1. Direct Keyword Traffic
The graph on the left (I) shows the client's rankings, observed weekly, for their primary keyword over a period of about 7 months. The right-hand graph (II) shows weekly search traffic for that same keyword:
A couple of notes: (1) this is a one-word keyword, (2) it's relatively common/popular, and (3) rankings over the 7 months fluctuated from 1st to 5th (I’ve inverted the graph to show 1st at the top). I'm a big believer in eyeballing data first, and I think the graphs show some degree of connection between ranking and direct traffic. The visual is a bit more compelling than the correlation (r = -0.31), although there is some relationship. The negative correlation is expected and reflects the fact that a low ranking is better (1 > 5, value-wise).
2. Long-tail Organic Traffic
From a broader SEO standpoint, though, how did the ranking of the client's most important keyword affect their overall search traffic? I've repeated the ranking graph (I) below, but added a graph (III) of weekly traffic from Google organic results over the 7 months:
Here, the relationship seems to fall apart (r = 0.21). Other than a late-summer dip, overall organic traffic from Google actually grew as their primary keyword rankings were falling.
What Does It All Mean?
So, do rankings matter or don't they? Well, in terms of direct traffic for the specific keyword that suffered the ranking drop, there’s certainly some effect. Studies have shown pretty reliably that search visitors focus (and click) most on the Top 3 results, with #1 getting the lion’s share. Outside of the single-keyword view, though, the situation gets a lot more complex. While the client's primary keyword is an obvious choice for their industry and is fairly competitive, the traffic for that keyword accounted for only 1.3% of their total organic traffic from Google over the 7 months. If we had chosen to take a very narrow view, and obsessed over their primary keyword, we would have neglected over 98% of their SEO results.Of course, I'm not saying rankings are irrelevant, just that you need to take a broader view, including:
- Tracking diverse, long-tail phrases
- Tracking search phrases actually used by visitors (not what you think they use)
- Measuring search analytics, such as total traffic from search
Interesting data. Certainly a single-point focus on rankings doesn't do any good.
But to really answer the question, "How Much Do Rankings Matter?" there are other things to consider. Namely, pinning rankings to traffic to conversions/revenue (or your suitable value metric) tells the real story.
A client of mine has a single #1 ranking that brings in roughly $3k of revenue per month. Based on estimated traffic share for that keyword, a fall in one position to #2 would drop revenue to a little under $2k. That's a real, instant impact on the bottom line.
However, for this client 65% of their revenue comes in through long tail rankings.
Rankings always matter - a single keyword shift, however, has to be measured against the gross result of SERP shifts like the "brand" Google update. What was the overall effect on traffic, and more importantly, revenues/conversions?
Don't fret over a few lost battles if you're winning the war.
No argument there. I don't want to oversimplify, other than to drive home a point, and that point is that you have to look at the big picture. What matters for any given site is unique, and what rankings drive traffic and sales vary wildly between niches and audiences.
It's interesting to note, though, that I've seen cases (individual, granted) where #2-3 actually outperformed #1. For instance, I had a client who accidentally stumbled upon that observation, first on PPC but eventually even in organics. It took a while, but we finally concluded that their primary competitor's site was so hard to use that people were clicking through to them, bouncing back to the SERPs, and then actually converting better on the client's site because of their initial bad experience.
Of course, I'm not suggesting that anyone purposely try to rank 3rd or rest on the fact that the competition sucks - it's just interesting food for thought, IMO.
That is definitely interesting. I've seen that in terms of ROI - especially in PPC where the cost of bidding up to the 1st position can destroy the profit margin. But never before with an organic ranking where #2-3 actually worked out better than #1. Normally the better ROI of a lesser position is largely a function of the saved costs of getting to #1.
Ties back into the need to look at the big picture and avoid snap judgments/action based on ranking shifts...
Note: @SEOmoz Devs, how come editing a comment often kills all paragraph separation and requires a second edit to repair formatting?
I've definitely found with PPC that you often get a better conversion rate for lower positions that at the top, for a whole variety of campaigns. Adding in the extra cost of #1, as MikeTek points out, and you've ended up with a less than stellar ROI (although having said all that, it varies from case to case, obviously).
