THE RKGBLOG

Quantifying Google’s CPC Decline: Don’t Blame Mobile

This piece originally appeared in RKG’s Dossier 3.2.  Dossier is a quarterly journal filled with original articles detailing best-of-breed strategy and encouraging a data-driven approach to online marketing.  To sign up for updates on Dossier or to review past issues, check out our page here.

Mobile cost-per-click has been a hot topic in the paid search industry as of late, particularly once it became apparent in the fourth quarter of 2011 that Google would likely post a year over year decline in overall CPC for the first time in over two years.  Google ended up reporting an 8% CPC decline in Q4 2011 and followed that up with an even steeper 12% decline in Q1 2012.

Two hard-to-dispute facts are leading many analysts to point the blame at mobile.  First, that the mobile traffic segment is growing faster than the search industry as a whole, and second, that mobile paid search clicks command far lower CPCs than desktop.  But does this story really hold water?

In the most recent RKG Digital Marketing Report, we captured the acceleration in the decline of Google CPCs, noting that we saw a 9% year over year drop in Q1 2012 vs a 6% decline in Q4 2011.  We also found that mobile comprised 13% of paid clicks in Q1 2012, nearly double the 7% level we found in Q1 2011.   So far so good, but let’s take a look at where that growth is coming from.

“Mobile” growth is being driven by tablets

When the original iPad launched in April of 2010 the entire mobile segment accounted for just a little over 2% of paid search clicks.  Still, RKG quickly recognized that tablet conversion metrics far outpaced those for smartphones and began efforts to segment the iPad from smartphones.  This was over a year before Google allowed for the targeting of tablets, in general, as a separate class of devices.  By the end of 2010, smartphone click share had risen to 5.3% while tablets held a 1.1% share.

Fast forward to the end of Q1 2012 and we find smartphone share has only risen slightly to 6%, while tablet share now stands at 7%.  While the iPad and, to a lesser extent, the Kindle Fire have become wildly popular, our segmentation efforts on account of the differing value of traffic across device classes has led to this wide disparity in growth rates.  This is critical to assessing the impact of the mobile segment on Google CPCs.  Smartphones could have 99% lower CPCs than desktop, but if their share hasn’t grown, it would have no impact on the year over year change in Google’s overall CPC.

How does Cost-Per-Click fare across device classes?

In reality, we found that smartphones commanded a 54% lower cost-per-click than desktop in Q1 2012.  A few factors go into this discrepancy in CPC:

  • RKG is generally segmenting out mobile traffic and bidding it lower across the board.
  • Branded keywords, which carry lower CPCs, make up a larger percentage of mobile traffic than desktop traffic.
  • According to Google, when mobile and desktop clicks are served from the same campaign, Google may apply smart pricing to lower the cost of the mobile traffic.
  • Poorer conversion for mobile traffic is likely decreasing competition.

Because of the iPad, the revenue per click our clients generate from tablets runs higher than desktop on average, but tablet CPCs still came in 11% lower than desktop in Q1 2012.  Our bidding systems are being more aggressive with the tablet segment, so this is a bit surprising.

Tablet CPCs come in lower than desktop even when we eliminate brand traffic from the equation, restrict our view to those campaigns that target both tablet and desktop at the same bids, or even individual keywords doing the same.  This suggests that competition for tablet traffic is lower even though its quality is better and/or Google is applying a smart pricing reduction to tablet clicks as well.

Doing the math on mobile

Google did not provide mobile costs broken out by tablets and smartphones in  Q1 2011, but we can easily determine the impact of the combined mobile segment on the overall Q1 year over year CPC decline.

We know that our CPCs for desktop fell 7.7% in isolation while our overall decline was 8.7%.  Therefore the combined mobile segment only tacked on another 1% to an already significant desktop decline.

What’s interesting is that CPCs for the combined mobile segment actually rose nearly 10%, thanks to the traffic shift to tablets.  Unfortunately for Google, the fact that tablets still command lower CPCs than desktop meant that mobile still managed to drag down overall CPCs, albeit to a lesser extent than you might have assumed given the coverage of the issue.

What else is bringing down CPCs then?

