Where does the eco-mode come from?
The eco-mode has been developed in line with Piano’s strategy to provide customers with the most advanced, qualitative, and ethical analytics solution, especially when it comes to sustainability.
Continue providing unsampled data as fast as possible while reducing the computing resources needed to provide the data, therefore minimizing the environmental impact on the usage of this data; this is what the Eco Mode is about.
How does the eco-mode work?
The eco-mode relies on a statistical function called HyperLogLog to estimate the approximate number of distinct values in a data set.
HyperLogLog is a state-of-the-art cardinality estimation algorithm capable of estimating distinct cardinalities of trillions of rows with an average relative error of a few percent.
You can read more about HyperLogLog and its many virtues on Wikipedia.
What are the benefits of the eco-mode?
When it's enabled, and thanks to the HyperLogLog function described above (How does the eco-mode work?), the eco-mode drastically reduces the computing resources needed to provide the data set requested by the user of Piano Analytics. Based on internal studies that are regularly monitored and updated, we estimate that, for the exact same data set requested, the eco-mode reduces computing resources by an average of 20%, and this can go up to 60%.
-
Environment
As the eco-mode allows reducing computing resources needed to provide the same data set, it, therefore, allows you to decrease the carbon footprint generated while using Piano Analytics. -
Speed
As the eco-mode approximates the number of distinct values in a data set, it, therefore, allows you to reduce the time needed to get your data, especially when it comes to a large volume of data requested or a large period of analysis (year-to-year comparison, for instance). -
Data Quality
As the eco-mode relies on a state-of-the-art cardinality estimation algorithm, there is, therefore, no compromise on the accuracy of the data, as the average relative error of our HyperLogLog implementation is estimated under 2%.
How and when to use the eco-mode?
The "eco mode" is extremely effective when it comes to counting the number of distinct values in a data set. It is therefore particularly suitable for processing metrics that rely on counting values such as
- the "Visitors" metric (number of deduplicated Visitor IDs)
- the "Visits" metric (number of deduplicated Visit IDs)
- the "Articles" metric (number of deduplicated Article IDs)
- And all the metrics related to them: bounce rate, time spent (visit), etc.
Metrics that don't need deduplication of elements are not approximated and show the exact same values with or without eco-mode. Here are some examples:
- Conversions
- Clicks
- Turnover with or without tax
Remember!
When the eco-mode is enabled, only eligible metrics are calculated using this methodology.Other metrics will be calculated without approximation, as they would be when the mode is disabled.
With its significant impact on performance, the eco-mode is recommended for the entire exploratory phase on large volumes of data (high-volume sites, extended analysis periods, significant cardinality, etc.). It will save you precious time in identifying insights while significantly reducing energy consumption. You will be able to deactivate the mode to launch the query which will allow you to obtain accurate results.