jim Collins. Collins is the author of an earlier book, Built to Last, which I haven't read, but this one is stand-alone (and, Collins considers it a prequel).
Good to Great is the result of about 5 years of research by a team of ~20 researchers. They first identified companies they felt had gone from 'good' to 'great', then identified comparison companies that had not made the transition, and then attempted to identify characteristics of the 'good to great' companies.
I would sum up the 'good to great' path as presented in this book as: Management hires the right people, creates a culture of discipline and focus on the core business and it competitive advantage, lets ideas percolate up to management, and leadership implements them, reviewing success and path frequently, and changing course as necessary.
THE PROCESS
The researchers first culled a list of ~1,500 public companies to find those that had gone from 'good' to 'great' by looking at performance over time (a period of average returns, followed by a transition period, followed by a period of above-average returns). (Of course, it was more difficult than that, for example, controlling for companies in industries that were experiencing industry-wide growth).
They ended with a list of eleven companies and comparison companies. Next was the arduous task of reviewing publicly available data on each of those companies - press releases, newspaper clippings, interview, and the like to understand what was going on within each of the companies during this time period. Did they have a change in leadership? Were they actively pursuing a new strategy?
THE RESULTS
Some sample characteristics of a 'good to great' company:
MY TAKE
As with many books in this genre, I was disappointed overall. Maybe I'm disappointed by the messiness of the real world or maybe I expect too much. But even with a sample size as small as eleven, I felt as though some of the characteristics were forced on some of the companies.
And, as is almost always the case, this book is not a how-to, it's a way of thinking. As the author states many times, the companies didn't realize they were in transition until it was complete. If this is so, can a company attempt to make the leap from good to great based on the path of the example companies? Does knowing that you're trying to transition impact the effectiveness of the transition? And even if it doesn't, implementing the steps will be difficult. How can a company determine what its competitive advantage is, what it can be the best at? Who are the right people?
GOOD IS THE ENEMY OF GREAT
The most powerful piece of the book, for me, was the statement 'good is the enemy of great'. I hadn't thought about it that way, but I find it very telling. It's just another way of saying that there's diminishing returns on the level of effort you put into something, but phrased that way I find it very powerful. How often do I settle for 'good' when I could have 'great'? And, when does it make sense to settle for 'good'?
CIRCUIT CITY, FANNIE MAE
I was quite skeptical of this book and its processes - perhaps because, reading it a few years after its publication, I have a vantage point that the author didn't. Two of the companies included in the eleven 'good to great' list no longer exist. Circuit City has gone bankrupt and Fannie Mae was bailed out by the government. Within less than 10 years of its 'greatness', these two companies got into serious trouble.
This certainly gives rise to some questions over the methods and results.
Good to Great is the result of about 5 years of research by a team of ~20 researchers. They first identified companies they felt had gone from 'good' to 'great', then identified comparison companies that had not made the transition, and then attempted to identify characteristics of the 'good to great' companies.
I would sum up the 'good to great' path as presented in this book as: Management hires the right people, creates a culture of discipline and focus on the core business and it competitive advantage, lets ideas percolate up to management, and leadership implements them, reviewing success and path frequently, and changing course as necessary.
THE PROCESS
The researchers first culled a list of ~1,500 public companies to find those that had gone from 'good' to 'great' by looking at performance over time (a period of average returns, followed by a transition period, followed by a period of above-average returns). (Of course, it was more difficult than that, for example, controlling for companies in industries that were experiencing industry-wide growth).
They ended with a list of eleven companies and comparison companies. Next was the arduous task of reviewing publicly available data on each of those companies - press releases, newspaper clippings, interview, and the like to understand what was going on within each of the companies during this time period. Did they have a change in leadership? Were they actively pursuing a new strategy?
THE RESULTS
Some sample characteristics of a 'good to great' company:
- led by 'level 5 leaders' - leaders who are humble, who think the company's success is more important than his own success, more 'plow horse' than 'show horse'.
- focused on getting the right people onboard and using those people to collectively find the path to greatness rather than dictating a strategy and using whoever's available to fulfill that strategy.
- determined what the company is passionate about and what the company can become the best at.
MY TAKE
As with many books in this genre, I was disappointed overall. Maybe I'm disappointed by the messiness of the real world or maybe I expect too much. But even with a sample size as small as eleven, I felt as though some of the characteristics were forced on some of the companies.
And, as is almost always the case, this book is not a how-to, it's a way of thinking. As the author states many times, the companies didn't realize they were in transition until it was complete. If this is so, can a company attempt to make the leap from good to great based on the path of the example companies? Does knowing that you're trying to transition impact the effectiveness of the transition? And even if it doesn't, implementing the steps will be difficult. How can a company determine what its competitive advantage is, what it can be the best at? Who are the right people?
GOOD IS THE ENEMY OF GREAT
The most powerful piece of the book, for me, was the statement 'good is the enemy of great'. I hadn't thought about it that way, but I find it very telling. It's just another way of saying that there's diminishing returns on the level of effort you put into something, but phrased that way I find it very powerful. How often do I settle for 'good' when I could have 'great'? And, when does it make sense to settle for 'good'?
CIRCUIT CITY, FANNIE MAE
I was quite skeptical of this book and its processes - perhaps because, reading it a few years after its publication, I have a vantage point that the author didn't. Two of the companies included in the eleven 'good to great' list no longer exist. Circuit City has gone bankrupt and Fannie Mae was bailed out by the government. Within less than 10 years of its 'greatness', these two companies got into serious trouble.
