(52)
There has always been a lot of discussion around the topic of conversion rates and if that impacts your Google rankings. Of course, the correlation versus causation item is on the top of ones mind when discussing such topics.
But I wanted to share with you a WebmasterWorld thread that tested this, obviously without being to isolate all the variables, on his own site.
- Dropped his prices on his site
- It led to a 2-fold increase in conversion rates from 2.89% to 5.68%
- Also led to a bounce rate drop from 39% to 26%
- Also more time spent on the site from 203 seconds to 290 seconds
Shortly after, he noticed his rankings spiked up higher in Google. Here is the breakdown he shared:
- widget (EMD): Friday: #97, today: #56
- widget product: Friday: #127, today: #39
- california widget: Friday: #12, today: #5
- widgets: Friday: #145, today: #82
He said, "It's almost crazy to think that we could be seeing any results this quick from this change but we have made no other changes to the site and we have done no promotion or link building since April, 2012." Could it be that Google is able to track this? Note, he does not use Google Analytics on his site.
Forum discussion at WebmasterWorld.
Image credit to BigStockPhoto for funnel
Comments:
Courtney Cox
07/24/2013 01:24 pm
It seems logical that he's seeing a spike in traffic because people are buying more often, which means they don't have to return to the search engine to visit another result. In that way, his conversion rate indirectly increased his rankings.
Chase Anderson
07/24/2013 01:56 pm
I've become more and more convinced that one of Google's quality measurements (especially for Panda) has become bounce to SERP. The challenge with this metric is that it would need to be compared directly to other competitors in the same SERP. The calculation needed to make this metric work could be pretty massive (explaining it's need to be calculated on a ~monthly process). Ultimately, a few days to see results, brings to question exactly how would Google might have changed Panda since they announced it will now be continually rolled out. I initially assumed that Google may have started to run the Panda calculation on separate spaces and released updates for those spaces in real time, something that might take days as opposed to months since the whole internet doesn't need to be analyzed to complete the process. Although it could also be that Google's previous stored values of bounce races for specific competitors may be used as a comparison at any time. If a site's bounce to the SERP is reduced, it may have a more immediate effect without the need to recalculate the entire SERP or internet. Very loosely thought out analysis, and without much data to go on. But if I were Google I'd be looking for this type of measurement to help answer the question of "Does this site work for people searching with this term".
Paddy Displays
07/24/2013 02:16 pm
he might also have earned some natural links by having cheap prices ( cheap price is link bait too)
Uri Lederman
07/24/2013 02:31 pm
There is no question that "time on site" & "Bounce Rate" correlate to quality site in Google's eyes.. but dropping his prices as the key ingredient to all this???? I don't know about that?? sounds like there was more too it.. :-)
adi
07/24/2013 02:35 pm
ummm that makes no sense at all. What are you trying to say?
adi
07/24/2013 02:38 pm
conversion rate no, time on site no, bounce rate possibly. 'Time on site' is subjective' . You could say that decreasing time on site is a good thing, as visitors are finding the information quicker due to an intuitive site structure. Conversely, more time on site could mean they are engaging more.. How would google know which one is better?
Courtney Cox
07/24/2013 02:38 pm
I'm saying that Google knows that if someone views your site from the SERP and has to return to the SERP to find another site that they had a bad experience. If visitors don't return to the SERP, they can assume they found what they needed. This is an indication that your site is better, and thus gives Google incentive to give it higher rankings, resulting in higher traffic.
Michael Martinez
07/24/2013 02:49 pm
Of course, other people could have changed their sites -- and rankings are not the same for everyone so ranking reports are useless and meaningless fluff.
