Hacking Amazon: Agencies Create Workarounds For An Emerging Ad Giant

Although Amazon is doing a good job responding to advertiser needs, “we’re on our own timelines,” Sullivan said. “Were accountable to our clients right now, so we have to innovative as much as we possibly can to deliver against the expectations set against us.”

Here’s how agencies are hacking Amazon’s ad platform.

Retro tech

Many buyers liken using AMS to what it was like running Google search campaigns in 2005.

Campaign setup, deployment and optimization are manual, as are selecting SKUs, changing campaign end dates and adjusting product display ad bids.

“Think of the most vanilla version of AdWords that you possibly can – that’s the current experience through AMS,” Sullivan said. “Anything moderate to advanced needs to be done through custom tools.”

To remove some of the heavy lifting, Performics built a dashboard on top of AMS called Caiman that pulls data from Amazon’s API and automates campaigns. Caiman also offers buyers more granular reporting so they can measure basic things like keyword performance.

“We were starting to have a significant amount of volume going through the platform, [and] we couldn’t handle it without machine aid,” Sullivan said. “For a long time, you couldn’t even do [dayparting].”

Dentsu performance agency iProspect has also rolled up its sleeves to manage AMS’s shortcomings. One tool it developed allows the team to pull granular reporting data and push out broad campaign updates on the fly. Another helps the agency make smarter selections for product ads by providing insights into keyword performance.

“So much of the work that is automated within paid search is not in Amazon,” said Jeremy Cornfeldt, president of iProspect US. “It takes a lot of manual work to get a campaign launched and to manage it through the process.”

Measurement is also still a big challenge within AMS. Although Amazon recently launched the ability to view trend lines and filter results by product, AMS doesn’t allow API access into all of its ad products. Because AMS stops reporting on campaigns after 90 days, it’s impossible to do attribution or historical analyses.

And data related to consumer insights, like search volumes, is nonexistent.

“I would love to get a sense of the volume of activity on Amazon,” Cornfeldt said. “How does it break out and differ by category? How does it differ on Amazon versus Google?”

Programmatic probs

It’s also early days for AAP, Amazon’s demand-side platform.

Like with AMS, campaign setup and optimization are manual, and reporting is limited on the platform, said Oscar Garza, head of media activation at Essence, which has built APIs that oversee the QA process in an effort to automate some of the grunt work.

In contrast, Google provides buyers with intelligence to help them know if they’re under-pacing a campaign, for example, or whether they have creative approval.

“These are things traders would need their attention drawn to early in the day,” Garza said.

But because it’s so new, AAP doesn’t offer buyers as much control as other DSPs to optimize toward specific KPIs, he said. And because the platform isn’t integrated with viewability and brand-safety vendors, it’s risky to buy outside of Amazon’s owned-and-operated properties.

“We’re very cautious about the media we buy and how we buy it,” Garza said. “It’s a quite a straightforward focus on selling for that Amazon conversion event.”

Like AMS, AAP’s reporting consists of “a couple of canned reports you can pull without a lot of opportunity to filter,” Garza said. In order to do advanced measurement and attribution, buyers would need access to AAP log-file data, which Amazon doesn’t share.

Many of AAP’s problems come down to being a new DSP in a mature market, said Dillon Lockett, VP head of display at iProspect.

“AAP is more cumbersome to navigate than other DSPs, and the UI is not as fast and intuitive,” he said.

But like search buyers, programmatic buyers don’t fault AAP too much for being a newbie.

“These are the things you deal with in software development,” Garza said. “It’s got bugs, and you have to squash the bugs.”

Enter ad tech

Not every agency has the intestinal fortitude to build its own technology. As buyers start to ramp their spend on Amazon’s ad platform, some are turning to partners for help – and the ad tech industry smells opportunity.

Kenshoo released its solution for AMS in September. The platform, used by around 400 brands, taps into machine learning to help buyers scale campaigns while offering more granular reporting. Kenshoo pulls in data about product and keyword performance through Amazon’s API to optimize bids specifically for the platform.

“One of the most important tenets of driving efficiency on Amazon is better understanding of the product,” said Oren Stern, general manager at Kenshoo.

Other startups, such as Downstream, are cropping up to cope with the nuances that come along with Amazon’s dynamic marketplace.’’

On Google, it’s possible to point to a static URL all day,” said Connor Folley, co-founder of Downstream and a former Amazon exec.

“On Amazon, you have tens of thousands of products, profitability constraints, ratings and reviews, and you have third-party sellers competing,” Folley said, noting that, at least for now, most ad tech innovation is happening on AMS because AAP isn’t stable enough to build software on yet.

Hope springs eternal

Buyers keep their fingers crossed that Amazon will provide more data, automation and reporting on its platform and eventually make it easier to measure how one type of advertising on Amazon impacts another.

But they also realize that, unlike Google and Facebook, advertising is not Amazon’s core business.

“They’ve had a remarkable ascent in advertising in terms of volume,” Sullivan said. “But Amazon wasn’t built inherently around the concept of advertising. If Google was born as an ecommerce business and backed into advertising, we’d probably be saying the same thing.”

This post was syndicated from Ad Exchanger.