This certainly gives rise to some questions over the methods and results.
Thursday, August 20, 2009
Web Analytics Integration
Judah Phillips recently wrote a blog post over at Web Analytics Demystified - Web Analytics Integration: Holy Grail or White Whale. In it, he describes the transition he's seen in the industry over the past few years around integrating web analytics data with data from other sources. In addition, he shares his thoughts on some of the potential integration points.
I wholeheartedly agree that there is much potential in this area. Web analytics data is powerful on its own - but combine it with data from other sources, and it becomes even more so.
Integrating data from any two sources can be difficult. Years ago, my project team was tasked with implementing a customer-facing feature that involved making certain data warehouse data available to the customer. If you're familiar with data warehouses, you'll recognize that this opposite from how a data warehouse is typically used. (If you're not familiar, you can find more information on wikipedia, but the basic idea is that the site data is the data used in for day-to-day operations, often customer-facing. The data warehouse takes the site data then reformats it, aggregates it, etc. to enable internal operations, often reporting). My project team was very familiar with accessing and using site data - and our data warehouse counterparts were very familiar with accommodating the new site tables. Yet, pulling data warehouse data back into the site data was new to both teams, and involved new processes and relationships.
I wholeheartedly agree that there is much potential in this area. Web analytics data is powerful on its own - but combine it with data from other sources, and it becomes even more so.
Integrating data from any two sources can be difficult. Years ago, my project team was tasked with implementing a customer-facing feature that involved making certain data warehouse data available to the customer. If you're familiar with data warehouses, you'll recognize that this opposite from how a data warehouse is typically used. (If you're not familiar, you can find more information on wikipedia, but the basic idea is that the site data is the data used in for day-to-day operations, often customer-facing. The data warehouse takes the site data then reformats it, aggregates it, etc. to enable internal operations, often reporting). My project team was very familiar with accessing and using site data - and our data warehouse counterparts were very familiar with accommodating the new site tables. Yet, pulling data warehouse data back into the site data was new to both teams, and involved new processes and relationships.
If integrating data stored in similar technologies with similar data maintained by teams with a close relationship can be difficult, then integrating data from disparate data sets can be even more difficult. Web analytics data is often captured, stored, and reported on differently than other data. Given this, it often takes focus and resolve to spend the effort to make this possible.
But, the payoff can be tremendous. More recently I was on a project team tasked with building out a system to collect click-stream data. Even before the Proof of Concept was initiated, one of the key requirements was that the resulting system be able to integrate with some of the existing data stores. And that proved to be a prescient requirement. Even before work on that capability was completed, we were discovering analyses that would benefit from combining the click-stream data with data from other data sets.
In summary, I agree with the author that integrating web analytics data with other data can be very powerful and I am glad to hear that he has noticed a shift in the attitudes and availability of features around this!
But, the payoff can be tremendous. More recently I was on a project team tasked with building out a system to collect click-stream data. Even before the Proof of Concept was initiated, one of the key requirements was that the resulting system be able to integrate with some of the existing data stores. And that proved to be a prescient requirement. Even before work on that capability was completed, we were discovering analyses that would benefit from combining the click-stream data with data from other data sets.
In summary, I agree with the author that integrating web analytics data with other data can be very powerful and I am glad to hear that he has noticed a shift in the attitudes and availability of features around this!
Wednesday, August 19, 2009
Target to take its website inhouse
I've known that Amazon offered a service where it powered the back-end of a third-party's website. What I didn't realize was that large companies with a strong online presence were still using the service in 2009.
When I first heard about the service, I thought it was brilliant. It was the early 2000's, maybe 2002 or so - the consumer marketplace was at an inflection point: e-tailing was becoming more popular and less of a niche market. But what would the marketplace look like at the end of the transition? Would traditional retailers change shape and exist in both the offline and online space? Could traditional retailers thrive in an online setting? Or would the market be divided into 'traditional' retailers and 'online' retailers?
Add to this uncertainty that creating an online presence - specifically those that include product inventories or ordering capabilities - can require new skillsets, technology, and hardware with high start-up costs. Add to that the high propensity for failure and the short window in which to make an impression in a changing market, and I thought it made great sense for companies to outsource their e-commerce operations to Amazon. I thought the service would primarily be an interim solution until each company built out its own e-commerce capabilities.
I hadn't thought about this service in years, until I came across an article which stated that Target would be discontinuing its partnership with Amazon in 2011 and taking its website inhouse.
So for nearly 10 years, the target.com backend has been powered by Amazon. I was completely surprised by this. But upon closer consideration, I can see how my thinking was biased.
My experience has been primarily with companies whose primary priority are its websites. In my personal experience, companies are filled with people who are technical (or used to be technical) and there are teams dedicated to disaster recovery, data redundancy, performance, security, up-time, etc.
Looking at it through that lens, building and maintaining a quality website seemed, if not easy, certainly not difficult. Add some resources with specific front- and back-end skillsets, expand a few data centers, add resources to the on-call schedule, and the site is up and running.
But, I suppose, not all companies are like the ones I've experienced. Building out a basic html site may be easy enough, but building out and supporting the infrastructure to maintain a highly-available, integrated, redundant, secure ordering site is complex. And for companies whose core business is distinct from that effort, it may not make sense to bring all that effort inhouse.
While this news rated just a few paragraphs in the local paper, I found it very eye-opening. And good luck to Target in this endeavor!