Durant Imboden
07/24/2013 03:20 pm
"I've become more and more convinced that one of Google's quality measurements (especially for Panda) has become bounce to SERP. The challenge with this metric is that it would need to be compared directly to other competitors in the same SERP." There's another problem with that metric: People bounce from sites and look for other search results for a variety of reasons. Our site, for example, is geared to people who are seeking in-depth information and enjoy reading (i.e., "Wikipedia people" as opposed to "Twitter people"). If 75 percent of searchers are "Twitter people," our bounce rate is going to look bad, even if the remaining 25 percent (the "Wikipedia people") stick around because we provide more comprehensive information than most of our competitors do. IMHO, using "bounce back to the search results" as a quality metric would make sense only if Google's personalization scheme took user behavior ("Is this guy a Wikipedia-type person or a Twitter-type person?") into account.
Chase Anderson
07/24/2013 03:31 pm
Think of it this way. You're competitors also receive both type of people. Do more people performing the query stick on your competitor? Even if you're the wiki and they're the twitter type content, if more stick with them, they're the better result. As a matter of balance, twitter people may be more opt to search one way, were wiki people search slightly differently. For example, homes for sale in Denver vs buy a home in Denver. One search will have one type of audience, the other will be slightly different. This same principle would work for 'pogo-stick' people also. These are those who click every result in the query and click through the SERP looking for just the right result. They're hurting everyone's bounce rate equally, unless they find what they want.
Chase Anderson
07/24/2013 03:34 pm
Change the color of your call you action button, it'll mean all the difference in the world. It's not at all unreasonable for price changes to drive these user metric changes.
Chase Anderson
07/24/2013 03:36 pm
That's obviously possible. But it's about as likely as someone at Google visiting the site, buying something, thinking the price point was so awesome that they then artificially boosted the site's rankings.
Gordon Campbell
07/24/2013 03:49 pm
Dropped prices may look like the content has been refreshed and could be short lived positive impact. Also, more time on site and more people buying from you means less return to SERP(NOT bounce rate) to search again which is a ranking factor...
DaveKeys
07/24/2013 04:33 pm
This topic is almost a red herring. We have a high tide to low tide adjustment in the Panda algorithm and probably a few aftershocks still happening from Penguin. No, the house did not fill up with 16 inches of water because you flushed the toilet. There is a flood outside!
sestuff
07/24/2013 04:49 pm
That person should have understood your first comment because it was clear enough. As for the metric, Yandex uses something called MatrixNet which sounds a lot like panda. Bing also mentioned using the bounce to SERP metric. Funny Google named the update after a Google engineer when they should have named it Yandex LOL.
sestuff
07/24/2013 04:53 pm
Exactly. I prefer SER because Barry gets straight to the point even though SEL is more in depth. IMO it doesn't mean that SER is better but if more people prefer SER when they search for a particular key term as opposed to SEL, SER should rank better for that term because it's the people's choice. Maybe not what is technically better or better written but what is better at solving the problem.
Barry Schwartz
07/24/2013 05:01 pm
Aww, I am blushing. :)
jimbeetle
07/24/2013 05:18 pm
Matt Cutts has consistently said that bounce rate would be a problematic metric and just a couple of weeks ago Google's John Mueller reaffirmed that Google does not use bounce rate: http://www.youtube.com/watch?v=__8fGDmBxW0&feature=youtu.be&t=44m50s
sestuff
07/24/2013 05:27 pm
They said they do not use bounce rate but that's different from bounce to SERP. Danny Sullivan once asked this specific question and Matt Cutts tried to flip things by saying that they don't look at bounce rate. Well Danny then repeated the question and said that he is not talking about bounce rate but bounce to SERP and Matt didn't deny it. I think he said that they may use it at some point.
