October 15, 2012
The most exciting AdWords feature to be rolled out in some time is Remarketing for Search Ads (RFSA). The “audiences” for this feature work exactly as they do with classic remarketing; the difference is that the ads may now appear in Google search engine results pages (SERPs), triggered by conventional keyword bidding.
“Behavioral targeting” has acquired a bit of a creepy reputation, seemingly overnight. It’s not clear that it’s inherently over the line. In the grander scheme of things, all advertising has the potential to be annoying, much like loud Broadway showtunes can be annoying in the wrong context. It’s a “feel” thing.
Some irresponsible marketers are cheering as they annoy people and laugh when they’re accused of “following people around.” But you can’t “educate” consumers out of feeling stalked. It’s a hard thing to measure, but every so often it pays to ask yourself if you’re using remarketing judiciously, or excessively. It can be as simple as using shorter date ranges or properly applying impression caps.
Last year, we heard a rumor that remarketing setup would become more powerful while also getting easier, in part because all the audience definition would be pulled from rich data available about users via Google Analytics. That would also simplify code installation.
It hasn’t quite worked out this way, but the principle of what is being deployed is similar to using Google Analytic’s power to set up audiences.
What Googlers are now calling the “new remarketing pixel,” or “smart pixel” is intended to go on every page of your site. That’s similar to how Google Analytics code is placed today. On the search side, this will allow more convenient audience definition. For example, you should be able to create an audience for “people who have stayed on the website longer than 90 seconds” to set a higher bar of engagement as part of your remarketing audience strategy. If not quite that, then you’ll certainly be able to specify multiple page visits, etc., with more ease as the interface is beefed up.
Reawakening Failed Keywords
RFSA should provide a nice way to scoop up a bit of conversion volume from keywords that fall into the “tantalizingly close to working” bucket. Like me, you’re probably sitting on paused campaigns in keyword areas that didn’t quite work. Let’s posit two main categories of “should work…really want them to work…but haven’t quite worked”:
- Diffuse-intent, broad words like “coupons” or “mortgage rate.” They’re not hitting your target CPAs. Too many people are coming in and using your site as a research tool. OK, CTRs, but a low conversion rate. You have to pause these words eventually.
- Keywords whose relevance (including CTR, but also “unspecified other” relevance factors) is deemed poor enough that the Quality Scores get set prohibitively and punitively low. Delivery is anemic; first page bids are in multiples of dollars, so you don’t see good results. Frustratingly, you can see the term is converting when the impressions and clicks do get delivered! But you’re unwilling to bid $4 on keywords that feel like “40-cent keywords.”
Often, the problem is a combination of these two problems.
How to Use It? Three Use Cases
The playing field tilts in your favor if you know the users have familiarity with your brand through previous interaction with your website. Consider the following examples:
- Category leadership words.Let’s say you’re in the gardening field, and you’ve had great success with product-specific keywords like “iris bulbs.” In addition, you have a passel of general category leadership ad groups – say “garden center” – that you’d love to keep in front of searchers. But they’ve failed in terms of ROI, so you shut them off. What if you keep showing them, but only to people who have previously reached the first page of the checkout process on your site (even as long as 400 days ago)?With the general searching public, you had been attracting too many clicks from people who were definitely only interested in a local, brick-and-mortar experience. With this remarketing audience, you’re showing ads to folks who seriously thought about buying online…from you. Big difference.
- Competitor words.The way Quality Scores work today, it’s costing more and more to bid on words associated with your competitors and their products. You’re Dodge, and you would love to show your competitive ads near search results when someone types in “Ford F150.” Those kinds of words, in the days when they were fairly priced, generated far higher ROI than the median ROI for the rest of an account. Too bad Google had to take away the punch bowl with such a clever Quality Score system. There will probably be numerous instances where this costly equation will improve for you if you’re just showing these ads to an audience of people who have previously visited your site. With future smart setup options, just show them to engaged visitors who viewed several pages of your automotive offerings.Especially in fields like B2B or higher-ticket retail (think cars, financial, motor homes, high-end travel), this strategy might prevail in spite of nasty-looking Quality Scores.The real kicker for the big companies that would love to try this out: they already spend millions of dollars on the multi-channel ads that get people to visit their websites in the first place, so the sites already generate very large audiences to use as remarketing fodder. In these cases, even highly specific subsets of these audiences amount to much more than trivial volume.
- Specialty products victimized by too many mass market searches.You sell high-end racing ski equipment, but in some product categories you’re inundated with non-buyers and cheapskates who visit the site and decide to spend all of $10. The problem is particularly pronounced on certain keywords, which (despite being relevant to some high-end racers) have suddenly become popular with a large influx of newbies taking up the sport. As a result, you’ve thrown in the towel on those particular words, simply pausing those ad groups.Now, consider reviving all of these products in a campaign that addresses only your remarketing audiences. Again, by showing the ads in high positions to such a narrowly constrained subset of people, you’re not getting perfection. But nor are you getting so many random $10 spenders. Turns out, this is an awfully nice way to remarket to “your kind of people” – a method that may allow you to reduce your reliance on (let’s say) remarketing in the display network by setting tighter impression caps over there so you tone down the “creep factor” while still doing something to leverage the benefit of past site visitation. You repeat your message, but without being perceived as harassing serious customers you want to treat with respect. More high-end conversions result.
It’s clear that Google has put some thought into avoiding excess “creep factor” when it comes to the remarketing ads that appear on its prize real estate on Google Search. Google’s working on editorial policies; for example, prohibiting advertisers from explicitly talking about “coming back” to the website. Instead, general references to offers, specials, and other new developments are recommended as subtler means of bringing people back.
This article was originally published on October 5, 2012 on ClickZ.com.