When I first heard about the service, I thought it was brilliant. It was the early 2000's, maybe 2002 or so - the consumer marketplace was at an inflection point: e-tailing was becoming more popular and less of a niche market. But what would the marketplace look like at the end of the transition? Would traditional retailers change shape and exist in both the offline and online space? Could traditional retailers thrive in an online setting? Or would the market be divided into 'traditional' retailers and 'online' retailers?
Add to this uncertainty that creating an online presence - specifically those that include product inventories or ordering capabilities - can require new skillsets, technology, and hardware with high start-up costs. Add to that the high propensity for failure and the short window in which to make an impression in a changing market, and I thought it made great sense for companies to outsource their e-commerce operations to Amazon. I thought the service would primarily be an interim solution until each company built out its own e-commerce capabilities.
I hadn't thought about this service in years, until I came across an article which stated that Target would be discontinuing its partnership with Amazon in 2011 and taking its website inhouse.
So for nearly 10 years, the target.com backend has been powered by Amazon. I was completely surprised by this. But upon closer consideration, I can see how my thinking was biased.
My experience has been primarily with companies whose primary priority are its websites. In my personal experience, companies are filled with people who are technical (or used to be technical) and there are teams dedicated to disaster recovery, data redundancy, performance, security, up-time, etc.
Looking at it through that lens, building and maintaining a quality website seemed, if not easy, certainly not difficult. Add some resources with specific front- and back-end skillsets, expand a few data centers, add resources to the on-call schedule, and the site is up and running.
But, I suppose, not all companies are like the ones I've experienced. Building out a basic html site may be easy enough, but building out and supporting the infrastructure to maintain a highly-available, integrated, redundant, secure ordering site is complex. And for companies whose core business is distinct from that effort, it may not make sense to bring all that effort inhouse.
While this news rated just a few paragraphs in the local paper, I found it very eye-opening. And good luck to Target in this endeavor!
Saturday, July 25, 2009
Twitter for Business
Earlier this month, I wrote a blog entry that was, in part, my ideas on how a company could successfully use Twitter.
Yesterday, Twitter launched "Twitter 101 for Business". It includes some more basic information, such as how to get started on Twitter, Twitter lingo, and Twitter best practices.
It also includes some suggestions on how to use Twitter for business, including case studies of companies successfully using Twitter currently. So, did any of my suggestions for how businesses could use Twitter appear in the Twitter 101 guide? Yes!
A number of the ideas I suggested were among those listed in Twitter 101, including offering deals via Twitter, reaching out to customers and potential customers via Twitter, and monitoring talk about your brand via Twitter.
I think "Twitter 101 for Business" is a fantastic idea, and well executed. I suspect many companies know they could/should be using Twitter, but were intimidated by the concept or didn't know enough about Twitter to be able to formulate an idea of how to use it. This site lays out, in a straight-forward way, both how to use Twitter and ideas on how to make the best use of Twitter. The case studies offer good examples on how companies are currently using Twitter.
Twitter 101 may provide a spark for some companies to get on Twitter!
Yesterday, Twitter launched "Twitter 101 for Business". It includes some more basic information, such as how to get started on Twitter, Twitter lingo, and Twitter best practices.
It also includes some suggestions on how to use Twitter for business, including case studies of companies successfully using Twitter currently. So, did any of my suggestions for how businesses could use Twitter appear in the Twitter 101 guide? Yes!
A number of the ideas I suggested were among those listed in Twitter 101, including offering deals via Twitter, reaching out to customers and potential customers via Twitter, and monitoring talk about your brand via Twitter.
I think "Twitter 101 for Business" is a fantastic idea, and well executed. I suspect many companies know they could/should be using Twitter, but were intimidated by the concept or didn't know enough about Twitter to be able to formulate an idea of how to use it. This site lays out, in a straight-forward way, both how to use Twitter and ideas on how to make the best use of Twitter. The case studies offer good examples on how companies are currently using Twitter.
Twitter 101 may provide a spark for some companies to get on Twitter!
Tuesday, July 21, 2009
Best Buy's Twelpforce
Earlier this month, I wrote about how consumer companies could use Twitter to their benefit. A timely article, Best Buy has announced its 'Twelpforce', which is a Twitter Helpforce.
Twelpforce is essentially crowdsourcing CRM using Twitter, an interesting approach.
Best Buy has created @twelpforce, a centralized Twitter handle for Customer Relationship Management (CRM) on Twitter.
A complete description is available here, but essentially, any authorized user can have his tweets show up in his stream and the twelpforce stream by adding a particular hashtag to the tweet.
Instead of the stream being monitored and managed by a single employee or group, as is often the case with 'help desks', it's monitored and responded to by whichever Best Buy employees are interested in doing so. Any Best Buy employee (who has signed up and been verified, I assume) can reply as @twelpforce.
There are two ways Best Buy employees are encouraged to use Twitter.
But, I find the second a bit creepy. I can understand responding to tweets about the company itself, but if I wrote a funny/joking tweet about needing another tv, I would not want a response from a Best Buy employee offering assistance!
Overall, an interesting approach, and I look forward to seeing how it turns out and what other companies do in this space.
Twelpforce is essentially crowdsourcing CRM using Twitter, an interesting approach.
Best Buy has created @twelpforce, a centralized Twitter handle for Customer Relationship Management (CRM) on Twitter.
A complete description is available here, but essentially, any authorized user can have his tweets show up in his stream and the twelpforce stream by adding a particular hashtag to the tweet.