Durant Imboden
07/24/2013 05:34 pm
"IMO it doesn't mean that SER is better but if more people prefer SER when they search for a particular key term as opposed to SEL, SER should rank better for that term because it's the people's choice." No, what it means is that Google Personalization should be able to tell if you're the kind of person who's likely to prefer SER's "quick meal" approach or SEL's "five-course dinner" approach. Also, the "people's choice" isn't always the right choice. More people buy the New York Daily News than The New York Times, but if I want comprehensive coverage and analysis, the Times is going to be more useful to me than the Daily News even if the latter is more popular. I suspect that, over time, Google Personalization will get better at determining what kind of content is likely to be preferred by a given searcher--and in what context. (Even a Ph.D. may be addicted to Top 10 lists, and even a New York Daily News, Mirror, or Bild reader may want comprehensive information when researching a serious topic.)
josh bachynski (SEO)
07/24/2013 05:36 pm
What was his percentage of Chrome, FF, PR Toolbar and logged in users, i wonder... ;p
jimbeetle
07/24/2013 05:40 pm
Thanks for that, I have to scout around and try to dig that up.
sestuff
07/24/2013 05:44 pm
Understand your point and you're probably right that personalization should play a bigger role but from what I can understand, you still rank for those queries so it's not like you have been knocked out of the SERPs completely - you're just not #1. My point does make sense though. Look at the SERPs as you do at the top 10 music charts - there is only one #1. Your point about more people buying the daily news isn't a good example because money makes the world go round. Why are movies that sell more tickets ranked higher than movies that sell less? Of course, I never said that bounce to SERP is the #1 factor but since we were talking about this topic, I simply stated that it would make sense to rank sites based on this formula. I have sites outranking me that suck and I know their bounce isn't better than mine so it's not the deciding factor in the SERPs. Authority and a bunch of other factors play a role as well. In your case, Wikipedia is likely outranking you due to authority.
jeje
07/24/2013 05:49 pm
If, and that's one big "if", these type of signals are indeed incorporated into the algorithm, this is not the type of things that can cause a huge spike in traffic, or removal of penalties/sand-boxes. Google just doesn't work on that sort of a constant basis. Some sites that deserve obvious penalties are around for weeks or months until they're spotted, while other fair quality sites that have done everything to recover from Penguin have still not received the redemption letter. So that thread teaches us nothing.
Chase Anderson
07/24/2013 06:04 pm
Thanks for the link but he says "We don't use anything like the analytics bounce rate". That's not to say they couldn't use a "Bounce to SERP" or similar metric. The difficulty with this metric he mentioned is that different queries have different bounce rate expectancies. So Google would have to look at bounce rates query by query and compare the results, accounting for positional influences. Doesn't this sound like something Google'd need to update ~monthly? Insane amounts of processing power and time would be needed to look at serp bounce rates in this way
Chase Anderson
07/24/2013 06:07 pm
Exactly, I believe this is where Google wants to be, but given the massive massive complications with this level of personalization, it's probably not close. It's also something that would probably freak people out more than they already are with Google privacy. Most people that notice Gmail ads targeted based on their email content causes them to pause.
Wheel from WMW
07/24/2013 06:21 pm
There is a flux as Google test which SERPs produces more ad clicks, this tells us nothing. Especially since you're talking about "improvement" to page page 6++
Durant Imboden
07/24/2013 06:28 pm
This discussion isn't about me, and in any case, I outrank Wikipedia for quite a few queries. I'd also point out that there's a big difference between a top 10 music chart ("Here are the biggest-selling albums") and a search-engine results page. Getting back to the topic of this subthread: "Bounce to SERP." One problem with using that as a metric is simply that some pages are digested more quickly than others are. Let's say that John Doe is researching the latest Panda update. He finds pages for SER and SEL in the Google results, and there's a good chance that he'll want to look at both. There's also a good chance that he'll bounce back to the SERP a lot faster from SER than from SEL, simply because Barry's posts tend to be short. Does that mean users are less satisfied with SER posts than with SEL articles? Maybe, maybe not. It depends on the user, which is why bounce rate or even bounce to SERP is a "noisy" signal.
Michael Martinez
07/24/2013 08:41 pm
Based on the information provided, no one knows what is likely, less likely, or more likely. This is just another case of people drawing conclusions on the basis of ignorance.