Google’s officially reported paid click growth rate hit 39% in Q1 2012, a level it hasn’t seen since the third quarter of 2007.  That surge in clicks coincides neatly with a drop in CPC growth, but the mobile segment just isn’t big enough to have that kind of impact yet and the timing just doesn’t fit either as the graph below suggests:

Clearly, other factors are at play.  Here are a few that we believe are substantial:

Product Listing Ads:

RKG has seen a steady and rapid increase in traffic to Google’s Product Listing Ads (PLAs) format since they first migrated to the AdWords platform in late 2010.  PLAs trigger almost exclusively for non-branded search queries — except for those manufacturers who sell their own merchandise online — and feature information about a relevant product, including an image and pricing details.

In Q1 2011, PLAs generated 3% of Google ad clicks with an average CPC that was 12% lower than standard text ads.  This year in Q1, PLAs held an 11% click share while CPCs where 18% lower than the traditional ads.

Excluding PLAs, our non-brand Google CPCs fell 5.2% Y/Y in Q1.  With PLAs, non-brand CPC fell 7.0%.

Brand shift:

In assessing how Google has managed to drive up ad click-through rates in the neighborhood of 25-30% Y/Y,  RKG has repeatedly pointed to a series of format changes over the last year or so that have given greater prominence to the paid listings while blurring the lines between the ads and the organic results.

Many of these changes, such as Sitelinks and other ad extensions that give more real estate to a paid listing, only trigger when the ad is in a “promoted” position in the top 2 or 3 spots above the natural results.  In turn, brand ads, which are more likely to be in top positions, receive a disproportionate amount of the benefit.  Brand ads also tend to have significantly lower CPCs than more competitive non-brand ads.

We found brand clicks to be up 51% in Q1, compared to a 34% increase for non-branded ads excluding PLAs.  At the same time, the average cost-per-click of brand ads was 85% lower than the average non-brand CPC.  Compounding this shift in traffic to lower priced brand terms, we also found brand CPCs falling significantly faster than other segments with a 27% year over year decline.

Why exactly this was the case is difficult to pin down, but it may stem from the brand click-through rate increase providing a relatively larger Quality Score advantage than non-brand terms received.  In other words, the click-through rate boost from Google’s tweaks has likely spread more evenly among competitors for non-brand phrases, so there isn’t the same drop in CPC from achieving a relatively higher Ad Rank.  Also, the overall shift to mobile, with its lower CPCs, has a higher impact on brand terms.

Whatever the case may be, the impact here could have been much larger, but brand costs only accounted for 5.5% of our total Google spend in Q1.

Budgets, Increased Rationality, FX:

RKG clients tend to be more flexible with their spending budgets than the average advertiser and they look to seize additional opportunities when they can be had at a given efficiency target.  Those with stricter constraints in place may have been caught off guard by the surge in Google traffic growth that began in earnest in Q3 2011 and continued through Q1 2012.

Greater sophistication in the marketplace may also be leading to an increased focus on, and the ability to more accurately measure, advertising profitability.  RKG clients have made appreciable year over year gains in return on ad spend over the last two quarters and that is partly driven by our own efforts to deliver stronger marginal returns.  Although probably not what they had in mind, Google has been a big help in illuminating inefficient marginal performance by making its bid simulator data available via API in mid-2010.

Finally, Google themselves have pointed to foreign exchange rate effects as a contributing factor to their CPC decline.  International revenue has comprised over half of Google’s total haul for years now, so it isn’t hard to believe that FX can play an appreciable role in how their finances look in a given quarter.  That said, RKG’s spend under management is very heavily weighted to North America and our CPC trends have still followed Google’s closely, so the FX impact on CPC appears to be minor compared to other factors.

Will Mobile CPCs ever catch Desktop?

It may only be having a minor impact on Google’s overall CPC trends now, but mobile traffic and its pull will only continue to grow in the months and years ahead.  On Google’s Q1 2012 earnings call, the question of if and when mobile and desktop CPCs would merge was raised and Larry Page had an interesting response.  He suggested that, in time, the disparity between mobile and desktop CPCs may reverse due to consumers spending most of their money locally, which is an area where mobile has the advantage.  He then added, “we’re very bullish about that.”

Bullish indeed.  Page didn’t give a timeline for this reversal, but in the near term, it appears unlikely that mobile, particularly if we limit our definition to smartphones, will catch desktop.  While there’s a lot of work going into making the mobile browsing and checkout process easier on customers, it remains far more tedious to shop and complete a transaction with our thumbs on a 4 inch screen than it is to do the same on a full size computer with a keyboard and mouse.