My theory as to why is that people become more discerning as they drop down the list. By the time they've hit #4, they've already seen a few sites, they're shopping around for the good deal or trustworthy site, etc. A big percentage of people who hit #1 are probably just casual browsers.
That's an interesting interpretation. I think sometimes people just give up - rationally, they would stop, compare #4 to 1-3 and pick the best one, but instead they just decide they're done and, if #4 is decent, that's where they buy. Tracking search user behavior (whether organic or PPC) is fascinating - people bounce in and out of sites like crazy.
Admittedly, PPC is a bit different, as the direct cost of being #1 is ongoing and very tangible. Getting to #1 in organic SEO certainly has a cost, but maintaining it is often cheaper than for PPC.
True, and I do stress that my own experience is only of PPC in this context.
Another idea (and this is completely speculative) is that sites higher up in the PPC rankings may be poorer quality. People being obsessed with being at #1 is often part of a pattern of behaviour of making money by driving a lot of traffic at a poor proposition, or having a poorly constructed campaign. The top positions on PPC campaigns are often taken by companies targetting terms that are too generic, etc, whereas lower down they may be better targetted and run by people making better use of their budgets by targetting a wider range of more targetted phrases.
My own anecdotal evidence of using the adverts as a consumer leads me to think that sites lower down the results are either a bit better or a bit more specifically targetted at what I want. But only as a very, very general rule.
It's great to see an example of the direct relationship between position and traffic (and as you say, the visual examples are much more pertinent).
I agree that it's important that you look at the long tail, but depending on your conversion criteria, your key, high priority keywords (i.e. those you have tested and know convert well) have to take priority. Obviously, if you've got an advertising driven model and only care about raw visitor numbers, that's not relevant, but in most cases it's very important.
Next I'd love to see a correlation graph between PR or mR and traffic! ;)
While I purposely picked the title to draw a little controversy, I want to make it clear that I'm not suggesting that rankings don't matter, just that you have to look at the big picture. Too many people are still fixated on 1-2 keywords and are missing the boat, not only in terms of things that might be going wrong but in areas they could improve and compete on.
I do find one interesting change has happened over time, though. In the old days, if you ranked #1 for a single-word keyword (let's say "computers"), you had a good chance of ranking well for other keyphrases that had "computer" in them. Not only was the long-tail shorter (people weren't used to typing 3+ word queries), but the algos were much less sophisticated. These days, ranking for that single-word keyword is far from a guarantee of ranking for other phrases that include that term. Each phrase is often a separate battle.
Interesting about that change. I've found with clients that targetting their main phrase does bring quite a good long tail with it... but then again, as a rule, I'm usually targetting less competitive terms and generally not a single term keywords.
Just a follow up - ran some data based on Greg Moore's comment below. Interestingly, traffic for all the phrases containing the primary keyword (singular and plural) correlated highly (r = 0.78) with traffic for the keyword alone. I have to admit, my impressions about that trend may be wrong.
There are, admittedly, a few keyphrases that are highly competitive by themselves and we have to do separate SEO for, but ranking well for our primary keyword still carries a fair amount of weight for related keyphrases, according to the data. Of course, that's still less than 20% of the overall search traffic, but it's certainly significant
Great post Dr. Pete,
I would argue to Mike Tek's point as well. While traffic from a rankings perspective may not fluctuate that much between the #1 and #3 spot; does the revenue or goals for the site change?
If the goal is to harvest emails or a certain number of downloads and the ranking changes that amount towards the negative, then yes rankings do matter.
I would also argue the point of quality traffic. Assuming the user typed in exactly what they wanted to see and the website was optimized for that keyword phrase (longtail or not) and displayed in the #1 position. The user will get what they want and so will the SEO - rankings matter.
We all know that the further down the list the less clicks; but why? Attention span? Change in search terms?
I say relevancy. Users visit the top terms on the first page because they know that the further they scroll down the worse the results get. A forced user habit. The only people who search down pages of SERPS are the SEO's who are trying to get their sites ranked :)
I've now officially written more words in comments than the original post :)
Just want to make it clear that I'm not bashing rankings in general - only saying that it's unhealthy to obsess over rankings for 1-2 keywords without looking at the big picture. I think you and Mike are absolutely right about making sure your results are in line with your goals.
