One of the powerful features of the Onefeed system is the optimiser. It is also one of the more complicated features of the system and many clients find it difficult to understand how it works. Ultimately the purpose of this tool is to help keep costs down and stop items getting significant volumes of clicks without orders. In the past merchants have spent a lot of money on items that they weren’t aware of, which resulted in hefty cost which could have been easily avoided. I will attempt to describe more about this now to try and clarify how it all works and how you can use it to your advantage.
Target Setting
The first part of the setup process for the optimiser is to set the targets at a category or product level or brand if you prefer to work by brands. All our merchants set the targets at a category level, however any targets set at the product level will take precedent.
Above is an example of the targets that have been set, each category can have a different value, they don’t all need to be the same values.
Rule Setting
Once the system knows what targets you are looking at achieving we can then set the way the optimiser interprets the click and sale data to the way it changes your feed and bids. There are 3 parts to this screen and they are explained here:
General
Target to use = Category / Brand – you can either use the category or brand targets. If you select Brand then the above table will change category to brand where you can set your targets.
Default Target = If a new brand or category gets added to your feed then it will assume this value
Target Metric = Cost of sale / Cost per acquisition – this tells the optimiser if you are looking to optimise based on cost of sale of each product or how much it costs to generate each sale
Date Range = This is the most important part of the optimiser to understand. The date range set here is a rolling date range, so each time your feed runs it will look at the last x number of days that you set here. For merchants with few clicks it doesn’t make sense to set a date range of 7 days as it will never remove anything, however with a long time period it will have more data to make optimisation decisions on.
Sales Activity
This sets the benchmark of when an item gets optimised. In the example the optimiser kicks in when an item reaches 1.2 times your target, so with a target of 4% the optimiser will kick in when the cost of sale for a product goes above 4.8% cost of sale, anything lower than this value it doesn’t do anything.
Products with no sales
In contrast to the above rule the optimiser can be setup to run on items that aren’t making any orders and don’t have a cost of sale. We find more products will be eligible to be optimised using this part of the optimiser.
The first part of the optimiser settings is the hardest to describe. This essentially will only kick in once the click costs exceed the metric multiplied by the target cost of sale. So using an example, if an item is £100 in value and the target for cost of sale is 4% then the optimiser will kick in once the costs meet £3.20 (which is 0.8 x 4%).
The other 2 metrics are more self explanatory.
Some other hidden features of the optimiser are that the products will automatically get added back into the feed if they start to sell. This is great for items that sell once they have been removed from the feed. It ensures items always remain within the cost of sale targets.
One other feature that needs to be explained is once an item is removed from the feed and doesn’t sell we don’t automatically add this product back into the feed. Most of the time if a product is not selling we don’t want them going back into the feeds when they may still not sell.
The optimiser has a few different effects when it is optimising which depends on the feed / partner the feed is for. Here’s a description of what happens for each partner:
Google PLAs – when an item needs to be optimised the bids get lowered to 1p. This means items can still be found but their rankings may mean they are very far down any product lists.
Shopzilla – As above, but instead of setting the bid to 1p we set the bid to 0, which has a similar effect to Google, but it means the items still stay on site.
Everything else – For all other sites the products are simply removed from the feed as changing bids has little or no effect.
