Brad Geddes / PPC Geek
Official AdWords Seminar Leader.
Author of Advanced Google AdWords.
Co-Founder, Adalysis.
(312) 884-9017
Brad Geddes's Theories on Marketing Evaluating the Impact of AdWords Sitelinks

Evaluating the Impact of AdWords Sitelinks

This is a guest post by Chad Summerhill, Author of the blog PPC Prospector.

Earlier this month, I was auditing data in our web analytics data warehouse when I came across some AdWords keywords that were being tracked improperly.

At first, I thought that AdWords was using my ad’s destination url instead of my keyword’s destination url.   After some investigation, it turns out I was wrong.

The issue was originating with clicks from AdWords sitelinks that have their own destination urls.  In my case, they just happen to look like the urls on my ads which caused my initial confusion.

While I was under the hood, I noticed that Google had disapproved my brand campaign sitelinks (without telling me).  Apparently, you can’t have your sitelinks all pointing at the same url (which is what I had done when I first got access to the sitelink beta).  If someone rats you out (a competitor I guess) then AdWords can disapprove them after a manual review process.

The truth is I just added the sitelinks without much thought.  Seemed like a no-brainer at the time—my brand ads would take up more SERP real estate. Yes! Sign me up.

After learning that the sitelinks had been turned off, I needed to understand the impact and determine if I needed to turn them back on.

Sitelink reporting in AdWords

There two places in your AdWords account for retrieving sitelink data: ‘Segment> Click type’, and ‘Ad extensions> Sitelinks Extensions’

By Click type Segment

Click Type Segment in AdWords

Click Type Segment in AdWords

 

When looking at your ‘Click type’ segment you can see the number of impressions, clicks, etc. on your actual sitelinks.  This view of how many searchers clicked on your ‘Headline’ vs. ‘Sitelink’ isn’t all that helpful when trying to understand the overall impact of having my sitelinks turned off.

Under the ‘Ad extensions’ Tab

Sitelink Extension Reporting in AdWords

Sitelink Extension Reporting in AdWords

 

When I turned to the ‘Ad extensions’ tab, I started to see a better picture of the performance boost of using sitelinks.  This view, reports all the ad performance metrics of ads clicked when sitelinks were displayed regardless of where the searcher clicked the headline or the sitelink. So, you can see the difference in CTR, etc. when sitelinks are present.  But it still didn’t really help me understand the overall impact of losing the sitelinks.

Trend your data to understand the impact of sitelinks

The best way I found to understand the impact of using sitelinks was to turn them back on and do a time comparison (pre-sitelinks & post-sitelinks).  This easily accomplished by pulling a campaign report segmented by day and analyzing it in Excel with a pivot table.  Of course, you can see your important metrics trended over time inside the AdWords interface as well.

Trended AdWords Data

Trended AdWords Data

 

In the image above the blue line is CTR and green is Conversions, so it looks like sitelinks made a big impact just as I had suspected.  In my case, turning on sitelinks improved my overall campaigns CTR by 40% while having no affect on my Conversion Rate.  Consequently, my overall daily-spend was up by 54%, so I decided to dig a little deeper.

Evaluating the impact of sitelinks on your organic traffic

Adding sitelinks makes your paid ad look more like an organic listing (at least I think it does) and it also takes up more real estate on the SERP, so I decided to look at what adding sitelinks did to my overall branded search traffic (both PPC & SEO).

This was easily accomplished using Google Analytics.

1.    Go to: Traffic Sources> Keywords (if you want you could also create an Advanced segment for just Google search)

2.    I filtered for keywords that contained my brand using a regular expression tip from Blue Glass.

3.    I then downloaded both the paid and the non-paid report and analyzed them in Excel.

What I found was very interesting.  Overall year-over-year branded visit growth didn’t change at all.  Growth just switched from organic to paid after I turned sitelinks on.

Effect of AdWords Sitelinks on PPC & Organic Search Traffic

Effect of AdWords Sitelinks on PPC & Organic Search Traffic

 

Now I need to decide whether or not to keep sitelinks running.  I’m almost certain that I will; the traffic is cheap and it seems like the right defensive move to protect my brand.