One specific area where I think a fixation gets unhealthy is actually when a client's main keywords of interest are the wrong ones. It's very hard to get someone to let go of obsessing over a #1 ranking, even if it's the wrong keyword. Sometimes, if you can show that they're missing the boat (that another keyword or focus would be much more relevant) with non-ranking metrics, the convincing gets a lot easier.
As for user habits, the eye-tracking data certainly backs you up. People have been conditioned to think that the higher results are more relevant. What all the data doesn't tell us, though, is what happens when those results aren't relevant? What happens when people bounce back and start going down the list, which happens all the time. The data on 1-3 being the "best" is certainly interesting, but it's also very simplistic. Real search behavior is incredibly complex, I find.
Regarding the lack of effect on the "longtail" traffic, your analysis assumes that overall long-tail search volume attached to the market was constant before and after the drop in the 1 keyword ranking... Is it not?
The longtail rankings could have dropped as well, but the market search volume spiked and you are now in a lower position that is getting 3.5% CTR of 150k instead of 3.75% of 125k.
Obviously some businesses are more seasonal than others but I haven't worked in any space where there wasn't any seasonality effects at all. Throw in social media effects and the recent economy and there is enough uncertainty to drive anyone crazy.
I enjoyed the analysis, but there seems to be too many moving parts to come to the conclusion that if the head drops the tail remains constant...
I would agree that the tail is more important than it every was though... That's in line with having great unique(specifi) content.
Cheers
I love the hard-core analysis I'm getting in some of these replies - you guys aren't pulling any punches :)
No argument there - I'm not trying to oversimplify and say this is absolutely conclusive, just show one case where ranking for what we think is the most important keyword for a client site is really just a small part of the big picture, and even as rankings for one keyword fall, other metrics might be unaffected or even increase. There are no doubt other influences going on, including seasonal factors. In many ways, I think that's one more reason why an obsession over just 1-2 rankings can blind us to what's really going on.
I Agree Dr. Pete...
I feel your pain when dealing with clients... Most tend to believe the answer to all of their problems is if they can just somehow weasle their way into the #1 spot for some industry specific or infactuated keyword (phrase). This is about all they understand when it comes to analytics and SEO - so they try to simplify their goals (to themselves) so they can somehow feel comfortable with measuring us (and paying us).
It's a struggle...
" if the head drops the tail remains constant... "
Ive been sitting on 2 graphs of long tail traffic for a month for 2 websites just because I REFUSE to believe they are true and im trying to figure out whats going on
It looks like for both sites Google has set a hard limit on the number of visits a day from non-brand terms of 5000 and 3000 each
the 3000 a day one has had massive structural, design, optimisation, you-name-it changes over the course of the year but there is a year long straight line at 3000/day from long tail
I thought I was going crazy cause this data doesnt make sense to me atm, but this would seem to support that the tail can remain constant through major changes
That's bizarre - you mean your organic traffic is stuck at exactly 5000 (or 3000, depending) every day? I can't say I've ever seen a completely flat graph like that - there are almost always variations, including weekday vs. weekend.
hm, well id paste some screenshots but it doesnt let me in the comments, and it looks, by the scarcity of youmoz posts, that the youmoz article monkey has been shot
will post in a QA when I get QA credits again
The other thing the data shows is the complete unimportance of number of indexed pages (to this site at least), looks like once the 100k important pages are indexed, having an extra million indexed or not makes very little extra traffic
btw, setting QA credits to a month from the last question asked instead of 1 question a month is sneaky and I dont like it (sorry, a tangent there but it really annoys me)
Wow. This is fascinating and there may be something to this. I just ran a quick analysis on my data (also a long tail site - sidereel.com) and while we see a significant breadth of terms driving traffic, our non brand terms, taken individually, also seem to plateau (even for top 5 results on popular keywords).
I'll need to look into this more. Could this be part of a sandbox like effect?
not for mine exactly - i can see the shoulder terms coming out of the sandbox very clearly.
It looks like the increase in traffic from shoulder terms, matches pretty well with a sudden decrease in longtail terms - giving me the same flat line over time, even though i now have shoulder terms added in.
