Over the past few months, Google has quietly been testing the performance of its various ad formats on the Shopping platform for multichannel inventory, which are those searches in which a user is eligible for both Local Inventory Ads (LIA) and traditional Product Listing Ads (PLA).
On November 10th, based on the test results, it moved to sunset the multichannel ad on desktop in favor of LIAs, and replaced the traditional PLAs which were previously occupying tablet multichannel inventory with LIAs.
With these changes, the user experience for multichannel inventory is now the same across all devices.
Uniformity Sounds Great, Right?
Not necessarily. For omnichannel retailers, the multichannel ad format was an effectively designed ad unit, functioning as an LIA when the user wanted to convert in-store and as a PLA when the user wanted to convert online (see below).
Clicks on the ‘In store’ tag directed users to the Google hosted storefront page, while clicks on the standard product link directed users to the product display page on the retailer’s website.
So, why ditch the multichannel ad?
In the days following the removal of multichannel ads, there were questions as to why Google decided to make this change. While the benefit of multichannel ads was in their flexibility, the rationale Google gave for phasing them out was based on increasing simplicity and reducing confusion for marketers.
Google cited two main reasons for the ad type sunset:
- The change simplifies user interaction by aligning the experience of shopping across all device types.
2. Reporting becomes easier for advertisers as the number of possible click-types declines from four to two.
In short, Google prioritized in-store conversions over online conversions for multichannel inventory.
Google said “experiments showed that making this change had minimal impact to merchants and users.” However, one large multichannel retailer found that during the 9-day testing period from November 2 to November 10, LIA ads converted online 25% less often than multichannel ads.
Despite the shift in traffic to LIAs, there was no lift in the rate at which desktop and tablet clicks drove store visits. With multichannel ads accounting for 1/3 of all Desktop clicks, the effects of this change were noticed on the bottom line.
Breaking out LIAs into a Separate Campaign Gives Marketers More Control – at a Cost
There are two courses of action omnichannel marketers can take in response to these changes with unique costs and benefits associated with each. A single campaign structure with LIAs enabled results in simpler management, but gives marketers less control of CPCs and user experience.
A dual campaign structure (with an LIA-only campaign and a PLA-only campaign) gives marketers more control of CPCs and user experience but requires greater management effort. Additionally, some Merkle advertisers that have switched from single to dual campaign management have seen higher CPCs for the new LIA campaign than what was observed for multichannel ads when the campaigns were combined. This increase was not the result of higher bids for the LIA campaign.
The key difference between structures is the serving criteria used by Google to determine which ad format to serve. In a single campaign structure, Google looks at factors such as user location, device, query intent, and time of day. In a dual campaign structure, Google has verified to Merkle that it tends to prioritize the bid over other factors. Therefore, in a dual-campaign structure, it’s up to marketers to ensure that the bids for each ad format are optimized by location, device, time of day, and query intent.
The key question becomes where and when does it make sense to sacrifice online conversions to drive users in-store by serving LIAs? Well, it depends on factors such as user location, device, time of day, and query intent.
The retailer described above elected to separate the LIA ad unit from PLAs by setting up separate LIA-only campaigns. This allowed for independent bids for each ad unit based on the omnichannel value driven by each.
The retailer also set geographic bid modifiers based on the rates at which each ad format drives store visits and online conversions within store radii buckets. With these store radii modifiers, the retailer chooses to show LIAs instead of PLAs only where the incremental in-store impact from doing so warrants a higher bid.
While there are benefits to having more control over each ad format, again, some advertisers do find higher LIA CPCs with this strategy. Since LIAs often require a higher bid than PLAs to be prioritized in the auction, we believe Google no longer applies the same discount for LIAs based on location, device, and query intent.
Ultimately, omnichannel marketers should decide, based on their own goals and performance, whether the benefits of increased control outweigh the costs of managing more ad units, as well as potentially increased LIA CPCs.
As frustrating as this change is for omnichannel marketers, it presents a new opportunity to segment campaigns by ad format and maximize the value of each ad format for the channels in which they are designed. Customizing bid modifiers by location and device based on omnichannel performance for each ad type is crucial for marketers opting into a dual campaign structure.
From an industry perspective, this is another clear step from Google in prioritizing the mobile experience for both users and advertisers. As mobile share of query volume continues to grow, we expect Google to make more decisions that emphasize continuity across devices based on the mobile experience.