Instead of the stream being monitored and managed by a single employee or group, as is often the case with 'help desks', it's monitored and responded to by whichever Best Buy employees are interested in doing so. Any Best Buy employee (who has signed up and been verified, I assume) can reply as @twelpforce.
There are two ways Best Buy employees are encouraged to use Twitter.
- Reply to user questions directed to @twelpforce.
- Best Buy employees are encouraged to search for, and respond to, tweets of interest. As an example from Best Buy's 'tip' sheet, a user tweeted that he needed a new tv (but in a whiny/funny 'the wife is watching 'Desperate Housewives' again, I need another tv!'). A Best Buy employee responds, states that he works for Best Buy, and can answer any questions the guy has about tv's.
But, I find the second a bit creepy. I can understand responding to tweets about the company itself, but if I wrote a funny/joking tweet about needing another tv, I would not want a response from a Best Buy employee offering assistance!
Overall, an interesting approach, and I look forward to seeing how it turns out and what other companies do in this space.
Sunday, July 12, 2009
Monetizing Real Time Data
TechCrunch has posted the 'Top 10 List' to monetization in real-time data from Ron Conway, an angel investor in Silicon Valley.
After reading through the list, 'real-time data' seems to be a euphemism for Twitter, which is, I suppose the only real-time data set available to the average business, and the largest one available to most others.
For reference, here is the list, directly from the TechCrunch article:
I would break the list down in the following way:
After reading through the list, 'real-time data' seems to be a euphemism for Twitter, which is, I suppose the only real-time data set available to the average business, and the largest one available to most others.
For reference, here is the list, directly from the TechCrunch article:
10. Lead generation
9. Coupons
8. Analytics
7. CRM
6. Payments - If I was at PayPal, I would be looking at this.
5. Commerce
4. User authentication - Corporate accounts want to pay.
3. Syndication of new ads - Twitter itself could just syndicate. Multi-billion.
2. Content advertising and advertising context and display
1. Acquiring followers
I would break the list down in the following way:
- What can third-parties do right now to monetize on Twitter? Given the current state of Twitter, there are certain things a company could start doing right now. (in ascending order of importance, as ranked by me, numbered from the original list)
1. Followers - This is the first step - creating a brand identity on Twitter. There are many ways to go about this, from having a corporate-branded account that tweets official, corporate output, to having a visible person tweeting on behalf of a corporation.
7. CRM - Again, there are many possibilities, from creating an official mode of customer interaction to searching for and responding to mentions of the brand on Twitter.
9. Coupons -Posting special events or offers on Twitter that make it worthwhile for a user to follow your corporate brand on Twitter.
10. Lead Generation - Who's searching for your product (or similar), or appears to need your product? - What can Twitter do to take advantage of some of the items on this list? I think that, if Twitter wanted to, it could monetize similarly to the way Google does.(in ascending order of importance, as ranked by me, numbered from the original list)
3. Ads - Paid Advertising/SEM, and Natural Search/SEO, on the search engine side could easily be modified for Twitter. Results could display for searches or individual tweets. Twitter search could even be modified to something other than 'most recent tweet' first.
2. Contextual Advertising - Google has search history to determine user context, Twitter has history of user tweets (what do I find interesting), followers (who do I find interesting), etc.
4. User Authentication - Verified Accounts, which allow public figures to 'claim' an official account, were a huge step in the right direction. I'm having a difficult time imagining 'paid' accounts, but if the verified accounts do not work out, they may become important. Or, perhaps, if paid accounts were certain benefits, such as priority in twitter queues or some method of discreet advertising.
Friday, July 10, 2009
Freakonomics - the book
'Freakonomics' is one of those books I enjoy re-reading every few years. I first read it when it was initially published, then again when I started working in business intelligence, and again just recently.
The book asks, and attempts to answer, a series of questions either using publicly available data or by performing experiments.
Some of the questions the book takes up include:
If teachers were to cheat, how would they do it and how could you detect it?
My first thoughts were retesting the classrooms with unbiased observers in the room or looking for groups of students whose performance pattern looked something like low, high, low in three consecutive years or examination periods. But, assuming you wanted to identify suspicious classrooms without retesting all the classrooms, and assuming you had access to the actual answer sheets (which they did), you could look for patterns in the answers. I hadn't considered that, but of course, as soon as the book started heading in that direction, I jumped to it as well.
Steven Levitt is an economist with a sociology bent. Levitt teamed up with Stephen Dubner, a journalist, and the partnership seems to work well. The result is an easy-to-read series of anecdotes written in a conversational tone with references to the papers or articles that have more detail - good for someone looking for a quick read but also someone who wants to dig into more detail.
The book asks, and attempts to answer, a series of questions either using publicly available data or by performing experiments.
Some of the questions the book takes up include:
- Do teachers cheat for their students on standardized tests?
- Do real estate agents' and their clients' interests align?
- What was the cause for the decline in crime in the 90's?
- What attributes of the childhood household inform future success?
If teachers were to cheat, how would they do it and how could you detect it?
My first thoughts were retesting the classrooms with unbiased observers in the room or looking for groups of students whose performance pattern looked something like low, high, low in three consecutive years or examination periods. But, assuming you wanted to identify suspicious classrooms without retesting all the classrooms, and assuming you had access to the actual answer sheets (which they did), you could look for patterns in the answers. I hadn't considered that, but of course, as soon as the book started heading in that direction, I jumped to it as well.
Steven Levitt is an economist with a sociology bent. Levitt teamed up with Stephen Dubner, a journalist, and the partnership seems to work well. The result is an easy-to-read series of anecdotes written in a conversational tone with references to the papers or articles that have more detail - good for someone looking for a quick read but also someone who wants to dig into more detail.