Granted at that level fo detail the data may be too granular to fir exactly, but its a really strange line non the less.
I don't think this is nearly enough information to make that judgement. You'd want to compare several keywords at once (perhaps different groups with different ranking trends). If traffic in general is increasing, or you're in a seasonal business, or there was a random fluctuation in searches for one keyword compared to another, you'd expect to see confusing data.
I don't think it makes sense to claim that you got as much traffic relative to searches when ranked #4 as when ranked #1, unless your copy is just eerily compelling.
I just want to make clear that I'm not trying to suggest that ranking #4 is just as good as #1 - I do think the first set of graphs showed some impact of ranking for that primary keyword. I just want to emphasize that ranking for any one keyword is only a very small part of the bigger picture.
Actually, I think things like seasonal business are a great point, but maybe not quite how you meant. Let's flip this around and say that you had a client who was ranking #5, and, after a lot of work, you got them to #1. They're naturally excited, but then they see traffic fall, because it just so happens that the new ranking kicked in right after the holidays and they have a seasonal sales peak. Is ranking #1 still good? Sure, but it might be dwarfed by other factors, such as seasonal fluctuations, offline marketing efforts, etc.
In my mind, that's actually another argument for looking at the big picture.
Excellent post- obviously you want to track the main keyword target i.e. "sunglasses", but as you mention a drop in rank for 'sunglasses' may not affect your overall business if in aggregate your long-tail 'sunglasses' words are improving. But try telling a client that right? Where it is impossible to check rankings for all long-tail terms one thing we've found helpful is to Filter analtytics data for visits or conversions by keyword. So Filter traffic for any keyword phrase "containing 'sunglasses'", so this will add up 'new sunglasses', 'women's brown sunglasses' etc. If you see a drop in aggregate traffic then that is a stronger sign to re-optimize than just seeing your rank drop for that one keyword. With people searching for so many long tail phrases, and the intense competition for one-two word phrases, one perspective is that SEO is more about optimizing for "topics" than "individual" keywords. btw- we've found these filtered reports are welcomed by clients and helps them understand the nature and potential of SEO.
That's where I think it's good to have a handful of metrics, so that you can tell clients a richer story. Even if ranking isn't falling, you often have a situation where you get someone to #1 and then they don't know what to do next (or don't think you're adding value). If you can show a client that, while their primary keyword stayed at #1, you also got their overall search traffic to double, that's a much more compelling story.
I would think that it all depends upon what the keywords you are targeting are.
If you are in the top 3 for "mortgages" being bumped out of the top teir would make a huge impact.
If you're in the top 3 for scoobydoo lunchboxes then you might not notice a huge difference, especially if you are receiving long tail traffic for "cartoon lunchboxes", etc.
And I agree completely with Mike Tek about the bottom line being all about conversions.
edit: Whoops, I thought I was commenting on the general thread, not Upwords comment. My bad.
Fascinating, excellent post. I'd love to hear how others handle the "broader view" - tracking diverse, long tail phrases...
My site (sidereel.com) has some similarity to the the one profiled, with thousands of long tail searches driving traffic to thousands of landing pages. Tracking and influencing our ranking across so much content is burdensome and time consuming, but also a critical bellwether for our business as a great deal of our traffic comes from organic search.
Our current process involves someone:
- pulling our top 500 search driven landing pages from GA with visits numbers (tragically, this report is sampled)
- pulling our top 500 keywords driving search traffic to the site
- running SEOBook's excellent rank checker Firefox plug in against the names of those pages and one variant keyword we track
- correlating it in Excel against last week (or months results), leveraging excel's auto-styling to highlight significant rank changes
The top 500 (of thousands) is kind of a tease, and not at all a complete view, but it's what we've got given the tools at hand. And certainly gives us a deeper into our overall SEO performance.
I'm intending to experiment with the Google Analytics API's porting directly to Excel (https://analytics.blogspot.com/2009/08/analytics-data-in-excel-through-our-api.html) to try to streamline and expand the process.
Man it would be great if someone built a tool that did this for me... (hint hint...) How are others doing it?
