The digital commerce landscape is not a level playing field. Everything is skewed in the retailer’s favor. Requirements, margins, stock availability, product visibility – all dictated by the Amazons, Walmarts and Targets of the world. Even the most basic elements are an uphill battle for most sellers, filled with unexpected twists at every turn.
Once sellers get a steady footing as it relates to the fundamentals (i.e. making their products available), they then need to worry about turning a profit. And as it relates to what’s truly profitable, ambiguity rules the scene. Nowhere is this more apparent than in the attribution processes these retailers have deployed within their advertising platforms.
Take Amazon and Walmart, the two heaviest hitters. There are a lot of front-end similarities that exist across their sponsored search advertising platforms. They both:
- Offer self-service, sponsored search functionality for sellers looking to advertise on-site
- Auction-based, CPC functionality similar to Google, Bing, etc.
- Use in-house tech to power and support their sponsored search platforms
- As opposed to using outsourced, rebranded technology
- Provide advertisers the option of creating campaigns via manual or automated methods
- Advertisers prefer manual, but deploying automated has its advantages as well
But once the conversation shifts toward backend functionality—how and when ads are displayed, how revenue is attributed, etc.—things veer in a very different direction. They both have drastically different ways of approaching backend metrics that tend to skew their results, making it difficult to compare performance across platforms.
To accurately compare both Amazon and Walmart’s sponsored search platforms, we’ll be reviewing a few approaches respective to each. In this series, we’ll take a look at:
- Auction and ranking dynamics
- Offline revenue
- ROAS and profitability
Auction and Ranking Dynamics
Auction and ranking dynamics set the stage for the divergent path these two advertisers take. In this instance, it’s the upfront serving functionality and placement options inherent to each ecosystem creating the divide.
Even at the fundamental level of auction type, these two platforms differ. Amazon takes the standard approach of a second price auction (2nd place bid + $0.01), whereas Walmart uses a first price auction (highest bid, as is) approach typically seen within display environments. From there, it only intensifies.
Amazon’s approach to ad ranking is designed to drive revenue, plain and simple. It lets advertisers occupy almost any spot within the top row of the in-grid search listings. There are also multiple, further down ad slots that can be occupied by any number of advertisers to go along with their sponsored display and digital display placements. See example below for the nonbrand query diapers newborn:
This creates an ideal scenario for larger advertisers looking to own the page. It ensures no ads (beyond their own) can sneak into the Amazon search results grid and steal revenue at the very bottom of the funnel, protecting their prior upstream marketing investments. Some advertisers are even willing to accept a loss on their brand term-specific investment if it means defending against conquesting.
For most advertisers, though, this format presents a bit of a catch-22. They know they need to defend their brand terms when consumers search for them, but:
- Are there always competitive advertisers looking to steal share?
- Is every impression up for grabs as Amazon would lead them to believe?
- If every impression isn’t conquested, are advertisers bidding against themselves?
Bidding on brand-focused auctions that would have shown their product naturally via an organic listing is a great example. Should these advertisers focus on defending their turf if, for a vast majority of the time, there is nothing to defend against? See the example below for the brand query huggies:
This is where the idea of incremental revenue enters the conversation. It comes up in relation to Amazon a lot. Advertisers want to know:
a) What revenue is incremental, i.e. what did we need to bid on?
b) What revenue would we have gotten organically, i.e. what did we not need to bid on?
When an advertiser owns their brand terms organically, they shouldn’t have to bid on them. But then again, there are instances beyond conquesting when it’s needed. For instance, advertisers could advertise on established/organically owned products to:
- Bring attention and potential sales to newer products
- Attempt to drive additional sales for overstock
- Push sales for less popular products
The thinking being, if consumers came in searching for a specific, popular product, then that product would a.) show up organically alongside the sponsored ad and b.) be bought anyways. The goal would be for the consumer to buy the intended, popular product and the advertised, less popular product.
As noted, there are ways to use brand term bidding to the advantage of advertisers. Knowing when best to do this is the difficult part. Advertisers should own their brand terms organically, but how do they know when they need to defend against conquesting? And how can one tell when those situations arise if Amazon is designed to show an ad in every potential auction?
It’s gaps like these that make Amazon a difficult sell for some advertisers. It’s a damned if you do, damned if you don’t environment, wherein a lot of the advertising spend is compelled by revenue-related FOMO. The idea of, “If you’re not there, your competition will happily take your place.”
But it doesn’t have to be that way.
Unlike Amazon, Walmart puts a premium on its organic listings. You typically won’t see a search results page dominated by paid ads. In most instances, consumers will barely know ads are present due to the restrictions within their auction dynamics.
Within Walmart’s search environment, a product must rank within the top 128 products to qualify for advertising. This means random advertisers cannot show up within the sponsored ads for a term they bear no resemblance to. It also means the ads that do appear will be relevant.
On the other hand, Walmart still doesn’t have the capacity to include negative keywords, which limits control on the advertiser’s end.
As of this writing, Walmart only allows advertisements to occupy specific slots within their in-grid listings. Positions 3, 5, 7, 8 and 12 are the only slots that can contain an ad. There have been rumblings of Walmart switching this approach and adopting a more Amazon-esque layout wherein search ads can appear within slots 1-12. Currently, most sponsored ads we’ve seen are still limited to specific slots. This approach corrals a few things that are typically rampant within paid search advertising:
1) Ensures organic listings own the top ad positions and help guide the results
2) Eliminates the need to bid heavily on brand terms out of fear of conquesting
3) Guarantees an organic-driven ecosystem due to ads only occupying non-premium positions
There can be a maximum of two in-grid ads per page within the search results for any given query. Even though ads can appear in one of five slots as noted, they will only occupy two of those slots at any given time. Again, this is barring any changes to Walmart’s overall approach to ad slotting.
Walmart’s ad slotting approach helps to alleviate this “incrementality” issue. The majority of clicks are driven by the first two results. Ads can only appear in position three at best, meaning both brand and nonbrand can drive incrementality.
Walmart’s self-imposed handicaps are another reason the ROAS-based metrics across Amazon and Walmart are not apples to apples. These factors help alleviate concerns around unnecessary competition, the need to defend conquesting, and incrementality, but they also weigh on Walmart’s ability to drive revenue and volume via sponsored ads. If they went the route of overwhelming their results pages with ads, which they may soon do, their volume and ROAS outputs would surely benefit.
To sum it all up, when it comes to auction and ranking dynamics, Amazon leaves the barn door wide open, whereas Walmart closes it a bit. This allows Amazon to drive impressive back-end metrics, but at what cost to true profitability?
In the next installment of this series, we’ll review the attribution processes of both Amazon and Walmart and shed some light on how each accounts for revenue and how these varying approaches alter outcomes within each platform.