In a wide-ranging keynote discussion at SMX Advanced in Seattle, Wash. last week, Google’s head of search ads, Jerry Dischler, talked new products and features and what the newest innovations reveal about where search marketing is headed. (Watch the full interview below.)
Asked what the big themes or trends were that drove many of last month’s product and feature announcements at Google Marketing Next, Dischler said changing consumer expectations and shorter sessions on mobile manifested itself in three key areas: audience, AMP and attribution.
Attribution: Supporting the move away from last click
The new Google Attribution product is the company’s answer to upper and mid-funnel blind spots created by last-click attribution. “Attribution is particularly important with mobile because in a last-click environment, marketers could be over-weighting desktop for example, relative to other channels,” said Dishler. “Marketers are saying 2017 is really the year of moving away from last click, and we want to be able to support that.”
Data-driven attribution, Google’s machine learning-powered attribution modeling, is emblematic of how the company is thinking about and incorporating machine learning into core background processes and marketing functions. When enough data is available, the solution automatically chooses an attribution based on numerous signals.
Machine learning: Less drudgery, more strategy
Asked about the role of machine learning and the inherent tension between automation and control, Dischler said, “In general we want to use machine learning to make advertisers’ jobs easier.” Advertisers can spend less time in the drudgery of manually setting bids and instead think strategically: What’s the most effective messaging? What are the highest value customer segments? How do we support new product lines and initiatives? How do we make the website better?
From a job perspective, search marketers will be able to move faster from looking at transactional value to customer lifetime value. “A lot of companies have not been able to get to that point because they are still so caught up in the mechanics of doing online advertising at scale.”
The industry has been discussing and experiencing the impact of automation for several years now, but the rate of pervasiveness of machine learning in search advertising has accelerated rapidly over the past year. For marketers that had been used to having a lot of manual controls and been the guinea pigs for machine learning, trusting the system and the algorithms can be difficult.
Don’t expect to see how the sausage is made
“We try to have controls when possible and transparency in reporting,” explained Dischler, “but not everything can be reported on. With deep neural nets, you can’t explain everything that’s going on. This is not just for search advertising, but is pervasive in industry. We’re going to have to be able to get comfortable with looking at inputs and outputs and not necessarily what’s going on in the middle. … If we do this right, we can arrive at better outcomes.”
As an example of this, Dischler pointed to the recent change to close variants for exact match. In testing that, Google says it was able to deliver 3 percent more conversions at a similar ROI. “We have to make these kinds of changes because it’s good for everyone. It’s better for users because they get better ads. It’s better for advertisers. And it’s better for Google.”
Audience ‘works for search’
Dischler said there has been a natural progression in using search data for targeting: RLSA, customer match, similar audiences, and now in-market audiences. “The reason why is because audience works for search.”
Asked what was in store for audiences in search, Dischler said there are a few other things in the works we’ll be hearing about, but also that Google has a lot of runway with current features — offering more in-market segments, building on similar audiences in terms of data to make them more actionable and flexible, for example. Google will be building on these existing offerings over the next six to 12 months.
AMP: ‘It pays to be fast’; improved tracking coming
“The idea of AMP,” said Dischler, “is to make the mobile web as responsive as apps — or more responsive than apps — while being a whole lot more flexible. And we want it to work for advertisers, too.”
How should advertisers be thinking about AMP versus responsive? “First and foremost, advertisers should have great mobile pages. Focus on that first. Then look at how long it takes for mobile pages to load. What does your bounce rate look like? What does your mobile conversion rate look like? It pays to be fast,” said Dischler.
There are several functions still to be supported in AMP, including some conversion tracking. In a current beta, advertisers can point their mobile search ads to AMP landing pages. Dischler says they are working to make all forms of conversion tracking work within AMP. “We have some work to do on the Google Analytics side, and we’re working through that, so expect to see some launches soon over the next few months.”
For those wondering if or when they should get started with AMP, Dischler encouraged advertisers to start experimenting with AMP with a fraction of their site to see how it does. On average, pilot advertisers saw 5x to 7x improvements in page load time and increases in conversion rates. “Granted, it’s more work in terms of landing pages, but it can be worth it.”
Online-to-offline attribution: Mapping ad clicks to store visits & purchases
Dischler discussed the two ways the company is measuring and reporting on store purchases linked to ad clicks. One is when retailers upload loyalty or other customer email lists to map ad clicks by signed-in users to purchases in aggregate. The other is a solution introduced last month that is only offered in the US. It matches Google data with financial partner data in an encrypted fashion that enables Google to extract aggregated numbers of store purchases and transactional value by store location. Those conversions will automatically show up in advertiser accounts when there is sufficient data.
That automation also occurs with store visits data. Google has measured more than 5 billion store visits to date and aims to get more granular. “Our goal with this is to make it available to more and more advertisers. As the stores get smaller and smaller and the transactional values get smaller and smaller, we want to do a couple of different things. We want to make it available in more countries. We’re deploying bluetooth beacons at scale so we can increase confidence levels that people are in a particular location.” This is something Google has been working on for more than a year now.
New Safari attribution challenges?
Earlier this month, Apple announced it will be building on its efforts to block third-party trackers in the next version of Safari. (I believe the change could potentially benefit Google.) Asked what impact that might have on advertisers generally and Google specifically, Dischler made the following points:
It’s early days, so we’re still working through it internally, as well as with Apple.In terms of Google Analytics, “… it’s one-key tracking, so you’re in good shape” if you’re using that. We’re still trying to figure out what to do with AdWords and DoubleClick conversion tracking and remarketing“From an industry perspective, Apple has been very clear about why it wants to do this, but there are unintended consequences. Cookies are a mechanism that gives users a high degree of choice. There are alternatives to cookies that give users no choice, like fingerprinting. We believe users should have a choice. The action from Apple is putting a lot of pressure on publishers and advertisers who are trying to run their business within a framework that works pretty well. I think it’s reasonable for advertisers to be able to track conversions, and I think it’s reasonable for advertisers to be able to access their customers on other sites. We’re going to try to work it out with Apple.”
Change is OK: The new AdWords interface
If you’ve recently received access to AdWords Next, the new interface, and found the change hard, stick with it, says Dischler. (And I agree from the experience I’ve had with it over the past year.)
“We wanted to be able to build a platform that on top of it we could build assistive flows to make it easier to do core tasks,” Dischler says. “We cleaned up core work flows, improved reporting and added new features.”
Dischler says advertisers find they can get work done faster. My favorite aspect so far are the visualizations in the Overviews that make it much faster to get insights; Dischler says to expect more of that. “We want to do more proactive and assistive things to offer a better experience and make it easier for people to do their jobs.”
Watch the full interview below.