Dr. Pete,
I am big fan of these data driven posts. I like that the take away is a bit hard to swallow.
Given your conclusion, what do you think would be the best metric to measure? Is there a uber-metric that takes into account the long tail?
Well, first a disclaimer: I don't believe in an uber-metric. As a usability professional, I'm very focused on conversion rate and other KPIs, but I can't name a single one that doesn't have limitations. You have to take a wide-angle view.
I should also say that I do still measure rankings - I just prefer to treat them as one piece of the puzzle. In addition to measuring a wider set of rankings, I'm a big fan of measuring overall organic traffic (by engine) and, where it's important, conversion by engine and keyword. This helps sort out (1) which rankings matter, and (2) if you're ranking for the right thing. Not only do I see too many people fixated on a couple of #1 rankings, but it's suprising how often those keywords aren't even very relevant to their goals.
I've had clients that are very focused on one main keyword and don't pay attention to the long tail at all. It's very frustrating because even after showing them that their main keyword only makes up about 5% of their revenue they still use that as their barometer of whether their SEO campaign is going good or not. Sure, it's still important to rank for your main keyword but it usually is only the tip of the iceberg as far as traffic and revenue.
DTC - I agree with you completely. Especially for a lot of mom + pop businesses, they see their 'trophy terms' as being the only thing worth considering when it comes to measuring the success of their SEO campaign. This includes ignoring the long tail altogether.
You can get around this by sticking to the good stuff - figures. If you can consistently show a client something solid, whether it is conversion data or increasing number of signups, then you shouldnt have a problem convincing them that your SEO work is paying off. It may come as a surprise to some, when they see that 'buy green widgets online' converts into signups way better than their 'blue widgets' trophy term.
Educating clients on how to interpret both rankings and reporting data is an important aspect of client management - as you want to avoid the client correlating their 'trophy term' rankings with your success (or failure) as an SEO.
I'm glad to see some data backing what I've thought about for a while. for one of my clients, their 2 word keyword was at a high of #3. They have since dropped to #9-14 fluctuating between there for about 2 weeks now.
However, they are still pulling consistent traffic for many long tail keywords, and overall traffic is doing well.
Rankings clearly do matter as proved in your first images. Ranking dropped and the keyword traffic on that keyword subsequently fell off. As you stated, overall traffic didn't drop much because your client was well set on other terms.
Likewise if rankings dropped on ALL of the clients keywords, they'd be in a world of trouble and probably out of business if they relied on search engine traffic to survive.
Rankings are everything, be it on 1 term or a million terms. If you can't be seen, you can't be visited, and you can't sell anything.
Again, I just want to be clear that I'm not claiming rankings don't matter at all - only that a narrow fixation on just one or a few #1 rankings can paint a very skewed picture and harm your long-term success.
I will argue, though, that we still have to move beyond measuring rankings, because we can't possibly measure all keyphrases. In this example, we're dealing with over 250K phrases for just 7 months, and this is a medium-sized site. Add to that Google's own data, which says that 20-25% of searches are entirely new, and our sites are potentially ranking every day for phrases that no one has every used prior to that day. Tracking all of those rankings is impossible. We have to start looking at search analytics and a wider array of metrics.
excellent!
Once again, it's the hollistic approach that prevails, not a single picked keyword or metrics attached.
This quite relates to the conversation I had recently with a friend who used to work for an "seo plant" (can't find a better descriptive word) where they streamline all their efforts and end up treating most of their client with "ranking guarantees". They'd sign let's say $1000/month deals with clients, for some phrases (let's say there were 10 of them) and would guarantee that chosen kws will rank in top ten over 30% of the time during the billing period. The client was getting tons of traffic coming from anywhere but those 10 kws (mainly long tail that was thought to be the least important) and would pay around 10-20% of the set fee, because that magic 10 kws were popping in and out of the top ten results.
Another great truth comes back once again: it's useless to guarantee rankings!
Agree, to my mind you don't need to track rankings you need to track ROI. I could deal with the clients "target keywords" declining in rank so long as my clients ROI from search was increasing.
If the client did like that i'd tell them to take their business elsewhere.