Web Analytics - Puzzle or Mystery - the article
Recently I came across an old article from a popular Web Analytics blog, Occam's Razor that is as applicable today as when it was published in 2007.
In 2007, Malcolm Gladwell published an article in The New Yorker expanding on the concept of 'mysteries' and 'puzzles' as originated by Gregory Treverton, national-security expert. From Treverton's definitions:
The poster from the Web Analytics blog argues that web analytics is a mystery and I would largely agree. In my experience, there are few puzzles in web analytics, mostly around data collection and data definitions, although many data definitions are mysteries. Interpreting and acting on the resulting data are most certainly mysteries. Yet, I've come across many people, especially those new to web analytics, or even outside it, who think it will be a puzzle - "If we could only report on 'X', we'd know what to do". I think that explaining web analytics as a mystery instead of a puzzle would go a long way towards helping people understand web analytics.
In 2007, Malcolm Gladwell published an article in The New Yorker expanding on the concept of 'mysteries' and 'puzzles' as originated by Gregory Treverton, national-security expert. From Treverton's definitions:
- Puzzles - have a correct, factual answer which is unknown because we do not have enough information.
- Mysteries - do not have a single, correct answer. We have a lot of information and require knowledge and judgment to make an assessment of the answer.
The poster from the Web Analytics blog argues that web analytics is a mystery and I would largely agree. In my experience, there are few puzzles in web analytics, mostly around data collection and data definitions, although many data definitions are mysteries. Interpreting and acting on the resulting data are most certainly mysteries. Yet, I've come across many people, especially those new to web analytics, or even outside it, who think it will be a puzzle - "If we could only report on 'X', we'd know what to do". I think that explaining web analytics as a mystery instead of a puzzle would go a long way towards helping people understand web analytics.
Friday, July 3, 2009
Amazon's Mechanical Turk - Artificial Artificial Intelligence
Recently, I heard a piece on NPR about Amazon's 'Mechanical Turk' service. It is an online marketplace for freelance human tasks (referred to as HITs, Human Intelligence Tasks).
Requesters post individual, standalone tasks to be completed for a pre-determined compensation. Users complete those tasks, and, if the work is approved by the requester, will receive the stated compensation, plus, optionally, a bonus.
This is a concept that I've wished for at various times in my professional life, and I am especially heartened to see a well-known name like Amazon get into this area.
First, some interesting background - historically, the mechanical Turk was a chess-playing machine that was, in fact, controlled by a human. Thus, artificial artificial intelligence!
I see two main reasons for this type of offering:
And, I've already earned $0.28!
Requesters post individual, standalone tasks to be completed for a pre-determined compensation. Users complete those tasks, and, if the work is approved by the requester, will receive the stated compensation, plus, optionally, a bonus.
This is a concept that I've wished for at various times in my professional life, and I am especially heartened to see a well-known name like Amazon get into this area.
First, some interesting background - historically, the mechanical Turk was a chess-playing machine that was, in fact, controlled by a human. Thus, artificial artificial intelligence!
I see two main reasons for this type of offering:
- An extension of the consulting / offshore work model - Perhaps because I come from a consulting background, I have a natural inclination to blended teams, incorporating groups with different attributes
and cost structures to deliver while minimizing cost.
This is a workforce that is truly flexible - the contract extends as long as the individual HIT (minutes or hours) with no ongoing commitment from either party. If the project is canceled or its budget cut, simply remove the listing from the website. If a particular worker does poor work, reject his submissions.
This is a workforce that is truly flexible - the contract extends as long as the individual HIT (minutes or hours) with no ongoing commitment from either party. If the project is canceled or its budget cut, simply remove the listing from the website. If a particular worker does poor work, reject his submissions.
Obviously this type of workforce will only be applicable to certain types of tasks - easily subdivided, smaller, requiring no special skills or training, and easily validated. But, from my own professional career, I can think of numerous instances when we could have used a workforce like this. I remember in one meeting saying something like 'we just need 10 people to sit in a room for 2 weeks and do [something].' We didn't want to take any of our employees or consultants off their current assignments, and anyway there was no need to pay someone a consulting rate to do this type of work. We considered requesting interns, but it wasn't much of a learning opportunity. We needed exactly the type of workforce that MTurk offers. - Humans remain better at some tasks - Consider the seemingly ubiquitous example of extracting garbled text from an image and entering it in a text box - relatively easy for a human, but difficult to write a program to do the same.
At the same time, even processes that can be fully automated generally require a human in the feedback loop before getting to that state. Human writes program, program provides output, human validates output, human modifies program, and so on.
Often, that feedback loop involves only the developer/tester, but in many systems, especially those that are self-learning or intelligent, and those in which the possible answer set is large or the possible answers unknown, the amount of validation may be very large.
- Do advertisements for certain flagged search keywords fall under any restricted product/service types?
- Audio transcription
- Translation
- Do user-entered images meet certain requirements?
- Find information from various sources and consolidate into a consistent format.
- Bookmark site X on digg/StumbleUpon, follow user Y on Twitter
- Write a review of website X on your blog
- Rewrite this sentence in your own words (must be statistically 50% different than original) (which I assume will be used to post multiple posts on a site that doesn't allow duplicates)
- Sign up as a subscriber on site X
- Fill out a survey online (which then took you to an advertisement and dropped a cookie)
And, I've already earned $0.28!