Michael Martinez
07/24/2013 08:49 pm
What Matt actually said, at both SMX Advanced 2012 and 2013, is :"To the best of my knowledge we don't use bounce rate to affect rankings".
Courtney Cox
07/24/2013 08:50 pm
Lol! Thanks, I thought so, too.
Chase Anderson
07/24/2013 09:14 pm
While nothing is certain until it's certain, it's not at all unreasonable to draw up hypothesis based on the best information available. We didn't visit space or circumnavigate the globe until well after it became accepted that the earth was a sphere. Assuming that a rank change from 97 to 56 was entirely caused by 41 webmasters making changes to cause their sites to decrease in ranks isn't very likely. (obviously this is an exaggeration of your previous guess). Moving from 12th to 5th however, isn't going to be due to 7 competitor's making simultaneous site changes, it wouldn't be logical to decide that was the most likely scenario causing the rank changes. It's not hard to mentally calculate reasonable and unreasonable causes when your basis is in reality. Realistic analysis of the situation saves a webmaster time spent investigating such a rank change. It's so unlikely that other webmaster changes caused his site to jump, so he doesn't need to give that theory any further thought unless all of the more likely scenarios have been eliminated. Do you actually perform SEO with such an existential view? SEO's operate on massive foundations of theory and uncertain principles - if you scoff at the dispersion of your "other webmasters" theory how do you ever measure success or failure when everything is so mystical and could be skewed by butterflies flapping their wings in Mexico?
Chase Anderson
07/24/2013 09:29 pm
Indeed, could be.
sestuff
07/24/2013 10:08 pm
I'm not a computer to remember things word for word lol.
sestuff
07/24/2013 11:15 pm
My last comment got deleted. Point I was making. Top 10 music is a very good comparison IMO and it's a lot alike. Only difference is one involves sales and the other involves clicks but both lead you to the product, but I won't drag this along - as you said, let's get on topic. Metrics can be measured very easily and it's not all that noisy when other factors come into the equation. Example, if you click back to the SERPs in the blink of an eye, how can you consume any info? It still takes longer than a second to read Barry's article. Using time on site is a good way to measure bounce to SERP. When measured on a large scale, it's easy to block out noise because it answers the question: do the majority of people remain on SEL or on SER? Does the user end up going back to the site after looking at all other sites? Yesterday I bought something online and bounced on a bunch of sites and even initially bounced from the site that I ended up buying from. IMO it was a good sign that I eventually went back to that site. On page activity is also a good signal. What does the person do on the site in those few seconds? Do they engage? Google can see what people are doing better than anyone. Well Bing is also capable since they own IE. In the end, seeing is one thing and the question still remains whether both can actually interpret the signals flawlessly which is why we are debating over this.
Google Scams
07/25/2013 01:52 am
stop listening to this nonsense, I went from #240 to #70! Wow look at me. That means almost zero visitors anyway. Look at reports from 2011 to last year and see how Google is killing our traffic, year after year. Google is all scams now to cut off our traffic and force us to advertise. Where do you think Google got 20+% more clicks on ads this year? By having Amit Singhal the crook manipulate search results and penalizing ours.
Arpit Agarwal
07/25/2013 04:19 am
Yes, There is too much fluctuation in Google ranking day by day. I noticed that one day any keywords goes up and another day it goes down. I can't understand about the process of Google to give ranking to any particular keyword. SEO is now become hassle for me. I am thinking to quit it, if this game is continued in future.
Vijay Sharma
07/25/2013 04:54 am
Its show that time on site is one of the big factors in ranking.
Soni Sharma
07/25/2013 05:42 am
That is all we call it engagement. higher time on site, low bounce rate traffic from multiple destinations will lead increase in ranking automatically.