Hi Pete, I have had similar experience. One of my clients had over 300,000 long tail keywords bringing in traffic, where only about 5% of the overall traffic was through the 50 main keywords that they had targeted.
Based on your theory that it is nearly impossible to actively manage 300K keywords (which I agredd with completely), what do you think is the best SEO strategy to get traffic for long tail keywords: Write quality content for these keywords, build intenral links for the long tail pages and external links for the home page (assuming that the top level domain will get the authority that will help us get these long tail pages ranked highed up for those keywords, even though the long tail pages may not have any additional external linksc coming in).This is the approach that I have taked thus far but feel that the strategy is almost like shooting in the dark where you dont really control much.
Or is there anything else that we as SEOs can do to better rank for long tail keyword?
Yeah, I think you're exactly right - targeting the long-tail at that level becomes about content marketing and SEO-friendly site architecture. You have to get people building deep links and get spiders to your content - those 300,000 long-tail phrases will naturally emerge.
Even you chose to spend the manpower to manually target 300,000 phrases and could somehow make it cost-effective (and, if you do, please tell me how), you'd still be missing out on the new search phrases people are using every day. Of course, you still want to target your core keywords and variants, but the larger part of the long-tail is going to emerge from your content.
Here is the example I use with my guys for long tail:
Keyword term: Affordable website optimisation specialist in Lancaster
Can assist in getting found for many variants of the phrase:
Affordable website optimisation specialist in Lancaster
Affordable website optimisation specialist
Affordable website optimisation
Affordable optimisation specialist in Lancaster
Affordable optimisation specialist
optimisation specialist
and so on. I am a little busy but you get the idea.
Great article. Nice to be backed up by SEOmoz. Keep up the great work.
a great post
Great to see the long tail benefits.
I agree the long tail keywords can often amount to a lot of traffic, and not only that are much more targetted traffic, as they are much more specific terms in most case.
Thanks Pete!
This is something that we have been telling our clients for a some time now but great debates like this really help showing our clients that although ranking for their primary keywords is important there is a bigger picture to consider. Indeed, it's about traffic diversification and writing great content to attract long tail phrases
Interesting post. I have been beating this to death with my clients b/c having tunnel vision about one or two keywords can really harm an SEO campaign.
Pete, such a fantastic post and VALUABLE discussion! I just had to comment so I could follow the ongoing comments!
It's so nice to see this example, this helped me lots to userdtand in an easier and clear way!
Great stuff, thanks for posting this, Dr.Pete
This is just one example or did it happen to lots of clients as well?
Imo, if you stay in top 3, that will be find, but if you drop out of top 3, you will suffer traffic loss in most of the cases.
If your main keyword is "camera," then your first chart shows what happens when you drop in the rankings for "camera."
What happens if you look in Google Analytics for organic search terms and type "camera" at the bottom, so you see all search queries containing the word "camera?" It would be interesting to know what happened to traffic from this group of keywords.
Love the questions I'm getting. I actually have this data, and it's an interesting question, so I ran some quick numbers. Not surprisingly, the traffic data for keyphrases containing a variant of the primary keyword falls somewhere between the two graphs. They depend on primary keyword ranking a bit less than the primary keyword itself but a bit more than overall search traffic.
I will admit, they track a bit closer to the primary keyword than I'd expect. In many cases, it feels like longer-tail phrases which include the primary keyword are getting more unique - you have to compete and ranking for them separately. In this case, though, traffic for the primary keyword correlates pretty highly (r = 0.78) with traffic for all keyphrases containing the primary keyword. Definitely food for thought.
One minor note - the phrases that include the primary keyword include my client's brand and domain name, so that may skew the data a little bit.
Great post, you show some good data and make some great points.
However, you negate your entire stance that ranking is not that important when you point out this is for a keyword only attributing 2% of your total traffic.
Do the same for a keyword bringing in 50% of your traffic, and I am SURE you will see data in completely the opposite direction.
Actually, that's more or less my point - the rankings drop wasn't nearly as disastrous as the client feared precisely because what they thought was their most important keyword actually accounted for only 2% of the puzzle. That's really the only point I'm trying to make - you've got to look at the big picture.