Wednesday, June 24, 2009
Flip Flop Flyball - the website
Came across an interesting site earlier today on baseball infographics.
The site combines two of my interests - baseball and using data in interesting ways, so I find it really enjoyable.
One of the entries from March of this year shows the direction the batter faces in each of the major league parks. I doubt it would help anyone win a world series, but something interesting to consider, and beautifully displayed.
It's not the heavy world of sabermetrics, but a lighter look at random baseball data - ticket prices, retired numbers, Kevin Costner baseball movies, etc.
I am especially impressed by two things. First, the author is using available data in interesting ways, to answer interesting questions. This is something I look to do, both in my professional and personal lives, and seeing other people doing it always inspires me to do more. Second, the author does a fantastic job with the visual display of the data. The visual display accurately depicts the data and enhances the viewer's understanding of what is being conveyed.
All in all, a fun little site!
The site combines two of my interests - baseball and using data in interesting ways, so I find it really enjoyable.
One of the entries from March of this year shows the direction the batter faces in each of the major league parks. I doubt it would help anyone win a world series, but something interesting to consider, and beautifully displayed.
It's not the heavy world of sabermetrics, but a lighter look at random baseball data - ticket prices, retired numbers, Kevin Costner baseball movies, etc.
I am especially impressed by two things. First, the author is using available data in interesting ways, to answer interesting questions. This is something I look to do, both in my professional and personal lives, and seeing other people doing it always inspires me to do more. Second, the author does a fantastic job with the visual display of the data. The visual display accurately depicts the data and enhances the viewer's understanding of what is being conveyed.
All in all, a fun little site!
Friday, June 19, 2009
Golf Statistics Article
I grew up in a golfing family, so I've been playing and watching golf since I was a little kid.
For most of that time, I've been tracking my golf outings. Always score, of course, but often putts, penalties, sand shots, clubs used, and whether a shot felt good. This season, I have been tracking the distance each shot goes using a gps device, by club, hole, and lie.
The data isn't very useful, given the big number of variables in a round of golf, and the small number of rounds I play each year. But, it has been interesting to look at, and has helped me finetune my play in small ways. (as an example, earlier in the season I was having a problem with three-putts. After looking through some rounds, I realized my medium and short length putts seemed fine, but my long putts were consistently leaving me with medium length putts instead of short putts.).
So, I was extremely excited to see an article in Slate about golf statistics!
From the article, I learned that the PGA employs ShotLink to record each shot made during its tournaments. With that wealth of information, there are many applications. For example, individual players can determine how they rank in types of shots compared to their peers to determine where to focus their practice efforts. They have also determined that if the hole were twice the size, poor putters would benefit more than good putters.
Mark Broadie, a golf researcher at Columbia, presented at the 2008 World Scientific Congress of Golf. A related paper, with much more detailed analyses is located here.
I would love to get my hands on the ShotLink data - there are so many interesting questions waiting to be answered!
For most of that time, I've been tracking my golf outings. Always score, of course, but often putts, penalties, sand shots, clubs used, and whether a shot felt good. This season, I have been tracking the distance each shot goes using a gps device, by club, hole, and lie.
The data isn't very useful, given the big number of variables in a round of golf, and the small number of rounds I play each year. But, it has been interesting to look at, and has helped me finetune my play in small ways. (as an example, earlier in the season I was having a problem with three-putts. After looking through some rounds, I realized my medium and short length putts seemed fine, but my long putts were consistently leaving me with medium length putts instead of short putts.).
So, I was extremely excited to see an article in Slate about golf statistics!
From the article, I learned that the PGA employs ShotLink to record each shot made during its tournaments. With that wealth of information, there are many applications. For example, individual players can determine how they rank in types of shots compared to their peers to determine where to focus their practice efforts. They have also determined that if the hole were twice the size, poor putters would benefit more than good putters.
Mark Broadie, a golf researcher at Columbia, presented at the 2008 World Scientific Congress of Golf. A related paper, with much more detailed analyses is located here.
I would love to get my hands on the ShotLink data - there are so many interesting questions waiting to be answered!
Tuesday, June 2, 2009
Bing, WolframAlpha, and Google
As an internet user, you're no doubt aware that google dominates the search market. If you have worked in SEO(search engine optimization), website traffic attribution, or web analytics, you may know more about how the search traffic breaks out. (Based on my professional experience, I distinguish google, yahoo, and Microsoft from 'other' search engines. If you'd like more details, comScore's April 2009 Search Engine Rankings is available.)
This week saw a flurry of tech announcements, including a new search engine (Wolfram|Alpha), and an updated/rebranded site (Microsoft's Bing).
First, bing.com, an updated/rebranded release of Microsoft's Live Search. Bing calls itself a 'decision engine', not a 'search engine', but the differences appear subtle to me. As far as I can tell, the launch is largely to get people talking about Microsoft search and to give themselves a cooler name that can be 'verbed up'.
Second, Wolfram|Alpha, a 'computational knowledge engine'. From the FAQ,
I tried out a couple different types of searches to compare the results.
Search for a city:
Search for me:
Search for stock symbol, date, etc.:
Winner: Wolfram|Alpha
To my mind, Wolfram|Alpha is a fascinating intellectual achievement, and I will certainly use it for specific types of information. Bing, on the other hand, I found less interesting, and for the bulk of my search I will likely continue to use google.
This week saw a flurry of tech announcements, including a new search engine (Wolfram|Alpha), and an updated/rebranded site (Microsoft's Bing).
