Amazon provides an automated ad targeting option on its Sponsored Ads platform both for Sponsored Brand and Sponsored Product ads. This is sometimes helpful, but over-reliance on automated targeting to generate an optimised keyword set can be expensive and, frankly, lazy.
There are numerous keyword generation tools available. The same is true for products/ASINs if that’s the way you want to target. And you can always start with the obvious and work out from there. Even if the only tool you have is the Amazon website search bar, you can generate good keywords. Let’s assume, for example, that we are advertising fence posts. I know little about fences, but if I type ‘fence post’ into the search bar I can see expansions on that term. Now I just have to pick the ones that are relevant.
Another 20 minutes with either a pad, pen and someone who knows the industry, or a keyword tool like Jungle Scout, and you’ll have a page full. If you launch some ads with some phrase or broad match terms, then you’ll generate more. You can view these and analyse their performance in the Customer Search Term tab. You’ll also be able to use negative targeting to remove some unhelpful terms that crop up. Very quickly you have something reasonably efficient that you can start to optimise.
Those who value targeting automation will say that we might have missed something brilliant…that people shopping for lawn seed might convert brilliantly for fence posts. They may be right, but based on Pareto’s principle, we know that ‘fence post’ will do most of the work. Anything extra we get from running automated targeting, that we wouldn’t have thought of without it, will just be the icing on the cake.
But there are people and agencies who take this principle too far. They find it easier to launch automated ads with a set of optimisation rules for harvesting the best keywords. The issue here is that these ads spend money fast, and only learn to be efficient over time. You can be £5k or £10k in and only just starting to have something approaching an efficient keyword set. The data below is from a client we won from one such agency. They relied heavily on automated keywords to create their ads and the return on ad spend was poor.
With our 100% manually targeted campaign the return on ad spend was twice as high from day one. This wasn’t a fluke, as the ROAS continued at a similar level ongoing. The ads were twice as efficient because of a bit of thought and planning.
Let’s go back to our hypothetical example. There’s no need to spend any money on the keyword ‘fence post’ to know that it should be in the keyword set. The time and money would be much better spent on optimising the bid. And no-one needs to wait for 10 clicks-worth of data/expenditure to know that ‘post box for fence’ is NOT going to generate any conversion for this product range. Putting ‘box’ as a negative phrase match up front will avoid wasted spend there.
If you are working with an agency or pitching your Amazon advertising business, ask about the agency’s approach to keywords. Healthy answers would include a clear articulation of what type of keywords are being targeted, with examples. Phrases like ‘automated ad targeting’, ‘you just have to keep spending and the algorithms will sort it out’ or ‘it’s all machine learning’ should cause some alarm bells to ring.