The heavily local intent of smartphone shoppers also significantly reduces the value of that traffic to online pure-plays, lessening the overall competition and, in turn, click costs.  Even for retailers with brick and mortar locations, tying offline sales back to mobile ads accurately and consistently is a major hurdle.  Google, of course, knows this and they will continue to try to help advertisers make those connections to revenue while also steering them towards assessing mobile performance by less direct measures of profitability.

By the time we all get this figured out, the definition of mobile may be quite different than it is now.  Two years ago tablet traffic was effectively zero, while the iPhone, which set the standard for all touchscreen phones to come, has been around for less than five years.  Who knows, maybe Google’s Project Glass program will actually come to fruition in another two years and we’ll all be sporting a head-mounted display and shopping with the blink of an eye – while taking great care not to walk straight into a streetlight, of course.

Ultimately, if Google wanted to increase CPCs tomorrow just to get all of the financial analysts off their back, they have the ability to do so even with the auction mechanism in place.  But Google doesn’t or shouldn’t care about CPC by itself, just as its advertisers shouldn’t.  Their concern is revenue and profits both in the short and long-term.  As long as internet commerce continues to grow and the effectiveness of our advertising continues to improve, Google will be well-positioned.

  • Mark Ballard
    Mark Ballard is Director of Research at RKG.
  • Comments
    8 Responses to “Quantifying Google’s CPC Decline: Don’t Blame Mobile”
    1. Great analysis and insight, Mark.

      I’m involved in a number of analysts calls and their disproportionate focus on the decline in average CPC is quite concerning. While CPCs are lower, the actual number of clicks being generated is far higher; hence the continued revenue growth,

      One area that you only slightly touch on is international. You mentioned FX impact, but average CPCs outside of the UK & US are comparatively low, and Google is seeing strong growth in these regions so that must be pulling down he average CPCs also.

      Regards

      Richard

    2. Mark Ballard Mark Ballard says:

      Good points, Richard. As I noted above, our numbers primarily represent North American performance and we are still seeing large CPC declines, but a shift to international markets with lower CPCs would bring Google’s reported CPCs down farther. Google does provide a US v international revenue breakout, but I don’t believe they do the same with clicks so it’s difficult to estimate the full CPC impact here. International revenue only moved up from 53% in Q1 ’11 to 54% in Q1 ’12 though.

    3. Terry Whalen says:

      HI Mark, great article – especially the concluding paragraph, which puts things in solid context.

      Question – aren’t display network revenues still growing faster than search network revenues? And wouldn’t that pull avg CPCs down? (Or are analysts focused solely on search-network-only CPCs?)

    4. Mark Ballard Mark Ballard says:

      Great questions, Terry. I believe Google’s display revenue is growing faster than search, but I don’t know what the CPC differential might be for them in total or exactly the % of their revenue from display (Techcrunch estimated 12%). We definitely see cases where a client’s display program is exclusive to remarketing, and those programs can have higher CPCs than paid search.

    5. Terry Whalen says:

      I think Display is growing faster than Search, and the vast majority of display is non-remarketing, with CPCs that are typically much lower than search CPCs.

      So, I think we can add Display as another downward force on CPCs (but upward force on paid clicks).

    6. Picking up on Richard’s point. I agree with you that it is strange the extent to which the stock analysts focus on CPC trends. I made the point on an investor call last week that “I don’t think Google cares what the average CPC is. It’s a KPI that is indicative of nothing. It’s like website conversion rates. If I could double traffic to your website cost effectively, but the traffic I bring converts at a lower rate, would you care about the impact on conversion rates?”

      Interesting that these folks who are high paid number crunchers don’t seem to get that.

    7. Thanks for the interesting analysis, Mark!

      I work with very few North American AdWords accounts, so I’m speaking from a Northern European perspective.

      I noticed a drastic and fast decline in brand term CPCs last November with almost all of our clients. We were conducting normal ad A/B tests as usual but I doubt all brands suddenly and simultaneously reached a CTR (or other) threshold where CPCs would drop, especially when QSs stayed quite static. Algorithmic update? I haven’t found out.

      One factor contributing to lower overall CPCs (via branded terms) is Google being so heavily “brand biased”, which we see every time we type any letter in Google’s searchbox and wait for the autocompleted suggestions.

    Trackbacks
    Check out what others are saying...
    1. [...] Lasten der organischen Ergebnisse geschaffen hat? Nicht nur gesunder Menschenverstand, sondern auch fundierte Studien legen nahe, dass die negativen Auswirkungen der Expansion von Product Listing Ads und Brand Ads auf [...]