How to Measure and Optimize Gmail Ads’ Performance

On September 1, the official Google AdWords blog posted that their ongoing beta, which according to the post was tested for several years, will be finally enrolled for all advertisers to use through their AdWords account dashboard.

Six months ago, my account manager from the Google AdWords team called and asked if I was interested in testing their new product. I was excited to test it and in about a week, our Gmail Sponsored Promotions (GSP), which was the name of the product while in beta, went live. After struggling to figure them out, I came up with a method to measure their performance and optimize creatives and targeting methods.

I must disclose that I’ve never run an email marketing campaign (apart from a newsletter for my old .com local blog I used to run back in 2013 — not available any more). After gathering some data about open rates and CTR from colleagues and the web about email marketing, I started analyzing the performance of the campaigns.

I understood pretty fast that the Gmail ads are not the same kind of direct response campaigns as Google search (obviously), and as the startup I work for is an eCommerce marketplace, it meant that I much preferred to buy search clicks instead of trying to direct top funnel traffic to an unfamiliar online booking marketplace. In addition, the fact that remarketing lists targeting is not allowed yet, I was about to shut the campaigns off for a long while.

But before giving up, I decided to change direction and try promote the supply side of our marketplace through Gmail ads. I decided to reach prospects by targeting Gmail users who received emails from online travel adenices (it’s called Domain Targeting, it’s available through the display keywords tab), and educate them about the opportunity they have to sell their reservations on Roomertravel.com in case they got stuck with a non-refundable booking. It is more of a branding/increase awareness kind of campaign, and I decided to follow Google’s famous “Brand Lift” strategy that is in use by huge brands as their YouTube video ads strategy.

I used specific generic keywords both in the teaser ad (or as it’s officially named in AdWords , “Collapsed Ad”) and in the expanded ad itself. By the way, I used generic keywords because I operate within a new and not yet familiar enough market. If that isn’t your case — you better be using branded keywords; otherwise your competitors will enjoy the fruits of your efforts. After the ads started getting traffic, I measured the search volumes of those keywords in my search campaigns. My best performing generic keyword increased 50% in impressions count during a two month period!

A “Collapsed Ad” in Gmail, or “Teaser Ad”

The “Expanded Ad” in Gmail

So this is a way you can measure the performance of Gmail ads. Insert relevant keywords (usually branded, and generic only if it’s a new market or you have zero competition) and track the impressions volumes of the same keywords triggered in a search campaign.

Now you ask: But how did I optimize Gmail ads campaigns?

If you still haven’t tried them out, I will tell you that they come with three new metrics that you can find at the modified columns dropdown. The three metrics are: Gmail saves, Gmail forwards, and Gmail clicks to website.

Another metric, a slightly more familiar one, is the CTR (just kidding, of course). A click in a Gmail ads campaign may occur in two situations:

1. The user opens your ad.

2. The user saves your email without opening it.

We have no way to know if a user saved your email with or without opening it first.

Having said that, I decided to treat the CTR as the, traditionally known by email marketers, “opening rate.” It means that the ad teaser (collapsed ad) was completely measured and optimized by this metric alone. An increased CTR wasn’t followed by a necessary increase in the Gmail saves rate, so I believe that most of your Gmail saves would come from an opened email, and that a CTR is a suitable metric to measure Gmail ads “open rates.”

The new Gmail Metrics in Adwords Reports. To add them go to the Columns tab.

Having a small budget, I was not ROI-oriented, so I only optimized for higher CTR. But I can definitely tell you that on this medium, as well as in search and display, the CPC was strongly correlated with the CTR, and a higher CTR significantly reduced your average CPC.

It’s only a little test I made. Not statistically significant. Do not rely on that, test it yourself as well. Always Be Testing!

Now here is what I did with the three brand new Gmail metrics:

To be able to attribute conversion for specific Gmail metrics (e.g., Gmail forwards), you’d have to collect tons of data and conclude that ad group X drove low CPA and had high a Gmail forwards rate. In that case, Gmail forwards would be an important metric for you and you should optimize the campaigns towards it. But if you’re operating on a low budget or using Gmail ads with a “Brand Lift” strategy like myself, it wouldn’t be effective to optimize by this method.

So what did I do?

Well, I combined a combo metric out of the three metrics. I instantly called it “Gmail engagements.” This combo metric, as you can imagine, is a simple sum of the three original ones.

Now every ad group, keyword, and ad had its Gmail engagements metric, and of course its Gmail engagements rate!

Here is the formula:

Gmail engagement rate = (Gmail saves + Gmail forwards + Gmail clicks to website) / Clicks * 100

That helped me to decide which keywords (mostly domain targeted keywords) were the best ones, and on a weekly basis I paused the lower 25% performing keywords. I also used it to optimize the creative. There are four ad types in Gmail ads and I used this method to figure out which ones generate the best Gmail engagement rate.

Now, you!

What types of optimization methods do you use for Gmail ads, or email marketing campaigns in case you haven’t tried Gmail ads yet?

Cheers,

Yuri