These days, though, I think you'd be hard pressed to find a site where one keyword accounted for 50% of search traffic, and if you do, than I'd argue that it's a pretty ineffective SEO strategy. Google's own data is clear - people are using longer search phrases every year and are getting more and more targeted in their queries. If you're not evolving with the changing needs of search users, you're missing out on a lot of opportunity.
Thanks for putting your point in better perspective for me. Now your entire post certainly has a great vibe to it in my eyes.
Again, thanks for posting hard data!
I would mention another important factor to be taken into account here which is personalization of search results. Depending on the keyword and history of searches, users might have received your link on different placements in SERPs.
Did you take this into account?
When collecting this kind of information for our clients' websites we can clearly see that rankings sometimes vary significantly - even from 1st to 10th place!
I think this may also affect relevancy of this research quite a lot...
Although there's certainly some noise in any given ranking these days, the drop and subsequent recovery were prolonged and verified in this case. Personalization is definitely having an impact, though, and is one more reason we have to look beyond rankings, especially any single, observed ranking.
Great post but I would be curious as to how conversion rates faired as rankings drop thus giving a coorlantion not only from ranking to traffic but from ranking to conversion rate.
To be blunt did/do you find higher rankings leading to better quality traffic.
I just noticed at other people asked this same question--fun, fun.
My general observation is that, while higher rankings lead to more attention and higher click-through, that doesn't always mean better conversion. Being #1 carries some inherent authority, but it's also where impulsive clickers instinctively go, meaning it can frequently carry a high bounce rate as well.
In this particular case, I can't pull the CR by the individual keyword in any way that would be useful - the data volume just isn't high enough. It's certainly an interesting question in the broad sense, though.
very cool charts the data is quite interesting i really wish i had taken stats in college so i could do my own analysis
Once again a great post. Most of them care about the site rankings for their brand terms and ignore the longtail keywords. Your analysis with the graphs is the rightway to explain this to a client. I'll be using this from now on.
If I understand the conversation, it seems to me that the message is once again: content is king. Juicy content with lots of unique info and long tail phrases can drive a massive share of your search traffic to your site. So create great content, right?
You'll have your prized keywords that drive tons of traffic and/or conversions, but don't forget about the solid foundation of the long tail (i.e., the big picture).
Whew ... love the debate. Rankings are definitely only one piece of the puzzle. You have to rank to drive traffic whether it's for one search term with a good amount of volume or 250,000 long-tail terms. Seeing the big picture really requires pulling rankings on hundreds of thousands of terms and being able to surface indicators accross that set of data as well as accross relevent groups of terms, kind of like you do with a PPC campaign. That kind of information can be really powerful.
It's really about (quality) traffic diversification. It's not smart to rely on only a few keywords to drive traffic. Sure, rank for some terms with decent volume, but devise a strategy for driving long-tail traffic. Then make sure you're originating traffic from sources other than search.
Good point! This is exactly what I've been thinking about lately too. Thanks!
Tracking diverse, long-tail phrases. Yes. I think it's important when we turn to long-tail search ranking.
This is something that we have been telling our clients for a long time. And it's actually based on advice we've received from SEOmoz over the years.
Rankings don't matter all that much. They are nice to throw in a report to show to the client, but at the end of the day, I know in my heart that ranking well for a few key phrases isn't going to matter all that much.
It would be interesting to see how this fluctuation affected sales instead of only traffic. If we saw that their sales didn’t dip then it could be a clear indication that this keyword doesn't convert, but then again if it does convert, how much sales volume does it actually push? What if this one keyword was responsible for 98% of sales? Although highly unlikely, it still good data to have when formulating an SEO strategy.
Great research!!
Thought I'd pull a few conversion numbers. I don't want to give out absolute #s (for client anonymity), but let's call the conversion rate for the primary keyword: X%. The CR for all keyphrases containing the primary keyword was about double (1.87X) over the 8 months, while CR for all keyphrases NOT containing the keyword was about one-third (0.36X).
Of course, the long tail here is very long, so that 0.36X accounts for quite a few sales. Unfortunately, the data for the phrases containing the primary keyword is a bit skewed, as it includes their brand and domain name (which, of course, convert well).
Very good article. It makes me have different idea about my website(Ineed-electronics.com) Page Rank.