Keyword Not Provided – Maybe Mobile Is To Blame?

SMX East - Up Close Logo

At SMX East yesterday, during the Trends in Mobile Search session, Google announced the addition of cross-device conversion reporting in AdWords and I believe this may be related to the loss of referral keyword data in Google Analytics.

The new method of ad attribution helps advertisers understand cross-device and cross browser conversions, and eventually will also include attribution for click-to-call and offline conversions—a panopticon of conversion tracking! It is great for advertisers, but I believe it may have pushed mobile privacy concerns far enough to cause Google to pull back the keyword data in Google Analytics (and may possibly have caused the recent hiccup in Google Webmaster Tools as well).

This new feature in AdWords was described by Andy Miller, Head of Mobile Search Sales and Strategy at Google. According to Andy, Google will be aggregating the cookie data that they collect from signed in users who are accessing the web while logged into any of their Google accounts or apps (this includes Google+, gmail and other websites and applications behind the Google firewall).

This is a pretty monumental change in the level of data and clarity that Google provides to advertisers, and really could be a game changer for digital marketing.

As the number of mobile devices increases, so does the potential number of devices where a conversion might happen. This means that customers may be researching a purchase on a computer and completing the conversion on a tablet, phone call or even offline. Previously this type of multi-device shopping/researching behavior could not be compiled together in a meaningful path or funnel across devices, but now it can, so there is clearly a benefit for marketers.

At a very basic level, this seems like a brilliant, winning proposition for Google. Building these conversion extrapolation models and reporting on this data has a clear value for marketers. It may make PPC advertising easier and more meaningful, since keyword-level data is still provided in AdWords (but as we know, it has been removed from un-paid referral data sources).

Mobile marketers are used to hazy keyword referral data in Google Analytics, and have been since the launch of iOS6, when all search-bar data went to [Not Provided], and appeared to be “direct” traffic, rather than a Google referral. With the  launch of iOS7, we briefly got accurate iOS referral data back (though keywords were still [Not Provided]), but this was only weeks before all keyword referral data—mobile and desktop—became [Not Provided].

To me, it is hard to believe that all of these things are not related. Now it seems like Google is building up the level of data that they provide to paying advertisers, but intentionally keeping it from other marketers who are unwilling to pay.

Since Google’s new conversion metric called “estimated cross-device conversion” ties together information from all the devices where people may log-into a Google account (tablet, phone, home computer, Web-TV, office computer, etc.), there is a serious decrease in user anonymity and the perception of privacy.

The public reaction to this new omnipresent reporting structure that will eventually even track when we make a call or go into a brick and mortar store could really be bad. Increasing ad targeting based on deep customer information sounds great to marketers; however, to politicians, journalists and your average person, tracking to this level of detail is not exciting, and is potentially really invasive.

Google’s push to Enhanced Campaigns earlier in the year was quite disruptive for many advertisers because they lost the some control over how their budgets were divided across desktop, mobile and tablet traffic—it removed the ability to actively device-target ads differently for mobile and desktop search.

As most digital marketers know, not all clicks carry the same value and without full control, many advertisers were frustrated. Many webmasters began trying to protect their desktop budget by changing the percentage of mobile ad serving on their desktop accounts to (-100%) in an attempt to subvert the new “enhancement.”

Now that conversions are tracked and attributed more accurately across different devices, maybe the disgruntled advertisers will be appeased. It is possible that the extra effort and protective campaign optimization for cross-device ad distribution will no-longer be as necessary.

Advertisers may even find that when conversions are tracked more accurately the enhanced campaign structure is actually more beneficial. It is hard to know, but perhaps the new conversion data will even justify the automated changes that happened to everyone’s ads when enhanced campaigns originally launched.

Proper conversion attribution is vital for calculating a true ROI, but lots of data is lost when searchers change devices or continue a search offline—so this is a welcome change– at least, on the first glance.

The problem is the use of the word “conversion” in the name. I think it is a bit “off” and may unfairly inflate success metrics that help Google sell advertising. This is because all calls and store visits will be considered a conversion, even if there is not a transaction.

Unfortunately, when there is no financial transaction, it is not really the ideal conversion that marketers want to track—it is more like a part of the funnel. If marketers start to use this “conversion” metric, their ROI information could be detrimentally skewed.

John Busby from Marin Software also spoke in the session emphasizing this point in Q&A, reinforcing the idea that additional tracking would be necessary to show data about true, transactional conversions and to get other information about the customer experience.

In the case of a click-to-call conversion “other information” would be specifics about the phone conversation or IVR interaction that took place –which is not part of Google’s new reporting. I agree and think this observation is even true in a broader context, because in-store visits and transactions will certainly still be difficult to track, analyze and attribute properly for an accurate ROI measurement. While it is a great boon in data, it will probably still be more “directional” than it is “precise.”

In the same session, Jim Yu from Brightedge talked about the trends in mobile keyword research and conversion. He explained that the average conversion rate on a smartphone is 0.3 to 1 for desktop but further explained that mobile conversion varies dramatically by industry.

According to Jim, the entertainment and media category average 1.6x better conversion on mobile when compared to desktop, and a couple other categories were also notably high. Since mobile search trends can change pretty rapidly, each mobile strategy must be planned to focus the effort and the budget where there can produce the most bang for the buck, so accurate reporting is essential.

Google’s new cross-device “conversion” estimates will be based solely only on information in the AdWords account, so they don’t reference industry averages or aggregated norms like the statistics that Jim presented — which could limit the amount of data that is provided to small campaigns.

The tool will only present you with the data if the system has a 95% level of confidence, so until there is enough device-level conversion data in your campaign, Google will leave the information out.

Hopefully, the new cross-device conversion data will help mitigate the negative impact some advertisers have experienced with Enhanced Campaigns and add legitimacy to the mobile marketing budgets that they have been suggesting. While the change may make life in the PPC world easier, if my suspicions are accurate, it certainly has made life in the SEO world a bit more difficult.

If indeed, this change was the impetus for limiting access to organic keyword referral data, the next big question is this: Will the addition of more accurate cross-device attribution drive more SEOs to start AdWord campaigns, or will SEOs do it simply as a means of buying back access to keyword data? Either way, it seems that Google will reap the benefits!