Now imagine what sitelinks could do for your non-branded, competitive ads (Google reports and average increase in CTR of 30%).  Sitelinks could help you take marketshare  from your competitors when you don’t own the SERP like you probably do with your branded keywords.

My next analysis will involve a non-brand, competitive, high-CPC keyword that I rank well for, in both PPC and SEO to see how much I cannibalize from SEO.  Maybe that’s a topic for a future post.

Opinions expressed in the article are those of the guest author and not necessarily bgTheory. If you would like to write for Certified Knowledge, please let us know.

No Comments

  1. speedypin
    March 17, 2011 at 1:38 pm · Reply

    Chad, thank you for this post. Everything you describe I have personally experienced using Yahoo PI (Paid Inclusion). Adding site links did increase CTR, but also increased our CPA (Branded and Non-branded campaigns). Like you, I believe Site Links make a paid ad look more organic. This is why we see the higher CTR (more tire kickers).

    I admit to having not turned on Site Links for any of our AdWords campaigns because of this previous experience. However, after reading about your experience, I feel reasonably secure with enabling them for our Brand Campaign to see what happens. This feels safe because the relative cost of branded keyword clicks is staggeringly minuscule vs. those non-branded.

    Agreeably, only testing will validate my theory that enabling Site Links on our non-branded campaigns will result in higher CTR and a higher CPA. However, I am not sure testing is worth the time because I am already of the opinion that we will ultimately end up paying for what would have otherwise been “free” traffic. Same visitors. Same page/landing page. Different response from the visitor? Hmmm…

    Let me (us) know if you end up testing this and the result. I would love to test this over the next few weeks and report back, but the already insane amount of potentially-damaging campaign factors in my niche make PPC a very challenging task.

    Thanks again!

    – Eric Itzkowitz

  2. Adrien O'Leary
    March 17, 2011 at 3:30 pm · Reply

    Hi Chad, thanks for sharing the details of your experiment. I was wondering whether you came up with a way to track CTR and conversions for each of the sitelinks. Being able to track the impressions for each sitelink is also something I’d be happy to measure. Thanks.

  3. brad
    March 17, 2011 at 6:44 pm · Reply

    Adrien, This is how Google suggests tracking sitelinks: https://adwords.google.com/support/aw/bin/answer.py?hl=en&answer=190706

    However, I don’t like that method as GA doesn’t support the ‘origin’ tag. Personally, what I’m doing is using the UTM_campaign as ‘sitelink’ and then using the ad type and other GA parameters for the campaign and sitelink name. Then I just look at the effectiveness in GA.

    Anyone else tracking sitelinks in a different way?

    • Adrien O'Leary
      March 21, 2011 at 3:06 pm · Reply

      Hi Brad, thanks for your comment. I usually to use the Analytics auto-tagging within AdWords. I’ve just started testing sitelinks tracking by simply adding ?utm_content=SiteLink on the sitelinks of a campaign. I’ll check back in the next ways if that works just as I’d like it to work.

  4. brad
    March 17, 2011 at 6:46 pm · Reply

    Speedypin – I’m a bit conflicted on your post. At times I agree with you and times I don’t (I’m kinda waiting for either me to make up my mind or for some really compelling argument one way or the other).

    The reason why is that a high CTR means a better QS and potentially lower CPC. I think if the account had QS issues – then I’d be more inclined to use sitelinks just to raise the overall CTR of the account. If the account didn’t have QS issues – then I’d be more inclined to follow your line of thinking.

    • speedypin
      March 18, 2011 at 12:42 pm · Reply

      Brad – Sorry for leaving out one of the most important factors (:.

      Our QS on brand is 10 for all keywords. Our non-brand campaigns are mostly 9 and 10 with a spattering of 8’s and occasional 7’s, depending upon how long-tail they are. As such, we pay relatively low CPC compared to many of our competitors. We definitely garner many impressions and clicks.

      Conversion issues? Our conversion for our brand campaign is very high so not much to do there. One of our other core campaigns (Countries) is currently averaging 10%+ and sure it could be better.