First, bing.com, an updated/rebranded release of Microsoft's Live Search. Bing calls itself a 'decision engine', not a 'search engine', but the differences appear subtle to me. As far as I can tell, the launch is largely to get people talking about Microsoft search and to give themselves a cooler name that can be 'verbed up'.
Second, Wolfram|Alpha, a 'computational knowledge engine'. From the FAQ,
It's a computational knowledge engine: it generates output by doing computations from its own internal knowledge base, instead of searching the web and returning links.The code was built with Mathematica and brings back memories of my college computer science classes! Wolfram|Alpha is not a replacement for google (or bing, or yahoo). For certain things, this type of computation makes sense and returns relevant information. For others, it does not even attempt to return results. In fact, Wolfram|Alpha results page offers a 'Search the web' option, which takes you to a google search with your initial search term.
I tried out a couple different types of searches to compare the results.
Search for a city:
- bing - returns the city website, then separate sections for hotels, restaurants, maps, newspapers, and jobs.
- google - returns a map and a list of links, including the city website, hotel chains, and a local college.
- Wolfram|Alpha - provides population, current time, nearby cities.
Search for me:
- bing - I know I'm not a celebrity, but I was disappointed in the results - the first results page includes my grandfather's 2002 obituary and 5 separate entries for my amazon profile.
- google - variety of results, mostly what I expected
- Wolfram|Alpha - no results
Search for stock symbol, date, etc.:
Winner: Wolfram|Alpha
To my mind, Wolfram|Alpha is a fascinating intellectual achievement, and I will certainly use it for specific types of information. Bing, on the other hand, I found less interesting, and for the bulk of my search I will likely continue to use google.
Saturday, May 30, 2009
Forrester Web Analytics Forecast
Forrester recently released its 2008 to 2014 US Web Analytics Forecast. From the executive summary
Although I didn't pay for the report, the author wrote an entry about it on his blog and there is an article over at readwriteweb, including some charts from the forecast.
And as someone who has spent time in web analytics at a company with a major online presence, I absolutely believe this statement is true (from the same readwriteweb article)
Forrester forecasts that US businesses will spend $953 million dollars on Web analytics software in 2014, with an average compound annual growth rate of 17%.Given that I believe so strongly in the power of data-driven analysis and specifically web analytics, I'm excited by this report.
Although I didn't pay for the report, the author wrote an entry about it on his blog and there is an article over at readwriteweb, including some charts from the forecast.
And as someone who has spent time in web analytics at a company with a major online presence, I absolutely believe this statement is true (from the same readwriteweb article)
Forrester also offers the prediction that, given the availability and affordability of web analytics data, a secondary market will spring up to offer services showing companies opportunities for using said data, thus bridging the "action chasm" between knowledge and execution.I only wish I were able to read the full article!
Friday, May 29, 2009
Facebook Apps Analytics via Omniture
Yesterday, Omniture "announced App Measurement for Facebook, a new solution that enables Omniture SiteCatalyst customers to measure the popularity and success of Facebook applications." (from the Press Release) (screenshots)
In March, Omniture announced Twitter analytics via its SiteCatalyst tool. (Press Release. I wrote a quick blurb here).
As a strong believer in the power of data and analytics, seeing a company with an established following making headway in the social media space is exciting.
In March, Omniture announced Twitter analytics via its SiteCatalyst tool. (Press Release. I wrote a quick blurb here).
As a strong believer in the power of data and analytics, seeing a company with an established following making headway in the social media space is exciting.
Thursday, May 21, 2009
I.B.M. Unveils Real-Time Software to Find Trends in Vast Data Sets - the Article
There was an interesting article in yesterday's Wall Street Journal about a new IBM software, to track and analyze data in real-time.
This sounds similar to a technology I have used in the past, although IBM seems to have built out the system to learn and find correlations in the data without guidance.
Real-time data analysis is often expensive compared to near-real-time or offline processing. Additionally, with complicated data, it would be a sophisticated and well-designed system to anticipate and correctly handle all data. So, a system that is self-learning would be advantageous.
I would be very interested in finding out more about this technology and its related product, System S. I was not able to find much information online, unfortunately.
This sounds similar to a technology I have used in the past, although IBM seems to have built out the system to learn and find correlations in the data without guidance.
Real-time data analysis is often expensive compared to near-real-time or offline processing. Additionally, with complicated data, it would be a sophisticated and well-designed system to anticipate and correctly handle all data. So, a system that is self-learning would be advantageous.
I would be very interested in finding out more about this technology and its related product, System S. I was not able to find much information online, unfortunately.
Tuesday, March 24, 2009
Recruiting via Twitter
I came upon an interesting article earlier today about companies using Twitter to post job openings.
I think it's an interesting concept, and a natural extension of networking and social networking sites, but I found the application, twitterjobsearch.com, a bit difficult to use.
As far as I can tell, you can filter by location and job type. Searching on 'Minnesota' got some targeted results, but also results from Atlanta, Houston, and the East Coast.
What I found more difficult was that the tweets weren't in a consistent format and many of them were missing information I'd find valuable - namely job title and location. Combined with the seemingly ineffective search, I think you'd spend a lot of time looking at postings that don't interest you.
An interesting idea that will likely improve over time.
I think it's an interesting concept, and a natural extension of networking and social networking sites, but I found the application, twitterjobsearch.com, a bit difficult to use.
As far as I can tell, you can filter by location and job type. Searching on 'Minnesota' got some targeted results, but also results from Atlanta, Houston, and the East Coast.
