Amazon PPC Campaign Set Up – Keywords First Approach

[et_bloom_inline optin_id=optin_3]

This is part of a series where I bring in my partner in crime Dr Ellis Whitehead whom heads up the team at DATAbrill. Ellis is a Data Scientist and Algorithm expert. We get into the details to give you an inside look at what we do and how you can apply the same process on your business.

A little bit of a background

  • Dr. Ellis usually works in the background at DATABrill, but today we would like to help you with running your campaign.
  • Dr. Ellis is from the United States but ended up getting married and settling down in Germany.
  • He started working in PPC and found his way into Amazon.

Processes that go into Campaign Creation

  • The real core of PPC are the keywords leading to sales.
  • Think of it in terms of keywords, second SKUs.
  • What you need to look at when setting up is: what kind of keywords ae relevant to the account:
    • High competition keywords
    • Long tail 
    • Middle tail
    • Main keywords
  • Some accounts get sales from any of the above categories – it depends on the account/ product.
  • We will be mostly interested in finding the main keyword so that the account will get the top of the search.
    • Use only one keyword per campaign so that you get a refined report and not all mixed up. Pinpoint the right keywords.
  • Sources of sales
    • Long tail keywords – if this is the main source of the revenue then a different approach needs to be taken.
      • Long tails do not have high search volume.
      • Low impression count
      • You need a lot of months to gather data on long tails.
      • Ex: you get 12 months of data — enough data BUT the conditions in your category have changed.
    • Focus on broad matches instead — analysis will be different.
      • Multiple broad matches will overlap depending on the query phrases they are matching.
      • You need to split data on these to make proper analysis. Use statistics to get the most probabilities.
      • It is quite complicated and high level.
    • Discovery process — go aggressively after broad keywords.
      • BUT: Amazon will give you traffic which will burn through your money so you have to aggressively negate this as well.
      • You need to be negating this as well so that you will not burn through money on ACost.
      • Update your negative keywords everyday.

Aggressively Negating

    • You will bur through cash at the very beginning but you will get leaner keywords as you go along
    • The key is to update the negation everyday. Until you get to the right spot.
    • The ad will then be running longer according to budget.
    • Some words you have previously negated will then b removed from negation according to performance.
    • This will help the campaign avoid over spending.
    • You can then start increasing the budget — when and if appropriate.

Close Matches, Loose Matches..

  • Example: Tea kettles in a campaign (example)
      • Get the placement statistics for top of search and other product placement.
      • When bid is set, we adjust according to conversion rate.
      • This is limited with how Amazon is set up.
      • Based on conversion rate and how much we are willing to spend based on conversion, we move the bid up or down.

Tips from DATABrill’s Learnings

  • There is a lot of advise out there based on random effects. If you want to do things systematically– DO NOT FOLLOW THOSE.
  • You have to make sure you can systematize — NOT RANDOM THINGS
    • Ex: How many SKUs should you have in an ad group.
      • Sometimes it works better to have many SKUs in one ad group.
      • You don’t have control and you don’t have the refined data — it is all mized up.
      • Solution: Have many SKUs under one ad group and follow 1 ASIN under one ad group so yu get better data.
  • It doesn’t matter how many campaigns you have — the important thing is one ASIN for one ad group.
  • Ex: Kettles – one campaign, but separate ad groups for each variants.
  • Advantage of more campaigns: fine grain data you can use for analysis.
  • Separate auto campaigns from non-auto campaigns.
  • Product targeting campaigns: separate those.

Bidding on Campaigns Overview

  • What is the conversion rate expecting and what is you budget?
    • These are the top items to consider.
    • Your computation would be conversion rate x budget (+ or – bidding on placement costs)

Connect with SellerSessions and DATABrill: