I’m not sure that anyone’s noticed, but it’s been a few months since we’ve published our monthly jobs report, and my blogging in general has dropped precipitously of late. One of my goals for 2011 is to begin blogging again and I’m starting with a much delayed jobs report from LinkUp for November and December of 2010.
The reason for the delay in publishing our jobs data is that we essentially overhauled the manner by which we were generating our monthly jobs report. As we’ve articulated in the past, LinkUp is a job search engine that indexes only jobs that are found on company websites. There are no job board listings included in our search engine, and we do not allow companies to post directly to the site. Rather, we index job listings every day from over 20,000 company websites throughout the U.S. and publish those job openings on LinkUp.com. As a result of this unique approach, we have completely eliminated much of the garbage that pollutes job aggregator sites such as Indeed and Simplyhired that publish job scams, phishing jobs, work-at-home scams, and other bogus and fraudulent job listings from thousands of job boards. In addition, because we only publish job listings from a single source (the employer’s website itself), there are no duplicate job listings that are so prevalent on Indeed and Simplyhired.
In any event, in July of 2010, we began an initiative to greatly accelerate the rate by which we were adding new companies to the LinkUp index. The effort has been very successful and we’re very pleased to report that the size of the index has increased substantially and continues to grow at a very rapid clip. One of the unfortunate consequences of the effort, however, was that our monthly jobs data was becoming less and less reliable as we added more and more companies to the search engine.
This seems a little counter-intuitive, because a larger database of jobs from more companies should have increased the reliability of the data, and in many ways, it did. But as a result of some very specific anomalies in the way that certain companies publish their job listings, the data was also being skewed in some strange ways.
As an example, there are companies that change the URLs for every one of their jobs every day, even if the jobs have previously appeared on their corporate site. The unique URL for a job is the primary means by which we arrive at a total job count for the month as well as track whether or not a job is counted as new. Job postings from companies that change URLs every day, therefore, were always showing up in our data as new listings and were skewing both our new and total job count. Luckily, only a small percentage of companies change URLs for their jobs every day, but as we added more and more companies in the past 6 months, that small percentage was beginning to have a meaningful impact on the monthly jobs data that we report.
In November, we decided that we needed to address that issue and a few others that were negatively impacting our data. We stopped publishing our monthly jobs report and took the time to thoroughly analyze what was going on with our jobs data. Over the course of the past two months, we have implemented a series of modifications and the result, we feel, is a much improved monthly jobs report. Once we had improved the accuracy and integrity of the data, we went back and ran the reports for November and December.
In November, new jobs on company websites throughout the U.S. dropped by nearly 50,000 or 9% from October, and total jobs dropped by 2,000, essentially flat from October. For the month, 41 of 50 states showed a decline in new job listings on company sites, and 28 of 50 states showed a decline in total job listings.
In terms of job listings by category, new job listings on company sites dropped 10%, with 27 of 32 categories showing a decline. Total job listings remained unchanged from October, yet 22 of 32 categories showing a decline in total jobs. Healthcare was really the only category to show a meaningful increase in both new and total job listings.
In December, the jobs picture didn’t improve much at all, with new job listings on company sites falling by another 38,810, an 8% drop from November. Total job listings on company websites fell by 15,353, or 2% from the prior month. Similar to November, 40 states showed declines in new job listings, and 37 states showed a decline in total job listings.
In terms of jobs by category, new job listings on company websites fell by 9% in December, with 25 of 32 categories showing a drop in new job listings. Total job listings by category fell 2%, with 21 of 32 categories reporting a decline in jobs.
So with the benefit of hindsight, how did our data compare with what the Department of Labor’s Bureau of Labor Statistics (BLS) reported? In November, the consensus estimate from economists was that the U.S. economy added 150,000 jobs in November. ADP’s report was equally as optimistic, as their data showed an increase of 93,000 jobs in the private sector. Unfortunately, the BLS reported that unemployment rose to 9.8%, with the private sector adding only 39,000 jobs. It was a surprisingly negative report, but very consistent with our jobs data for the month.
In December, economists were again expecting a gain of 150,000 jobs for the month, and ADP’s report showed a significant increase of 297,000 jobs in December. Unfortunately, however, the Department of Labor reported that the U.S. economy added only 103,000 jobs in December, a far weaker number than expected. While unemployment fell to 9.4%, the drop was due entirely to the growing number of people who indicated, for a variety reasons, that they had stopped looking for work. So like November, our jobs data showed a far weaker labor market than what ADP and economists were forecasting, and once again, our data provided a pretty accurate picture of what actually happened during the month.
LinkUp will release its January jobs report on Wednesday, February 2nd, two days prior to the release of the Department of Labor’s report scheduled for Friday, February 4th.