2017 Amazon PPC Roundup with Brian Johnson and Danny McMillan (Part1) – Session 021

[et_bloom_inline optin_id=optin_3]

In this session Brian Johnson brings us the latest 2017 Amazon PPC roundup on the updates from Amazon and what it means to your keyword optimisation and amazon ads.

Brian has been in ecommerce for 11-12 years, was a student of AMS2 (Amazing Selling Machine 2) and with his engineering background he delves into testing everything.

Feb and May Amazon updates

  • We have seasonality that affects buyer behaviour
  • Amazon started to rollout in Feb 2017, the change to the organic algorithm and specific to the indexing changes
  • Took about 3 months to rollout this change out in total

Amazon PPC roundup – Indexing and algorithms

  • Indexing is an organic algorithm function, not a sponsored function
  • Sponsored has it’s own queues to test for relevance
  • The A9 algorithm is the organic indexing
  • This indexing tests the relevance of the product for what is being searched for – it’s a pass or fail in that you’re indexed or not for this search phrase
  • Search position ranking is different – where you might be on page 1 for a search term versus beyond page 20 as an example

Change in Number of Keywords Indexing

  • Back in Feb-Apr the change we saw was a difference in the number of keywords that we were indexing for
  • If we’re not showing up in organic search then you’re not there to make sales
  • There were far fewer keyword phrases that products were showing up for
  • It was confusing when we looked at our test results – was it category specific, was it number of chars, was it the make up of the chars (punctuation, commas, etc)
  • Was it including or excluding description fields

Amazon PPC roundup had inconsistent Testing Results

  • The testing was inconsistent across accounts making it difficult for us to identify the algorithm changes
  • This made it difficult for to provide advice – we gave general rules to stay within but people had to test it for their individual products
  • Rule: Try to keep under 250 chars for all backend search term fields – don’t rely on the backend fields but use to supplement your indexing from title and bullets
  • Organic algorithm uses Title, bullets and backend search term fields
  • The sponsored relevance match has to do with the title, bullets and front-end description

Backend Search Terms

  • Looking at a service that enables you to view the “Backend search terms” of your competitors – you can Google for this service – consider this a grey hat service and Amazon will close it
  • Identified that big sellers (in this product) were using frames in a top-down approach. Broadest two-word search phrase at the top and then nest them underneath – repeating their main keywords down
  • Brands that would rank the same product in every country
  • But this approach seems to throw out the original thought processes
  • (don’t use commas, don’t repeat search terms, etc) for what good practice should be

Amazon PPC Roundup Best Practice Guidance

  • Amazon is always telling us to do one thing, then another person tells us another. Which source do you trust is upto you, but views will differ.
  • Focus on full phrases separated by a space or the individual search term fields
  • Brian has run a thousand or so tests for academic research
  • This gives large scale patterns and a large data set to understand how the algorithm works
  • Alexa woke up here, listening to keywords so Brian had to unplug her 😀
  • Back in April / May Brian was viewing the lack of consistency here across listings with a mix of commas, repeating words, exclamation points, semi-colons and nobody knew what they were supposed to put in the backend
  • This was not very reliable

Bottom Line is…

  • Carrying that forward – we don’t have an answer on this yet
  • Amazon rolled out the 5,000 to 400 chars which was officially announced last month
  • Some 3 months after they did it (pretty normal for Amazon)
  • More people are teaching and implementing that the backend search term fields are being updated to be more succinct
  • It would now be a more valid experiment to do a sample of a 1,000 accounts backends again to look for the patterns and differences been top and poor sellers
  • Would expect – Top sellers will rely on their title for their most targeted keyword phrase indexing
  • The bullet points come net
  • If anything else is not being indexed or the seller doesn’t want in the front-end they can put it in their backend
  • The sellers will have gone back and updated them based on the 2017 knowledge and best practices
  • This means that a test now would give better patterns recognised since Amazon’s update.