• Jeff Collins

    Isn’t it sad that as advertisers we have to come up with workarounds in order to get the granularity we need at the mobile level when it could be, and is currently, easily available to us? I can’t believe there isn’t more of an uproar about this change. When implementing these workarounds our clients’ accounts are going to become way more disjointed and convoluted than they were when we had mobile segmented campaigns.

  • http://twitter.com/Kevin_Lee_QED Kevin Lee

    Interestingly, at Didit.com we adapted a click routing technology initially designed for landing page testing to route clicks based on device. Tat way if the advertiser doesn’t have an adaptive design site or can’t handle dynamic click routing or personalization of the user experience on the site but does have a tablet-specific site, we can route the clicks appropriately.

  • http://twitter.com/bgtheory Brad Geddes

    Hi Fred,

    So one of our first enhanced campaign tests is doing the segmenting of keywords by bid multipliers trying 10% and 25% increments (so we went from one mobile campaign to about 20 in some cases). The issue isn’t the setup. The issue is trying to manage the structure post launch. Having an easy way to look at the CPA data, the current multiplier, then seeing what keywords need to switch buckets is not well supported to any software, and is very tedious in Excel. I agree its a workaround; but in some cases, its so hard – I can see where advertisers will give up trying to manage it and then Google will just see people using their system and think its all OK.

    The other issue is that CPA bidding does not do a good job of taking device into account. I did an audit recently where the company forgot to put the conversion tracking script on their mobile website, and campaign was targeted to all devices. They were still receiving 25% of all traffic to the mobile site, even though there had never been a mobile conversion in 6 months. CPA bidding works well if you are good at segmenting the data into multiple campaigns. If you don’t segment, it can still work well; but the results won’t be nearly as good as someone who did segmenting first.

    Any thoughts on how to manage keywords by bid modifier moving forward?

  • http://twitter.com/ShockleyAu Shockley Au

    Hi Fred,

    Great post! I know that all the changes coming via Enhanced Campaigns has search marketers looking for ways to retain control of running how their campaigns run, although to be fair, there is also a lot of exciting features coming with Enhanced Campaigns.

    With regards to delivering a tablet/mobile experience to a user based on device, assuming your site isn’t responsive, would it make sense to use dynamic serving ( https://developers.google.com/webmasters/smartphone-sites/details ) rather than a ValueTrack parameter?

    It would just be a section of code you add site-wide, and you don’t need to manually {device} tag all your destination URLs in AdWords, or set up profile filters to clean up all the ‘?device=’ from your Analytics reporting.

    Looking forward to hearing your thoughts!

  • http://www.facebook.com/kerschbaum Joseph Kerschbaum

    A question on Solution 1, Duplicate Campaigns: Can you decrease desktop/tablet campaign bid modifiers by 100% (“Then, by setting the mobile multiplier of the ‘Desktop’ campaign to -100%, all mobile ads will be served by the ‘Mobile’ campaign. “)? Therefore, with this change, you have created a mobile-only campaign? When I access this screen within AdWords, I see my mobile modifier but I don’t see a modifier for desktops. Screen shot of mobile tab: http://screencast.com/t/ibphY7kmU

  • http://www.toptiertools.com/ Frederick Vallaeys

    Hi Shockley,

    You could definitely do dynamic serving at the page level but then you’d have to maintain a list of user agents which comes with its own set of challenges (https://developers.google.com/webmasters/smartphone-sites/detecting-user-agents). I guess we all just need to figure out for ourselves what level of inconvenience we’re willing to deal with in exchange for the possible benefits.

  • http://www.toptiertools.com/ Frederick Vallaeys

    Hi Brad,

    You’re right that the maintenance of the extra campaigns is going to be tedious if you find that conversion rates fluctuate significantly and you need to move keywords back and forth between campaigns with different mobile bid multipliers. Have you found that you can still get good-enough results with fewer than 20 tiered campaigns? The fewer tiers you run, the less you’ll need to move keywords around but it will of course come at the cost of optimal ROI.

    I haven’t seen any of the bid management systems build support for this strategy into their tools but maybe Marin can take some of its IPO money and build this for us :-) Or for those with access to the API, they can build this themselves. And for those without the API, AW Scripts can be used to build a semi-automated solution.

  • http://www.toptiertools.com/ Frederick Vallaeys

    Hi Joe,

    No, there’s not desktop bid multiplier unfortunately… The idea is to force the “Desktop” campaign to be desktop only by setting a -100% multiplier for mobile. Then by setting all the bids in the “Mobile” campaign lower than those in the “Desktop” campaign, Google would most likely only serve those ads on mobile devices since they would never outrank the equivalent keywords from the desktop campaign where all the bids are higher.

    Hope that makes sense…

  • http://twitter.com/bgtheory Brad Geddes

    Hi Fred,

    For our accounts that are currently segmented by operating system and device; we’re going to lose so many conversions (our CPAs range by more than 10x (roughly $10 to $120) based upon operating system & carrier); that we want super fine control in order to see if there’s anyway not to lose too many conversions.

    We’re examining all the search query data between operating system and device to see if there are subtle differences by demographics that we can take advantage of in order not to lose too many conversions. We’ve found some already, but its a lot of comparison work.

    In those campaigns, I think we’ll end up with many tiers of bids.

    For others, its a practical matter of using 25% ranges not because we want to use such large ranges; but because of the ongoing management issue.

    We’ll probably build this reclassification function with the API; but first we need to manually bid it for a while to understand all the rules before we lay out the system specs.

    When you think of these tiers. Are you taking the CPAs, normalizing the data by device as if the adjustment wasn’t used; and then classifying the keywords into tiers for ongoing management? Or, are you taking the actual CPAs, comparing it to your target, and then reclassifying based upon current targets?

  • Adrian Huth

    talking with Google reps has sure been pleasant lately let me tell ya. I feel bad for them actually as they are now being paid to push something which the advertiser pushes right back on and now best practices become workarounds to them shoving non converting traffic down our throats. every week we basically have the same conversation now and make no progress on growing the accounts.

  • http://twitter.com/ShockleyAu Shockley Au

    Agreed. Personally I prefer responsive web design, although that often requires rebuilding a website. New websites should be built responsive from the start, though.