Google Update Speculation: What Happened on November 14?

MozCast is a website that monitors changes in Google’s rankings. Activity for each day is tracked as a temperature reading. There was a temperature spike on November 15, and no signs of a glitch. The fact this appeared across multiple IPs and a look at MozCast’s 30-day history indicates something happened.

There has not yet been a confirmation of an update from Google, and chatter among webmasters was not any different than usual. Other flux-tracking tools, however, also showed one-day spikes in activity, including SERPmetrics and SERPs.com. Since the different tools exhibit measurement differences, secondary metrics are useful in identifying large-scale trends. On November 15, shifts search results for the top 10 domains, that day, jumped significantly.

The same day, there was a jump of 3.2%, quite substantial from a historical perspective. This was seen in the total search results representing sites such as Amazon, Wikipedia, and eBay. Interpreting these data can be challenging. Algorithm updates are often beneficial to larger brands, but introducing penalties for low-quality sites can simply eliminate them and leave room for others. One position, at #10 for example, drops out, and the next one down fills it and is more likely to be a larger site.

So far, the trends appear to point to an unusual one-day change and a historically high value for the “Big 10”. The values are based on data compared since April 2012. No causes are currently evident as to why such increases occurred that day.

Experts looked into keywords, but saw no clear patterns to determine if there was an update. Wikipedia picked up top 10 listings, but the big domains seemed to gain traction because other sites fell out of their positions. On November 14, a search for “famous footwear store hours” resulted in FamousFootwear.com being in the top seven spots. A shuffle was seen on November 15, however, as the main site took up just the top four spots. Wikipedia picked up a newly opened slot as did others.

Investigating the Possibility of a Glitch

Google reported a DNS error on November 14. While such issues are not uncommon, many webmasters and SEO experts also had a similar problem. Experts have speculated as to whether a bug could actually cause rankings to fluctuate widely. Such a bug occurred in April in which many domains were misclassified, causing noticeable changes to rankings. One would see a reversal in the case of a bug, appearing as a secondary spike a day or so later. As with the recent spike, there has been no pattern reversal.

Tracking daily ranking changes is difficult. One example illustrates how dynamic the Google search algorithm now is – the story in which a woman, claiming to be the daughter of Andy Kaufman, said the late comedian is still alive. The day before captured several informational sources in the top ten, showing Google viewed the search as generic. After the news broke, well-known news sources such as CNN, Wikipedia, Fox News, US Today, and NY Daily News appeared.

A concept known as Query Deserve Freshness factored in. This means a static query changed quickly because of new information. The newly available data and content impacted the search results as a factor of Google determining value in real time. This probably did not cause the overall shift, but single events show how dynamic daily queries can be.

Identifying the Cause

Rankings fluctuations are hard to explain. The only evidence to look at it is typically the results; one has to measure positions, their changes, and what domains or URLs are listed. Using this method, domain diversity can be tracked. Profiling a major update is difficult because changes in rankings are primarily a result of the sites, and their link profile characteristics, in question.

Natural fluctuations should be considered as well. The holiday season normally brings shifts in competition. Events such as Black Friday cause fluctuations in commercial SERPs too. What measuring the November 14-15 shift does is show a lack of roll back. The pattern looks like an algorithm update, without the hard facts to prove it.