The Benchmarks – Retargeting Part 3

In our series so far we have addressed what Retargeting is and how it works (A Marketer’s Guide to Retargeting – Part 1), as well as how marketers are currently using retargeting, and what the trends and challenges are associated with that (Retargeting Best Practices – Part 2).

In our entry today, we’ll look to the benchmarks in search, social and display retargeting, as well as make some final conclusions.

SEARCH, SOCIAL AND DISPLAY RETARGETING BENCHMARKS

Researchers at Marin Software Inc. investigated the collaborative power of running search and social ads simultaneously compared to marketing each channel in isolation.

The results of their white paper “The Multiplier Effect of Integrating Search & Social Advertising” showed that consumers who clicked on search and social ads were more likely to buy and spend more. They also concluded that search campaigns were more successful when they were managed alongside social campaigns.

With this in mind, we decided to investigate if a similar outcome resulted when marketers retargeted across multiple channels. More specifically, we compared performance for advertisers running retargeting campaigns on Facebook and the Web concurrently, compared to advertisers who were only retargeting in one channel alone. We also looked at performance metrics for advertisers retargeting on Google search via Google Remarketing Lists for Search Ads (RLSA), although this data was extracted separate from the social and display retargeting efforts.

Google RLSAs drove higher Click-through Rates (CTRs) and lower Cost Per Clicks 1

retargeting search CTR

As expected, RLSA campaigns performed better than non-RLSA campaigns, with RLSA campaign Click-through Rates (CTRs) 2-3× higher compared to non-RLSA campaign CTRs. While it’s not a shock that the search retargeting campaigns performed better, one of the more exciting trends was the growing performance gap between RLSA and non-RLSA campaigns over the 3-month period.

There are a few reasons that could have added to this performance gain. Marketers could have enhanced their audience segmentation efforts by creating more targeted, actionable segments; they could have optimized budgets by raising spend on the higher performing segments, and minimizing spend on worse performing segments; or they could have improved their creative strategy and customer messaging to drive better results.

While the early results have been promising, and illustrate that there could be extra performance benefits to be gained as advertisers continue to identify new methods to leverage and optimize their RLSA campaigns, it warrants further study to see if these trends persist.

Retargeting Search CPC

1 The search retargeting data was compiled by taking a set of Marin Software users and comparing RLSA campaign performance versus non- RLSA performance within that set.

2 CPC numbers are based on an index. 100 is baseline.

Another interesting takeaway was that RLSA campaigns have better CTRs, but at the same time, they also returned those CTRs while delivering those leads at a lower CPC2 compared to non-RLSA campaigns. At first glance, this seems contradictory, because the basic RLSA bidding strategy is to apply a bid boost to RLSA audiences. If marketers are increasing bids on RLSA campaigns, CPCs should then be expected to be higher.

Yet, there could be a couple of influences at play here keeping the CPC down:

The RLSA list is a much more targeted audience, so there could be less contest for those clicks. Similarly, because RLSA CTRs are higher, the better performance may also be increasing the quality score for those ads, leading to potentially higher placements at relatively lower costs. Furthermore, because this data was pulled from a set of Marin Software clients, they may be garnering some of the benefits of the bidding algorithm, which allowed them to algorithmically optimize bids for their RLSA campaigns instead of applying a blanket bid boost. The key outtake is that RLSA has been delivering incredible performance. Not only have marketers vastly improved engagement through RLSAs, they’ve done so while driving down the cost of each engagement.

Marketers who retarget on both Facebook and Display enjoyed better performance

As outlined earlier, when Marin Software Inc. looked at the performance data for marketers running search and social (non-retargeting) campaigns, they found a multiplier effect when marketers optimized those campaigns holistically. When they looked at data of advertisers on their own Perfect Audience platform, they also uncovered a comparable effect when they were retargeting on both Facebook and Display channels, in comparison to only retargeting on just one of these channels.

Retargeting Facebook CTR Comparison

The statistics above shows that marketers who were retargeting on both Facebook and the Web at the same time, enjoyed higher click-through rates on Facebook compared to a group of similar marketers who were retargeting only on Facebook.

