Feb 32011

Online Sales Potential by Browser: Safari Users and Upgraders Lead

At RKG we are experimenting with using Google's Enhanced CPC option because it attempts to better optimize paid search bids by utilizing factors that advertisers aren't given the option to control fully themselves.  Among the signals Google takes into account with ECPC is the type of browser being used and we felt it was a good opportunity to review our own data on the subject to estimate the potential impact of making bidding adjustments based on browser usage.

Looking at a sample of our retail clients we aggregated results for the most common desktop web browsers in use and came up with normalized results for sales per ad click (average set at $1.00) as well as the overall traffic share for each version:

We can immediately see there is a real opportunity for advertisers to segment their spend by browser version or, more realistically, let Google try to do it for us.  For every dollar being spent to draw in an Internet Explorer 6 user, we would be willing to spend $2.20 for a Safari 5 user as we would expect them to spend over twice as much following each ad click.

Why do Safari Users Convert Better?

Demographics likely explain much of the difference as Mac users are more likely to belong to high income households.  We also find that, in general, users of up-to-date browser versions tend to convert better and Safari 5 is Apple's newest version.  Users of the newest full release of Internet Explorer (IE8), also demonstrate much higher SPC than IE6 users, while Firefox 3.6 users convert better than Firefox 3.5 users.

Using an older browser may be an indicator of the consumer's comfort level with technology or even their internet connection speed (browser file sizes generally range from 15-25 MB).  Someone using the browser that came pre-loaded on their 3 year old machine may be savvy enough to price compare online, but they may still prefer to purchase in-store at higher rates than average.

Browser functionality is certainly a major factor here as well as older browsers lose the support of developers and simply can't process web pages as fast or as well as their newer counterparts which hurts conversion rates.  For advertisers working in a multi-channel tracking environment, javascript execution times can also mean the difference between crediting an order to say, paid search, or crediting the same order to natural search or to no specific channel at all.

What Impact Might We Expect from Google Enhanced CPC Utilizing Browser Information?

Although there are huge differences in sales per click across the browsers above, the most popular browsers are naturally huddled towards the average, diminishing some of the segmentation potential.  Applying a simple model for the expected change in traffic given change in CPC, we would expect to be able to increase revenue by around 4-6% at the same efficiency by taking into account browser version in PPC bidding.

That's a bit idealized though.  It should be noted though that Google's ECPC option will only alter a bid by up to 30% and, in some cases, such as for Safari 5 users, we would want that percentage to be higher in order to fully account for the differential in traffic value.  It's also unclear how deeply Google is looking at this information.  We hear that they do consider the browser release version, as opposed to just browser type, but we can't say how fine those distinctions are.  Lastly, ECPC looks at a number of other signals and browser is not completely independent from those other variables.

Ultimately, we're happy to see Google working to help advertisers improve traffic segmentation, even if it's out of our hands more than we prefer.  For well established PPC programs, one time projects that have the potential to generate revenue gains of even just 5% or so can be difficult to come by, so we would encourage experimentation with ECPC and we hope you'll see even better results than that.


15 Responses to "Online Sales Potential by Browser: Safari Users and Upgraders Lead"
Hi Mark, really enjoyed the post. By using Google Conversion Optimizer (ECPC's big brother) Google isn't limited to the 30% increase. Of course, you would have to be more granular with your account structure to have the appropriate control over your bids (i.e. head terms in very small ad groups or even by themselves). And I’m not sure how many bid management platforms can set a CPA bid at the ad group level. I wrote an article not too long ago about combining 3rd party bid-management and using ECPC that your users may find relevant: http://www.wordstream.com/blog/ws/2010/09/01/enhanced-cpc-bid-management Thanks, Chad
Mark Ballard says:
Hi Chad, I really enjoyed your post as well. Having the 30% limit is probably comforting to some compared to giving Google more free rein under Conversion Optimizer. Also, while it's certainly possible to reorganize a program to fit a CPA model for Optimizer, as you've shown, it's going to be a big challenge for many. If Google is unwilling to give us more control over these variables or it is simply not worth it to them (since few would use them), a nice middle ground may be the ability to tweak that "enhancement level" for ECPC based on results.
Great post here Mark. It's very interesting to see conversions vary not just by browser but browser version as well. Time to pull out the same type of report and see if the results would be consistent with yours. Cheers!
As Mark points out, when wading into the product, it is comforting to have a +30% guardrail. It's worth noting though that Google is also reducing bids in many auctions when eCPC is utilized. Furthermore, these bid reductions can be to the tune of -100% in some extreme cases (bid to $.00). So for clarity, the full limitation on eCPC bid adjustments range from -100% to +30%.
Geordie says:
Hi Mark, Great points here. Do you have any feedback as to whether Enhanced CPC has worked for your clients as yet? What would you say is the best way to test its effect accurately? Cheers! Geordie
Mark Ballard says:
Geordie, we're looking to follow up on both of your questions in more detail soon, but the short version is: We do see Enhanced CPC working, but the effects aren't huge because we were already taking into account many of the more important factors Google is using like the network and matchtype. Since ECPC is basically an ongoing test within the campaigns opted into it (a portion of bids are left untouched as a control) the best view of the data is probably from Google if you can get them to provide the necessary breakouts.
Geordie says:
Thanks Mark, With a lot of the campaigns I manage, one of the challenges with Conversion Optimizer (and I suspect with this too) is the fact that Google undercounts conversions by about 20% on average. With conversion optimizer, you can account for this by raising the target CPA in step, but it doesn't look like that would work with eCPC. I guess everything's relative, i.e. one keyword's eCPC bid adjustments would be relative to its overall conversion count vs. an equally flawed, more poorly converting keyword. Same difference.
Peri says:
I would be very curious to see how this a years time has changed these figures.
Peri, I'm not sure when we'll get around to this exact topic again for the RKGBlog, but we did cover some similar ground here: http://www.rimmkaufman.com/blog/kindle-fire-conversion-rate-worse-than-iphone/08022012/


Check out what others are saying...
[...] This post was mentioned on Twitter by Rimm-Kaufman Group, Long Tail Tool Maven, Long Tail, SEM Software, Meta Tags and others. Meta Tags said: Online Sales Potential by Browser: Safari Users and Upgraders Lead: At RKG we are experimenting with using Googl... http://bit.ly/g6kxAY [...]
[...] above. And they’ve got a great post demonstrating those types of questions where they look at increased sales due to customers using specific browsers. [...]
[...] reading Mark Ballard’s post on the Rimm Kaufman blog about the possibilities with Google’s new “Enhanced CPC” [...]
[...] on March 5, 2011 by admin   ·   No Comments After reading Mark Ballard’s post on the Rimm Kaufman blog about the possibilities with Google’s new “Enhanced CPC” [...]
[...] reading Mark Ballard’s post on the Rimm Kaufman blog about the possibilities with Google’s new “Enhanced CPC” [...]
[...] published tons of research on conversion rates by device showing iOS devices convert better than others which suggested to us [...]

Leave A Comment