This week, we reached our goal 10 positive reviews for “LUPO: The Space Adventure” in Amazon UK and started experimenting with driving traffic to the listing page. We also received the “Co-Created with City of Espoo Schools” certificate for the project we ran with several schools and the technology training company Mehackit last spring.
Apart from these milestones, it was another learning experience week for us.
With the 10 positive reviews, we started experimenting with ways to drive traffic to the product listing page in Amazon UK. Amazon suggests running a campaign with automatic targeting for the first two weeks to let Amazon’s AI determine search keywords and sponsored products placements. We did and we did learn something, though not much about keywords. Clearly more effective was the placement of our product to suggested similar products list, when Amazon customers are browsing some other product. Unfortunately, Amazon does not reveal the identity of the other products. Amazon “search term report” looks like this (partly shown, color highlights mine):
Notice that we didn’t spend much money on this so the total spend is very low on every line, but we did get some insights. Firstly, the only keyword showing up in the report is “card games”, a very broad term and an unsurprisingly poor CTR or click-through-rate. It did generate some sales, but the poor CTR resulted in a high ACoS or Advertising Cost of Sale of 63.7%, which is not sustainable, so we are not likely going to use this keyword in future advertising.
The second thing to notice is that most sales came from placements. The garbled text in the Customer Search Term column on most lines shows that this is an actual product rather than generic search term and Amazon intentionally hides the identity of the product from us. This is understandable, because both we and the other product are paying customers of Amazon. It does leave us with a challenge though. To do effective advertising on Amazon, we need to guess those matching products and find out narrower search terms, so that we can spend our very limited advertising money more wisely. We did another couple of experiments for these purposes.
To find out which other products resonate with our potential customers, we simply guessed a few products and did test searches on Amazon. We also searched our product on Amazon and looked at what Amazon would propose as similar products, sponsored or not. We now have a short list of guessed similar products and are now testing the top candidates over the next few days.
For search terms, we used Google AdWords. We simply tested various words and phrases, looking for highest CTR and reasonable CPC or cost-per-click. Below a sampling of findings.
Now, we are testing these on Amazon, again with a small budget and monitoring performance daily, adjusting accordingly.
We also did some social media experiments as well as ad tracking experiments, but more on those later. I will report on our learnings from time to time, when I feel that we have learned something that might be of general interest. Note, that much of what I reported above could have happened in a corporation, just with bigger budgets. To remind you that we are still an early stage startup, I’ll just note that in addition to reaching milestones and doing advertising experiments, we also did plenty of hustling and fire-fighting related to things like loose shipping packages used by Amazon, back-end software issues impacting our listings, etc. In a startup, you need to be a jack of all trades.
Next, looking forward to more event marketing experiences, this time at the German Comic Con in Berlin over the weekend. Hope Gerhard & Teppo will get a chance to see some of you there!
Author: Nouri Mikko Werdi