On this page
You can create a Rule Engine strategy to identify top performers or bottom performers in your account.
Once you've created a custom strategy, you can start adding conditions to help you find the top and bottom performers. You can also add these conditions when customizing existing and prebuilt strategies.
To build the conditions, you can follow these steps:
Select the metric you want to use to identify your top performers, like Clicks, impressions, etc
Select the operator “in top” or “in bottom”
Add the number of entities you want to see, i.e., 15 percent (to see 15% entities) or 5 value (to see 5 entities)
You will be able to create a condition like this:
This will help you get the results based on the performance data of your account, the system will compare the performance of keywords and show you the top 10% with the highest cost.
Logic of these conditions
The system will look at the performance of the entities (the scope at which you create a condition, like keywords, campaigns, etc.) and will rank them to get the top or bottom performers.
The top 10% or bottom 10% entities are identified from all the entities that enter the strategy’s scope. The ranking is given based on the metric you choose, like clicks, cost, conversions, etc.
Other conditions to filter out entities based on some thresholds, like Cost > 5 or Impressions > 100 do not disturb the ranking. For example, if you have 100 keywords, the top 10% would show 10 keywords in the result.
If the strategy has any pre-filters applied, the entities entering the scope would change and so the results would change too.
When the top/bottom condition is used with other conditions like Click > 5, the result could show lesser than N or N% entities. Let's understand this with an example of a rule that has conditions
Keywords with Cost > 5 and
Keywords in top 10% of account
Assuming that the account has 100 keywords, out of which only 7 keywords have cost > 5 in the last 7 days. In this case, the system will evaluate the top 10% of 100 as 10 keywords and then filter out the ones with a cost < 5. So, in the end, it will result in only 7 keywords (and not 10).
If there is more than one entity with the same performance data, the system will check the cost to decide the higher rank, and if the cost is also the same, it will consider impressions to break the tie.
In a situation where both cost and impressions are the same, the system will sort the entities for their id (Keyword id, Campaign id) and rank them.
How will the combination of other conditions work with these?
As mentioned above, the top and bottom analysis is done on all the entities. Now, if you have other conditions that would filter out some entities, it will be on top of the ranking. For example, if you have a rule like the following:
This is how the system will evaluate this rule:
Let’s say you have 50 keywords where 30 keywords have received 0 clicks in the Last 120 days. This means that only 20 keywords have Clicks > 0.
As mentioned above, the top/bottom conditions would apply to the entire set of entities, so 10% of 50 keywords would be 5 and this is what the system will show you as a result, and not 10% of 20 keywords.
However, if the case here is that there are only 4 keywords with Clicks > 0 in the Last 120 days; then the system would only show you 4 keywords as result. This is because, even though 10% should show you 5 keywords, one of them is not satisfying the other conditions; so only 4 will be shown.
How will different rules with these conditions work?
The system will filter the matching entities in rule 1, before moving to rule 2 - just as it would do today for any strategy in Rule Engine. Let's take another example here to understand the flow:
The system will rank all the keywords based on impressions, in this example. Let’s say there are 1000 keywords in the account.
First rule - 10% keywords, would show 100 keywords in the result. If we had to understand by rank, these 100 keywords would be of rank 1 to 100 by impressions.
Second rule - 20% keywords, would be 200, which is rank 1 to 200. This is because, the ranking would be done on all the entities entering the scope. So, 20% here would mean 200 keywords from the top, that is keywords with rank 1 to 200 by impressions. However, the top 100 keywords (ranked 1 to 100) are already filtered out by Rule 1, so the output of Rule 2 would be the remaining 100 keywords ranked 101 to 200.