Adwords Search Query Reports, US vs. The World: The Followup

A couple of weeks ago, I wrote about a client of ours and the vast difference in Adwords search query report numbers for the same 2 keywords in the US vs. in Germany. It was the most-commented post on this blog in recent memory – not only did people commiserate, but they asked great questions to try to get to the bottom of the mystery.

I was able to answer several of the questions right away (although, of course, there are no bad questions, and I thank everyone who commented for helping me think everything through). But 3 questions came up to which I didn’t have the answers:

•    Is the client brand a German word?
•    How does traffic break down by device?
•    How does traffic break down by Google vs. search partners?

The first question I had to look up, but quickly realized that the answer was no, it’s not a German word. Of course I had to run some reports to answer the second and third questions, and the results were interesting indeed.

Searches By Device:

I re-ran the SQRs for each country, and segmented impressions by device. I suspected there might be a big difference, but alas, the two countries were nearly identical:

us by device
germany by device

Germany did have a slightly higher percentage of mobile searches than the US: 11% vs. 9%, but the difference isn’t statistically significant. Clearly, devices are not the reason why US search queries were so much higher.

Searches By Network:

I noted in my previous post that session-based queries were high in Germany compared with the US, and that prompted someone to ask about distribution by network. We’ve all seen questionable websites, sites that aren’t really search sites, lumped into the search network. Let’s take a look at the segmented report data:

us by network
germany by network

Bingo.

I have to admit, I was stunned to see the difference. I knew that the search network was probably more robust in the US than in other countries, but this is downright horrifying. Only 2% of the impressions in Germany on the brand terms came from search partners, but fully 30% in the US were from partners. And just to refresh your memory, the majority of matches in the US were broad match, as compared with Germany where the match types were more evenly distributed.

I smell a rat. I’m still not convinced that we’re getting fair treatment here in the US.

And for heaven’s sake, if you haven’t already done so, please go sign the petition to allow separate bids for search partners. This is something I’ve wanted for a long time, and clearly it’s long overdue.

Or perhaps Google is deliberately holding out on giving us separate bids for this very reason? What do you think?

Related Posts:

Comments

  1. *smug face*

  2. Lisa Sanner says:

    Great sleuthing Melissa. Have to say I’m not shocked by the results. Darn search partners, would be nice to call a rat a “rat” or a cute little pet depending on the outcome (conversions). Everyone -please sign that petition. Power to the people!

  3. Somewhat anecdotal evidence from me here in the UK – but from what I’ve seen, CPA bidding *seems* to reduce bids for search partners when the conversion rate is not as good…

    • Melissa Mackey says:

      Good point Jordan. We’ve been using CPA bidding more and more because of that, even though it can sometimes reduce overall volume. I’d still like the option to set search partner bids though!

      • I agree, separate bidding on search partners would be better.

        It can take a fair bit of time by the looks of things to reduce bids by which time you’ve already wasted a lot of money when you could have started out with lower bids for search partners.

        Martin – ‘auction level bidding’ – is that what it is called (different bids for different search networks)? (I thought your keyword/ad entered into an auction every time prior to being shown anywhere so all bids are ‘auction level’). Anyway – I know what you mean…

        • Not sure it’s called ‘auction level bidding’, but I think it comes close. When an ad is eligible, Conversion Optimizer calculates the chance of a conversion and bids accordingly. While everyone else has to bid at keyword level, Conversion Optimizer sets the bids right before the auction occurs. So if there is enough data to suggest that search partner clicks have a 20% lower chance to convert, then the bid would be 20% lower. The same adjustments can be made with regards to search query (instead of just keyword), location, browser, etc. That’s why the tool, despite all limitations, is so powerful…

    • Yeah, that’s how it should work – bids at the auction level…

  4. Very interesting to see the difference (from the other side, in my case ;)
    A possible reason for this may be that the search engine market is more diverse in the U.S. compared to other western countries like Germany. Comscore had Google at 96% market share in Germany (other sources report something like 80%), leaving little for other search engines, including search partners. The charts above might be a nice visualization for Google’s monopoly in some places…

    • Melissa Mackey says:

      Absolutely agree, Martin – it goes both ways! A monopoly can be a bad thing, but so can a bigger search partner network. :) Thanks for all your insight – it has been really helpful in digging down to what’s really going on here.

      • Thanks Melissa, I enjoy being able to talk about German PPC for a change ;) By the way, the nice thing about Google’s monopoly is that you can focus on one engine. When I get a glimpse of the whole BingAds universe I’m somethings thankful that I don’t have to know all of that, too…

  5. Matt Graves says:

    I just tried this yesterday, Keword and Ads Tabs (it works on all but kw and Ads is granular) > Segment – Top vs. Other. What I thought was Avg. Pos. in Google was actually Avg Avg Position with a meddle of Google and Search Partners. The Top vs. Other also nice on the Keyword tab because the option to Segment by Networks is not offered.

    • Melissa Mackey says:

      Yeah, all of the Segment reports yield some pretty enlightening stuff. All PPC managers should look at this info on a regular basis.

      • It also depends I guess on what you are advertising…

        Here in the UK if you are advertising the sort of products that people search for on ebay and amazon then you can get a crazy number of impressions from search partners (are amzon and ebay as big in Germany, and are they search partners with Google in Germany?).

        Here is an example for impressions from one of my campaigns in the UK for last year – Google Search=959,000 – Search Partners = 3,477,000 (clicks were fairly similar). The same campaign in the Republic of Ireland actually had fewer impressions from search partners than from Google. You could draw a tenuous and not very scientific conclusion form this that people are much more likely to search for products on Amazon/Ebay in the UK than they are in (Republic of) Ireland. Maybe similar in other countries?

        • Yes, Amazon and eBay are just as big in Germany and this sounds very familiar. With the brand not being a Germany word it’s unlikely to be included in an eBay category name. Otherwise, with the breadcrumb trail being used as a substitute keyword, impression numbers can go through the roof, just because of eBay.

  6. Sorry – my replies seem to be all over the place…

  7. Yay search partners?!?!?

    Thanks for digging into the data, Melissa.

  8. Nice follow up, Melissa. Thanks for keeping us all in the loop. Signed the petition! Crazy partners…

  9. I just came across this article. I honestly think the difference is easy to explain and comes down to language, culture and education. For starters, German is a much more structured language and conforms very strictly to it’s rules which are revised regularly and the subject of public debate. Not only that, but it has about half as many words as English. In that context, there is less scope for misspellings, mistakes and variants. And indeed I do see far, far less misspellings in my German PPC accounts than my US ones.
    Add to that, the average German is far better educated than the average American. (I’m a Brit, but live next door in the Netherlands and have no vested interest in this, nor anything as a Brit to boast about.)

    Moreover, Google’s algorithms certainly understand languages in different ways. Until recently, using their keyword tool in Dutch repeatedly gave me phrases including prepositions as though they were adjectives and I watched Google Search struggle to figure out what would have been a simple query in English. It’s much better now.

    Finally, of course less Search Partners. Google Search is almost ubiquitous in most European countries, having a market share in the high 90s and not so many alternatives to go to if we wanted to.

Speak Your Mind

*