Continued Volatility Shows Google Still Tweaking Algorithm
This post was co-written by Nathan Safran and Assaf Karmon, a lead engineer on Conductor’s Searchlight platform.
On February 24th Google announced an algorithm update designed to “find more high-quality sites in search”:
Our goal is simple: to give people the most relevant answers to their queries as quickly as possible. This requires constant tuning of our algorithms, as new content—both good and bad—comes online all the time.
To gain a better understanding of how the change impacted the SERPs Conductor studied the search visibility for more than 2,000 top ranking ehow keywords(position 1-5). Using Searchlight, Conductor’s SEO Platform, we tracked week over week rank movement, starting from the week of January 30th, to establish a baseline in the SERPs in the weeks prior to Google’s update, and expose changes occurring after the algorithm rollout. Although we restricted the study to a single content farmer, inferences can likely be drawn to similar content farms who, like ehow, have a mix of relevant and non-relevant content.
Looking first at volatility in the SERPs, we found an increase in the time period after Google announced their algorithm change on Feb 23rd, as the percentage of keywords with rank movement increased from 27% to 36%.
Increased Average Rank Movement:
Although the number of keywords that moved on Feb 20th, as described in the previous chart, was similar to previous weeks (nearly identical, in fact, to two weeks prior) the keywords that did move did so far more than previous weeks. On average, there was a 4 position drop on Feb 20th, several days before the official Google announcement on Feb. 23rd, presumably as they began to rollout changes. On Feb. 27th we saw a partial recovery from the previous week’s decline as the keywords moved up by nearly 2 positions. This suggests that Google is continuing to tune the algorithm changes over time.
Average Movement Hides Large Changes by Smaller Group
Digging deeper into the rank movement, in the weeks leading up to the algorithm changes, virtually no keywords moved 3 positions or more (up or down), while on Feb 20th nearly 7% of keywords dropped by three ranks or more. However, the following week (Feb 27), was a ‘recovery week’ as half of the terms that had dropped the previous week by three positions or more regained their rankings. Overall, 17% of keywords rose by 3 positions or more and the percentage of keywords that dropped by 3 positions or more returned considerably closer to pre-change levels.
Significant Movement for Those Keywords Impacted by Google Changes
Looking at the movement of keywords that moved by 3 positions or more week over week, in the weeks leading up to the change(Feb 6 and 13) they held steady with an average movement of 3 positions. On Feb. 20th,the average drop increased to 74 positions, 24x more than in previous weeks. The following week the group experienced something of a recovery by increasing 17 positions.
Of those keywords that dropped, 27% did not recover at all or in any significant way. Interestingly, looking at those that did drop by 3 positions or more on Feb. 20th, we found that they recovered even more than the overall group average, with an upward movement of 26 positions the following week.
Analysis Shows Fall Followed by Partial Recovery
Ultimately, the totality of the analysis above shows that at the time of the algorithm update, nearly 7% of keywords analyzed dropped by 3 positions or more, nearly three-fourths of them then recovered, while 27% did not. The week over week volatility also seems to suggest there is some ongoing tuning that Google is continuing to make with the algorithm and it will be important for all SEOs to keep an eye on this going forward. We’ll keep an eye on this and update with new developments.
Keywords That Fell Weak on Relevance
To flesh out characteristics of pages that dropped versus those that did not, we looked at several pages for keywords whose rankings fell and recovered, and compared them to pages that fell and did not recover.
This ehow page was ranking in position 4 on Feb. 13 for the term ‘New Years eve table decorations’, yet dropped entirely off the rankings map the following week, and did not subsequently recover on Feb. 27th.
The searcher is looking for fairly specific information–‘New Years eve table decorations’–while the ehow page was more generically aimed at ‘How to Decorate for New Year’s Eve’. The content was fairly thin too, and clearly specific to general new years eve celebration tactics rather than the specific table decorations the searcher was looking for.
Here is ‘step 3’ of the process describing how to decorate for New Years eve from the page–clearly it has little to do with actual New Years Eve table decorations:
3. Hire a balloon company to install a balloon drop to be released at midnight. There are stores that sell do it yourself drop kits, but they are generally small and often don’t release correctly. It’s best to leave this kind of thing to a professional.
Contrast that with this ehow page that was ranking in position 5 for the same query on Feb. 13th and moved up to position 4 on Feb 27th. The heading at the top of the page is “Table Decoration Ideas for New Year’s” and the content seems richer and clearly more closely linked to actual searcher intent:
Martini Glass Table Decorations
A martini glass centerpiece is both versatile and reusable for most other celebrations. For New Year’s Eve-appropriate table decorations, get some clear, oversized martini glasses. Choose a color scheme like blue and silver, and purchase some round ornaments according to the color scheme for your party. Fill the oversized glasses with the ornaments, and use some bubble wrap, foam, basket fillers, star garland, ribbon or glitter to add more pizazz to the decorations.
This theme—a drop in the rankings for pages ranking for queries only loosely connected to searcher intent–is one we saw persist in our analysis of pages that dropped. This seems to suggest that Google’s algorithm update is indeed a step forward in providing users with more relevant and targeted search results (although there have also been reports on blogs and across twitter that it may be casting too wide a net and penalizing arbitrarily in certain cases). And, as described above, the findings seem to indicate the algorithm change penalizes on a per page basis—if the content is not relevant for the query the page drops–rather than across an entire domain.
So whats this all mean for the SEO professional?
- Don’t Panic: If your site has been affected by the change, remember SEO is a long term endeavor. Make sure you understand what is really happening on your site before you start making any major changes. Technology will help you determine if there is an issue with one or two keywords versus a site wide issue, and PPC can be used as a short term gap filler as you evaluate the extent of the impact.
- Break Out the Good Old Fashioned Elbow Grease: Low tech tools should be paired with high tech tools to gain a full picture of what is occurring on your site. Spend some time with your web browser and do a relevancy audit to determine if there is a common theme for the keywords/pages that have dropped in rank (as described above in the ehow thin content example). Once you have that figured out, you can move forward on a plan to fill in content holes as needed and start the move back up the rankings.
- Plan for the Future: If history is any indicator, this algorithm change will not be the last one to hit the search engines. No SEO likes getting surprise calls from the boss asking why online sales are down; an SEO platform will alert you when rank drops occur giving you the visibility you need to determine when the landscape has changed. And, the trend with many of the last major algorithm changes has been to place an increased value on quality content (see Mayday change) so if you are not already devoting time and energy to enhancing your site content you probably should be.
Let us know in the comments if you are seeing something different on your site.
Senior Research Analyst