Retargeting Display CTR Comparison

The outcomes were similar for marketers who concurrently retargeted on both Facebook and the Web, compared to similar marketers who were retargeting only on the Web. In this scenario, the performance gap for display retargeting was even higher than the performance gap for Facebook retargeting data above.

BEST PRACTICES FOR RETARGETING ACROSS CHANNELS

Best Practice: Use search intent to build better cross-channel intelligence

One of the key benefits of including the search channel in cross-channel retargeting campaigns is that marketers can use search intent data to their advantage to create more targeted, higher-value retargeting lists.

To show you this idea in action, take the most basic retargeting scenario: A user visits a company’s web site, clicks through to the product page, and then leaves without taking an action. At this point, the company has the perfect chance to retarget those users through search, social, or display.

Retargeting Scenario

Usually, that’s where the case for retargeting begins and ends. However, without any information about the user’s intentions, it’s almost impossible to say what value the user actually has, and to predict what the right next step is. A product page visit by itself doesn’t guarantee that the user was ever interested in purchasing the product; the user could have been doing research to learn more about the product type, comparing the product’s features and price point to another similar product from another brand, or may have happened on the product page by mistake. Ut’s possible that they may not even be in the market for the product at all.

Without understanding intent, a marketer can’t be confident what the right communications message should be, what creative elements to include, or whether the user is even worth targeting at all.

However, if you can layer search intent on top of your basic retargeting lists, it will allow you to break them down and refine them further to give you more insight and increase the likelihood of conversion. By knowing that a person came in on a particular term, such as a price comparison term, the marketer could know what elements of the retargeting campaign they should modify to suit, whether it’s the retargeting window (shorter windows for those further along the buying cycle), the creative (featuring the product the user viewed), and the messaging/promotional language (a discount code or comparative claim). This level of information gives a marketer the absolute best chance of speaking directly to their individual customers in the most appropriate way.

Best Practice: Use Search Intent Data to Build User Identities

In a similar way, marketers can group certain search terms together to develop comprehensive user personas, even without adding data from third-parties in the mix.

For example, a women’s fashion retailer may have two main user personas they target:

Trendy and fashionable, price is no obstacle Classic and traditional, price-conscious Marketers could use this information to group a set of keywords that fits with those distinct target groups:

For the first user persona, keywords associated with luxury and high-fashion brands, specific product lines, and modern design language could be packaged into a dimension to better classify the persona. For the second user persona, keywords associated with classic brands, traditional colors, and price or discounts could be bundled into another dimension to separate this persona. Through this exercise, the marketer can then take advantage of these insights when they are ready to retarget these users across search, social or display.

An example – two users visit the “Ralph Lauren” section of a retail web site – a brand that could be considered both trendy and classic. One user enters the site using a search term for a color that is particularly on trend. The other user enters the site using a search term indicating some budget constraints. By using the search intent to segment these two consumers into the different persona buckets, the retailer can then retarget to them with the most appropriate creative and targeted language. The first user might be exposed to product ads featuring the “it” colors of the season, while the second user might see product ads that also highlight an offer for free shipping.

Best Practice: Maximize reach and optimize returns by selectively combining retargeting and negative retargeting across different channels

A common use case for Google RLSA involves increasing visibility to previous site visitors by widening the keyword list an advertiser might bid on. So for example, an advertiser who normally wouldn’t bid on a more expensive, broad term such as “NYC hotels” could decide to do so for the subset of users that had previously visited the advertiser’s web site. While this is one of the more effective ways to leverage RLSAs, it comes with a couple of drawbacks:

It doesn’t drive a whole lot of volume. It requires that the user first visit the advertiser’s web site, and also enter a specific search term the advertiser is targeting. More clearly, this approach doesn’t reach new users. While it may be an efficient tactic, it’s not one that will drive a lot of new business growth. For large retailers, spending in broad, higher-cost keywords like “NYC hotels” is necessary to obtaining new business and maintaining brand awareness. But what if you could increase your audience reach and be more efficient at the same time? Isn’t that the best of both worlds? This is where a blend of retargeting and negative retargeting can help generate not only return business, but reach new consumers more effectively as well.

