Back in August, it was announced that Google would allow device bid adjustments on computers and tablets, and not just on mobile devices. With this announcement came the capability to segment campaigns once again by device. But creating new device specific campaigns is a lot of work and can potentially overcrowd your account. You need to determine if this tactic is right for you. Should you be segmenting your campaigns by device?

To answer this question, a few other questions need to be answered. First, why would you segment by device? You can already bid differently per device. Well, how drastic are the bid differences for what you’re willing to bid across the different devices? Ad copy, for now, is fine, right? We have regular text ads and mobile text ads. Remember, expanded text ads are taking over and it may be that ONLY expanded text ads will be running come January.

With expanded text ads there are no mobile specific ads, but rather mobile URLs to send mobile users to specific mobile pages. With that said, you need to ask, is my mobile specific copy crucial to the success of my mobile advertising? Finally, how crowded is my account and will the extra build hurt or help the organization of the account? Adwords allows 10,000 campaigns per account, but that doesn’t mean you want to have 10,000 campaigns in your account.

Will the segmentation of devices help or hurt the account structure you have in place? Do you or your client like to visualize device performance often? With these questions, we should be able to answer if you should be segmenting your campaigns by device.

How Drastic Are The Performance Differences By Device?

When I am looking into device modifiers, I check one of two statistics. For lead generation clients I look at the conversion rate differences between devices because if the conversion rate for mobile is half of what we’re seeing on desktop, it would make sense to pay around half per click in order to land at the same CPL. For ecommerce, I look at revenue per click numbers with the same thought in my mind.

With this being said, take a look at how desktop, tablet, and mobile perform in your account. Below is an example.

Device Performance

In this example, CPL looks to perform better on desktop and mobile, while tablet shows a much higher metric . However, this is due to the fact that we’ve been able to pull bids back on mobile devices, while we have not been able to pull bids back on tablets. With the capabilities to pull back on both, the CPCs on tablets and mobile devices should better align moving forward, and therefore even out the CPLs. For this reason, you should make adjustments based on conversion rates and not solely CPLs.

With this specific situation, we can keep everything in the same campaign, but it may be best to create individual desktop and tablet/mobile campaigns due to the rate at which they convert. Why would you need to segment these? Similarly to the review on RLSA Efforts having their own campaigns, it is much easier to get the entire story when items are segmented.

For example, let’s say in the campaign listed above we have a 75% impression share. Our desktop impression share and average position could be 100% and 1.0, while our mobile and tablet IS and average position land at 50% and 3.0. In this case, all of our growth availability when looking at keyword level bid changes would be on mobile devices, but when we’re looking at data at the aggregate keyword level, we’re not necessarily segmenting this granularly. So, we raise the bid and therefore get more mobile and tablet traffic, which does not perform as well as the desktop traffic we’re getting and could potentially pull our numbers down.

With this analysis, the recommendation would be if the conversion rates on one device to another are close to double, it may be smart to segment devices into their own campaigns rather than relying on bid modifiers for bidding purposes.

Do You Use Mobile-Specific Copy? Is It Crucial To The Success Of Your Mobile Advertising?

Most would say that mobile copy is the biggest reason you should segment mobile into their own campaigns. Expanded text ads are not mobile specific and in January they will more than likely be the only option to use. However, others would argue (including maybe Google) that the biggest reasoning for mobile specific ads in the first place was to send those users to a different URL – and this problem was fixed with the mobile final URL launch. Well, what are your tests telling you?

It can be tough to segment your mobile copy if the copy or the landing pages aren’t helping the results. If you always had mobile landing pages in place and then tested “order from your phone” type copy, it would help you make a decision much easier.

In many cases, users launched mobile specific ads with mobile terminology along with the mobile landing page right off the bat in order to follow best practices. In this case, you can segment device performance, and add device preferences as a column in order to create a pivot table similar to the one below with mobile performance segmented out.


The results above show when we have presented a mobile specific ad on mobile devices compared to when we have shown a regular text ad on mobile devices. When testing ad copy, I typically look at conversion per impression data, while I look into conversion rate data when doing a landing page test (for lead gen clients).

In this particular situation, the mobile ads have mobile-specific copy and mobile-specific landing pages. With that being said, the higher conversion rate but much lower CTR on the mobile-specific copy shows me that the landing pages have helped performance a lot, while the copy appears to have actually hurt performance, even if it may have helped conversion rates slightly. Thus, we’ll be fine transitioning to expanded text ads and just adding the mobile final URLs, and so no segmentation would be needed at the campaign-level.

If these results showed that the CTR on mobile devices is dramatically higher when a mobile-specific ad was shown, then you would have to consider setting up a mobile-specific campaign in order to write the mobile ads within the expanded ad copy.

Will The Extra Build Outs Help Or Hurt My Account Structure?

The final question before making a decision on device specific campaigns would be how it would help or hurt the structure of your account. At Hanapin Marketing, many of our campaigns are built upon certain naming conventions that drastically help with account organization with the use of text-to-columns for reporting. In this case, you would just add devices to that naming convention. But do you already have RLSA-specific campaigns built out? Do you have multiple locations with their own groups of campaigns? When do the extra builds stop? These are questions you need to be asking as you make these types of decisions.

Is it more important to segment campaigns out by certain regions, by audiences we are targeting for remarketing, or by device? Is it plausible to do all 3 segmentations? Should I do different segmentations for different areas of my account? Using segments and dimensions gives you the data to look into the performance differences between different regions, performance differences between different devices, etc. Look into which segments are seeing the most drastic differences in performance and make changes accordingly. Maybe your account has little enough categories (ie: one product sellers) that you can segment everything out and have a naming convention that looks as: Network_Location_Device_Match Type_RLSA/Non-RLSA.

It all depends on how your account currently is structured and how segmenting more granularly by device would impact performance. There is a chance that your account has 3x the campaigns if you segment campaigns out for desktop, mobile, and tablets. Be sure to make the decision based on statistical backing and then move forward if the numbers make sense and the structure is manageable.