Dave
07/25/2013 05:47 am
How a site that is barely ranking in top 30+ is getting traffic from search engine? Organic Traffic drops tremendously if you drop down from top 3 position, even in top 7-10 position you get very few traffic. How come he has increased his conversion ratio. Is he talking about conversion through Paid search that helped him to gain ranking? Or, it's only just the Google dance that used to happened 2-3 times in a year but now the dancing happens daily or with a refresh you get hell new sites and few old site vanish in ranking. Next day or week those new sites will be replaced by another bunch of new low quality sites that don't have a contact page to contact them. In the end user will be forced to click on Ads, as people who are spending on ads will definitely have contact page for sure There is no stability in Google ranking. It looks like they are testing on money keywords how bad organic search result will lead to click to their Ads.
somebody
07/25/2013 10:24 am
agree. they not spammers, they are scammers! Just guy find a money cow. if company owners not fire up him and mr cutts, money cow will dead soon.because of short-sighted management. adwords already not convert well both for webmasters and advertisers. Untargetted ads, clicks displayed and disappears, lot of clicks in $$$ niches for 0.01, etc.
boba
07/25/2013 10:29 am
google can think by another way. May be they just analyze toolbar and chrome users, their action on site. +social signals, comments, etc. Anyway, they make stupid rotator which will never stop rotate sites. Top seats reserved only for youtube/wikipedia/yahoo/amazon sites + for internet coalition
hohla
07/25/2013 10:32 am
just drop price for new bmw model to $1, and tomorrow you will have top10 in google for all cars-related queries. This is what they understood under high quality site.
asa
07/25/2013 10:35 am
it only mean what google not using this metric from analytics. why they need to do it, if toolbar and google chrome spy on surfers?
Chase Anderson
07/25/2013 01:43 pm
Even though Google, with toolbar and chrome data, would see that data set as too small, and demographically not a great representation of their entire search population. If there is one thing you can be relatively sure of, it's that Google would demand the best data, even 95% good wouldn't be good enough. Chrome, Android and Toolbar don't represent 95% of searches, not even close. They must use the data from the search itself.
Michael Martinez
07/25/2013 01:55 pm
I'm not either. I did a Google search for the quote, found it on multiple sources, and decided not to provide a link.
mark
07/25/2013 02:17 pm
If he isn't talking about paid search then how would Google know his conversion rates if he does not have analytic's in his site?
sestuff
07/25/2013 07:38 pm
Yeah I know. I didn't want to do a Google search ;) As for your quote, I was looking at a different one where Matt responded to the bounce to serp metric with something like: no we don't use analytics data. Which doesn't really provide an accurate response to: do you use bounce to serp.
Michael Martinez
07/25/2013 09:02 pm
I think it's accurate enough. The only bounce rate they would be able to measure for every Website in their index is bounce-to-SERP. So, as far as Matt Cutts knows (and he asked the search team about this at one point), they don't use (any) bounce rate in determining rankings. People who want to hold out hope for some sort of bouncing metric being used can always speculate that bounce-to-SERP might be one of the signals that Panda (or even Penguin) looks at -- but I see no way to confirm such a speculative guess.
Michael Merritt
07/25/2013 09:26 pm
I think Google must have some way of tracking at least some proportion of click-thrus directly from the SERP itself. If you've ever noticed the URL bar when you right click on a link, it changes from the destination site's URL to a Google URL, essentially routing you to the site through Google. It usually happens in the blink of an eye, but can occasionally take some time. I think this routing is partially to prevent malware infection, but I have to think they're also collecting some data on click-thrus, and perhaps setting some cookies. Then if the person bounces back and clicks another site or refines their search, they know the person was dissatisfied with the specific result.
davidfricks
07/26/2013 12:47 am
There was an update yesterday (Wednesday 24th) around 2pm PST. When did you see the change?
Uri Lederman
08/02/2013 09:48 pm
I hear what you are saying... but from my testing... a reduced price does not make you click a button... if you are looking to buy... this would make you click the button... not a reduction in price... this person seems to have an eCommerce store selling physical products.. Again.. just my testing, but reducing the price on physical products WILL NOT double your conversation... maybe giving it away...