Take our example above of the travel retailer bidding on a high-cost, generic term like “NYC hotels” on Google. The advertiser needs to bid on the keyword to appeal to new customers, but would prefer to avoid incurring multiple clicks on the same high-cost ads from the same users.

Retargeting Audiences Retargeted

How is this achieved? A marketer could bid on the initial click and pay the high CPC fees. Then, once the user has clicked on the ad and visited the advertiser’s site, the marketer could immediately switch to negatively retargeting that user on Google to minimize costs. At the same time, the marketer could start retargeting to this user on the web through the display ad exchanges, and via social channels, such as Twitter and Facebook. In this way, the marketer can achieve a sizeable increase in reach and frequency, while remaining visible with the potential customer in a much more cost-efficient way than if the marketer had relied on search alone.

Best practice: Leverage search insights to cross-promote and upsell across channels

Another typical RLSA use case involves negative retargeting on those brand terms to reduce marketing spend on navigational terms, particularly considering larger brands often have good visibility in the organic search results for any brand-related terms. This is fine if reducing cost is the primary objective, however, it could be an opportunity overlooked to simply negatively retarget those brand terms. Brands often report higher returns when bidding on their “brand” terms compared to “unbranded” terms. This is because the user that searches for and acts on the brand term is familiar with the brand, and closer to making a purchase decision than a user who types in an “unbranded” term and may still be in the earlier research phase. If any part of the 80/20 rule applies, it makes sense to try to reach those brand-aware users with marketing that could help increase business with them.

What if you could take the first-party data you’ve collated about your customer and use that to help cross-promote or upsell products or services that the customer may not have been aware of, but could be in the market for?

Take the example of a consumer who clicks on a branded search ad, visits the website, and purchases an item. Based on this sequence of actions, we could surmise that the customer is familiar and comfortable with the brand, and this could be a perfect opportunity to upsell or cross-promote. Rather than simply negatively retargeting “branded” terms for this user, the marketer could modify their creative strategy to cross-promote a related product or service. Or if the marketer found that the customer hadn’t yet registered an account, it could offer an incentive to return to the site and create a login to increase the likelihood of future engagement. By testing different approaches, marketers could actually turn cost minimization opportunities into revenue generating opportunities.

Best Practice: Sequentially message across channels to move customers along the buying journey

While sequential messaging can already be attained by leveraging different targeting windows, including search insight data into sequential messaging retargeting campaigns enables marketers to more methodically tailor messaging strategy, rather than simply relying on assumptions around time lapsed, phased periods, or impression sequencing.

For example, search is often an entry point for potential customers researching a product or a category. An advertiser could identify visits from this type of user by classifying a set of keywords as “research stage” keywords. Subsequent ads targeted towards this user might be focused on promoting educational materials or research tools, or strengthening the brand image.

The user may then go back to Google with an alternative query in mind, at which stage the advertiser could have categorized another set of keywords to identify a potential visitor as “buying stage.” Then, ads across Google, Facebook, Twitter, and the Web could be focused on moving the potential customer further along the buying cycle, perhaps by featuring product images, “best in category” claims, or promotional language to drive the user back to the site and complete the purchase.

By gathering and mixing insights from each channel, marketers can access a more complete view of who their potential customers are, and how they’re interacting across all of their marketing campaigns, in order to drive increased returns.

CONCLUSION

At this point, we’ve gone through a mass of different topics from trends, to benchmarks, to best practices. The best practices outlined in the previous section can be effective tests to help marketers make their retargeting campaigns more sophisticated. However, it’s important to note that many marketers are still value testing retargeting as a strategy, and simply getting the first few audiences set-up and retargeting campaigns running across search, social, and display can provide considerable returns.

At Digital Fuel, we can help you with your retargeting campaigns to ensure that they achieve your goals cost-efficiently. We are well equipped to retarget across multiple advertising platforms and networks, plus multiple devices such as mobile, tablet and desktop computer. We can retarget to your existing customers or an offline acquired prospect data pool, such as surveys, data collection etc.

Please CLICK HERE TO GET IN TOUCH

Author: Digital Fuel

Words by the Digital Fuel team members