      We use our well-ranked SEO pages as our landing pages. Agreeably, many would say this is generally not a great idea (too many distracting on-page features and functionality). However, I have tried various A/B tests using custom landing pages, but they only seem to work for a while then die leaving me right back where I was or worse. Fun, eh? LOL!

      To be fair to anybody reading this, my experience in this niche is certainly atypical. The prepaid phone card industry is just a monster. Running a PPC advertising campaign(s) for a company within it while achieving a + ROI is very trying. Further, there are many other variables (opposing factors) influencing our industry, which I have not mentioned. At any rate…

      Yesterday, I did turn on site links for our brand campaign and will report back with my findings as they come about.

      Thanks for everything you do for our industry!

      – Eric Itzkowitz

  5. brad
    March 19, 2011 at 1:34 pm · Reply

    Speedypin – I see. If you’re doing tests and the organics do work better than I can see your point a lot clearer.

    I think that’s why testing is so essential. There’s ‘best practice’ and then there’s results.

    Every once in a while best practice doesn’t work (I have one account where the homepage is the best and no matter what the homepage always wins – 150k+ words all going to the homepage), and if that’s what your tests show – then you need to follow the results and revenue.

  6. bloomarty
    March 25, 2011 at 10:57 am · Reply

    Hi Chad,
    thanks for sharing these results. Your conclusion about sitelinks stealing traffic from organic without any growth overall makes sense. It’s consistent with research that shows that organic + ppc often doesn’t provide additional traffic compared to organic results alone. However, some argue that overall profit goes up if you combine organic and ppc (see http://pages.stern.nyu.edu/~aghose/paidorganic.pdf). I wonder if sitelinks would help with conversions as well…

    Anyway, your post made me think. You wrote that your overall campaigns CTR increased by 40%, which is probably the right number to look at. Still, I’d really like to know about the direct impact of sitelinks on CTR. The 40% increase represents an average: some ads had sitelinks, some did not. The ads without sitelinks probably didn’t have any increase in CTR, so the ads with sitelinks would have had to compensate with higher increases to reach the average of 40%.

    Now I had an idea that involves some simple math.

    The sitelinks extensions reporting in Adwords tells you, how many clicks and impressions your ads with sitelinks generated. The regular statistics also tell you your overall clicks and impressions. By subtracting sitelinks impressions from overall impressions you can find out how many impressions without sitelinks you had. The same goes for clicks. This lets you determine your CTR with and without sitelinks.

    Looking at a post-sitelinks timeframe, lets say we have a CTR of 30% with sitelinks and a normal CTR of 10%. Of course, these numbers don’t really tell you much. To qualify for sitelinks, an ad has to be on a top position, which already guarantees a much higher CTR.

    You can also find out the percentage of ad impressions with sitelinks. Lets say, 50% of impressions came with sitelinks. Now lets look at a pre-sitelinks timeframe. We know that 50% of impressions would not qualify for sitelinks, and 50% would have. So lets divide impressions accordingly. For the impressions that never qualified for sitelinks, CTR never changed – it was always at 10%. With this, we can calculate the number of clicks for these impressions. We can also calculate the number of clicks for the impressions that later qualified for sitelinks: Total clicks minus the number of clicks from the non-qualified impressions.

    From there it’s a small step to the CTR of the impressions that later qualified for sitelinks. We can compare this CTR to the CTR with sitelinks and we know the exact impact of sitelinks.

    I did these calculations for a branding campaign where CTR increased from 29% to 33% after sitelinks were implemented. The share of sitelinks impressions was 42%. CTR with sitelinks was 52%, without it was 19%. This lead to a CTR for sitelinks-qualified impressions before the implementation of 42%. So the CTR increase with sitelinks was (52/42 – 1 = ) 23%.

    I’m not sure if anyone else cares about the actual CTR increase of sitelinks, but I’m always interested in how these things work, so I thought I’d share.

    Looking forward to the post about non brand campaigns…

    Have a great weekend

    Martin

  7. needmoreppc
    November 14, 2012 at 1:34 pm · Reply

    Any clarification on the ?utm_content=SiteLink issue?

    Has anyone tested this further?

    I’m seeing a nice uplift from Sitelinks, but I want to be able to track it closer and I’ve been mulling this one over for a while.

Leave a Reply