<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>2Shanghi - Doug Sibley's on Moving to Shanghai &amp; Other Thoughts - Latest Comments</title><link>http://2shanghai.disqus.com/</link><description></description><atom:link href="https://2shanghai.disqus.com/comments.rss" rel="self"></atom:link><language>en</language><lastBuildDate>Sun, 11 Apr 2010 09:43:44 -0000</lastBuildDate><item><title>Re: (Micro-)Outsourcing Everything</title><link>http://blog.dougsibley.com/?p=49#comment-44319632</link><description>&lt;p&gt;Building lists from scratch is an order of magnitude harder in crowdsourcing than augmenting an existing one.  So, it requires a bit more handholding by the requestor.  Having said that, here's one basic approach.  &lt;a href="http://www.smartsheet.com/blog/brent-frei/self-building-lists" rel="nofollow noopener" target="_blank" title="http://www.smartsheet.com/blog/brent-frei/self-building-lists"&gt;http://www.smartsheet.com/b...&lt;/a&gt;&lt;/p&gt;&lt;p&gt;There are several others that work well depending on the list being built, for example, creating a psuedo-seedlist also can work well:  &lt;a href="http://brentfrei.typepad.com/blog/2009/06/list-of-fly-fishing-stores-in-wa-state.html" rel="nofollow noopener" target="_blank" title="http://brentfrei.typepad.com/blog/2009/06/list-of-fly-fishing-stores-in-wa-state.html"&gt;http://brentfrei.typepad.co...&lt;/a&gt;&lt;/p&gt;&lt;p&gt;-Brent&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Brent Frei</dc:creator><pubDate>Sun, 11 Apr 2010 09:43:44 -0000</pubDate></item><item><title>Re: (Micro-)Outsourcing Everything</title><link>http://blog.dougsibley.com/?p=49#comment-44290706</link><description>&lt;p&gt;Hmm, Brent, I've been reading a bit more about you're ability to making from-scratch lists with smartsheet but I don't see a post on walking through the mechanics of it.  I'd like lists of unique items (like all the English language magazines focused on China) but I don't want to pay for the top 5 magazines 50 times each and miss the bottom 200.  Is there an easy way to do this in smartsheet?&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dougsibley</dc:creator><pubDate>Sun, 11 Apr 2010 00:05:35 -0000</pubDate></item><item><title>Re: (Micro-)Outsourcing Everything</title><link>http://blog.dougsibley.com/?p=49#comment-44289494</link><description>&lt;p&gt;Brent,&lt;/p&gt;&lt;p&gt;Thanks for your detailed reply.  I suppose you could eliminate the steps for the virtual worker but then again you could do the same for the intern; in my test I simply followed the demo and asked for some bits of information on five companies.&lt;/p&gt;&lt;p&gt;You mention that the tasks range in complexity but even at five cents though, the allotted time is 20 seconds (in a $10/hour example) which is not a lot of time to read pricing information (let alone visit a few web pages).&lt;/p&gt;&lt;p&gt;I gave smartsourcing a try last night and unfortunately it's still a bit rough around the edges.  I gave examples of five companies asking for the CEO's full name (not the parent company's if applicable), Industry (pick list), and website URL.  One of five was right, another two were close, and the final two were just copy-pastes from the other one (so for four of five rows the CEO's name was identical - all should have been different).&lt;/p&gt;&lt;p&gt;I really wanted this to work better than it did.  I also really wanted to be able to use it to generate lists (e.g. all the English print magazines in Hong Kong).  Alas, I think it's back to the intern route for me.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dougsibley</dc:creator><pubDate>Sat, 10 Apr 2010 23:43:55 -0000</pubDate></item><item><title>Re: (Micro-)Outsourcing Everything</title><link>http://blog.dougsibley.com/?p=49#comment-44197594</link><description>&lt;p&gt;Doug,&lt;/p&gt;&lt;p&gt;Regarding your question about the math involved.  The answer is sort of buried in the article and one of the comments following it.&lt;/p&gt;&lt;p&gt;A very simple metaphor would be this:&lt;br&gt;Our intern and the virtual worker differ in that the intern needed to open a drawer, pull out a flash card, record what was important on it, and then put it back and repeat.   The virtual worker simply sits with his/her hands on the key board and types as someone else holds cards up for them at whatever pace they like.&lt;/p&gt;&lt;p&gt;When you do run a few tests on Smartsourcing, you'll begin to see how efficiently the technology involved in the whole process eliminates wasted motion.&lt;/p&gt;&lt;p&gt;The intern was handed a sheet of company names.  So the steps for each were something like:&lt;br&gt;- Copy the company name and paste it into a Google search box&lt;br&gt;- choose the company website from the results list&lt;br&gt;- copy the url and paste it into sheet&lt;br&gt;- find the company description page on the website and determine the appropriate 2 sentence description, then copy and paste it&lt;br&gt;- find the products page and...  so on&lt;/p&gt;&lt;p&gt;By pre-building some concatenated Google search urls with the company name, and for each subsequent step, you can eliminate several steps for the virtual worker and put them right where they need to be to do the work from the onset.&lt;/p&gt;&lt;p&gt;Per the article:  "...the 4 tasks per company are a roll-up summary of the actual task process. Given the nature of the data being asked for, some parts were easier than other parts. "Does this company offer a GL module?" was dead simple and fast, whereas "Give an average price paid for the product" had varying degrees of difficulty in coming to an answer. So, in some of the task categories, we asked more than one worker to answer the same question and averaged the answers. The price paid per task ranged from $0.01 to $0.05 depending on the difficulty. Therefore, the overall average time to complete and cost per task are a rollup of what was actually varied task prices and more than 600 tasks."&lt;br&gt;&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Brent Frei</dc:creator><pubDate>Sat, 10 Apr 2010 08:56:38 -0000</pubDate></item></channel></rss>