What I found more difficult was that the tweets weren't in a consistent format and many of them were missing information I'd find valuable - namely job title and location. Combined with the seemingly ineffective search, I think you'd spend a lot of time looking at postings that don't interest you.
An interesting idea that will likely improve over time.
Friday, March 20, 2009
Differentiating Online Ads
Recently, Google started offering behavioral ads in addition to its contextual ads. There was an interesting article in the NYTimes blog section this morning about it.
Google allows you to opt out of its behavioral ads, and also provides generalized information on why you're getting specific ads. The author of the blog suggests going further, to display exactly why the ad is displayed and why it's displayed the way it is. For example, whether the image, text, or price displayed are variable, and if so, exactly what information about you was used to determine its value.
What I found most interesting about the article was a quote from a Google representative which suggests that few consumers understand the difference between contextual and behavioral ads.
It's simple (or it can be, at least). Contextual targeting uses information about what you are currently doing. For example, search for 'Prius repair' and you'll get ads about local repair shops. Search 'Prius 2009', and you'll get ads about new Priuses. Or in an offline sense, an ad for sugary kids cereals during Saturday morning cartoons.
Behavioral targeting uses information about what you have been doing. For example, if you've been browsing for new Priuses, you may see ads for cars even when you're browsing a news site. In an offline sense, this could be like if you go to the same waitress every Saturday morning, and one week she offers you a coupon for being a loyal customer.
Of course, there's a difference between online and offline. Most people fully understand, or *could* understand offline collection methods. But 'online' is a black-box to many. I think the author's idea is an interesting one, if unlikely to come to fruition any time soon. I would probably use it, but most wouldn't. I think the most benefecial aspect would be one mentioned in the article, that companies would behave more ethically or risk public press because of watchdog agencies and the media having access to the information.
Google allows you to opt out of its behavioral ads, and also provides generalized information on why you're getting specific ads. The author of the blog suggests going further, to display exactly why the ad is displayed and why it's displayed the way it is. For example, whether the image, text, or price displayed are variable, and if so, exactly what information about you was used to determine its value.
What I found most interesting about the article was a quote from a Google representative which suggests that few consumers understand the difference between contextual and behavioral ads.
It's simple (or it can be, at least). Contextual targeting uses information about what you are currently doing. For example, search for 'Prius repair' and you'll get ads about local repair shops. Search 'Prius 2009', and you'll get ads about new Priuses. Or in an offline sense, an ad for sugary kids cereals during Saturday morning cartoons.
Behavioral targeting uses information about what you have been doing. For example, if you've been browsing for new Priuses, you may see ads for cars even when you're browsing a news site. In an offline sense, this could be like if you go to the same waitress every Saturday morning, and one week she offers you a coupon for being a loyal customer.
Of course, there's a difference between online and offline. Most people fully understand, or *could* understand offline collection methods. But 'online' is a black-box to many. I think the author's idea is an interesting one, if unlikely to come to fruition any time soon. I would probably use it, but most wouldn't. I think the most benefecial aspect would be one mentioned in the article, that companies would behave more ethically or risk public press because of watchdog agencies and the media having access to the information.
Thursday, March 5, 2009
Twitter Analytics via Omniture
In addition to its usual web analytics, Omniture has started offering Twitter analytics, which I think will be a powerful addition to a company's business intelligence.
The Twitter analytics space is young, but with an experienced brand like Omniture joining the fray, I expect interesting things.
The Twitter analytics space is young, but with an experienced brand like Omniture joining the fray, I expect interesting things.
Tuesday, March 3, 2009
Blink - the book
I was really disappointed with "The Tipping Point' by Malcolm Gladwell but had heard good things about "Blink: The Power of Thinking without Thinking", so decided to check it out. I am glad I did.
The basic premise is that our intuition, our snap judgments are often correct, even when we can't explain why or how we came to that decision.
Some of the book was anecdotes. A fake piece of artwork that curators spent months investigating and felt was real, while art critics knew instantly that something was 'off' about it. A 'fantastic' artist who every music manager instinctively loves, but who doesn't play well to focus groups. I don't like anecdotes. They can too easily be cherrypicked.
The author does use some more scientific observations, which I really enjoyed. A researcher who videotapes short interactions with couples and is able to predict which will get divorced based on very short segments, focusing on traits such as defensiveness and flexibility. A study that showed that it took participants drawing 50-80 cards to recognize that red cards caused losses and blue cards caused profits, yet their body 'knew' and started reacting within 10 cards.
As usual, I would have liked more proven studies and less anecdotes, but compared to other books of its type, I was pretty impressed.
The basic premise is that our intuition, our snap judgments are often correct, even when we can't explain why or how we came to that decision.
Some of the book was anecdotes. A fake piece of artwork that curators spent months investigating and felt was real, while art critics knew instantly that something was 'off' about it. A 'fantastic' artist who every music manager instinctively loves, but who doesn't play well to focus groups. I don't like anecdotes. They can too easily be cherrypicked.
The author does use some more scientific observations, which I really enjoyed. A researcher who videotapes short interactions with couples and is able to predict which will get divorced based on very short segments, focusing on traits such as defensiveness and flexibility. A study that showed that it took participants drawing 50-80 cards to recognize that red cards caused losses and blue cards caused profits, yet their body 'knew' and started reacting within 10 cards.
As usual, I would have liked more proven studies and less anecdotes, but compared to other books of its type, I was pretty impressed.
Subscribe to:
Posts (Atom)
