How to use Amazon's Ground Breaking Query Performance Data

How to use Amazon’s Groundbreaking Query Performance Data

In the past few weeks, Amazon has unleashed an incredible amount of new data on shopper behavior and performance, chipping away at the long-believed notion that Amazon performance is a black box. One that we're particularly excited about is the Query Performance Analytics. You can find this in Seller Central under Brand Analytics.

This dashboard lists hundreds of search terms relevant to your brand, followed by the full funnel of consumer behavior following it. For each search term, viewable across weekly time periods, you can see: • Impressions in search

• Clicks from that search term (we’ve confirmed this includes both paid and organic clicks) • Cart adds (a never-before-seen metric on Amazon)

• Purchases, including median price

These are hard numbers (actual count), not just illustrative trend or relative rank typical with past Amazon data. Furthermore, you can see things like ship speed, as well as the big hitter: % brand share of each metric.

Understand the value of a single search term. We've never before been able to say exactly how valuable a term is. Now we know how many purchases it drives, and the shape of the funnel from impression, to click, to cart, to purchase. Branded, category, and competitor terms have very different funnel shapes: for example, category terms tend to have much lower share compared to branded terms. This helps us focus on which terms matter for us to track rank, share of voice, or ad performance.

Determine the value of investing in capturing more search. I think this is pretty groundbreaking: you can size up the opportunity of a single term. Look at the purchase count, and then look up the #1 % conversion share on Brand Analytics for that term - in theory, that's the ceiling of return on investment if you were to be the dominant product for that search term. Multiply that % by your AOV for the term, and you have your size of prize. In my opinion, this demonstrates that most terms above 5,000-10,000 in search rank aren't really meaningful for low ASP items - in that range, you're talking about $5k in incremental sales per month in many cases.

Measure the impact of your experiments. This data suggests you should change or fix many things - namely, your content - so you can click and convert better on each individual term. Make adjustments to your content, and then track how the funnel changes over time. The fun part is you can actually attribute $ or hard conversion numbers to the changes you make to your pages and/or ad creative. We generally hear about a lack of rigor in experimentation around content (or variations! - for another post :) ) on Amazon. This can change that!

We're still digging into these dashboards to validate the data and make sure we understand the numbers in front of us. It seemed like they improved in accuracy once the rollout of the mobile app sessions happened a couple weeks ago.

As with all performance marketing, we always need to be testing, experimenting, and showing our work (and measuring impact). This is a great way to do that - and you should start asking questions about these dashboards!