When is burying products useful?
We strongly believe in behavioral driven ranking in Voyado Product Discovery, where the ranking proven to yield the best sales figures for a given context is the one to present. However, for some scenarios the product discovery engine does not have any information it can utilize. Burying products allows you to provide that kind of additional insight to the engine.
Key use cases:
- Bury suppliers' products
- Bury a product type
- Bury seasonal wear
How does burying products affect my site?
Within most algorithms, the sales figures of a product are taken into consideration when determining its rank.
The functions on your site that are powered by Product Discovery and support the bury functionality will, during the lifespan of being buried, temporarily decrease the sales figures of every buried product. Buried products will lose positions in search results, category pages and other locations where burying is supported. Original sales figures for the product, phrase matching in search and different trends will however still affect the ranking to some extent.
Which features support burying products?
Burying products is fully utilized in:
- Product lists, normally used on category pages, when using relevance or sales as sort order
- Search hits, when using relevance or sales as sort order
- Recommend based on product
Burying products is partially utilized in:
- Product lists, normally used on category pages, when using personal as sort order
- Autocomplete
- Did-you-mean
- Product suggestions
Functions that partially utilize burying products, use product bury on product backfill.
What happens when burying a product twice?
Products will be buried in the same way as when being buried once.
Why can't I bury even more products?
There is a limit of 100 simultaneously active sets of buried products in each market. If this limit is reached, check if any other bury for that time period could be removed or re-scheduled.
We do not recommend having too many active buries at the same time. Too many buries will likely give you unintended outcomes and require a lot of manual labor.
What do the statistics for buried products show? (version 3)
The top row of statistics shows the aggregated number of interactions on any included buried products. For example, if products P1, P2, and P3 were buried, the number of displays would increase by 3 on a page load where all three products are displayed. If the page was loaded once more in the session, the number of displays would be increased by three yet again. The same principle applies to clicks and sold units.
If impact measurement is enabled for the bury, the second row will show the impact for buried products. This means that it will show how much less interaction for each metric the bury has contributed to. Normally these figures are negative, they can however be positive in cases where not much data has been gathered. For example -67% in displays means that the displays of the products being buried have decreased by 67% in visibility.
What do the statistics for buried products show? (version 4)
If impact tracking is enabled for the bury, the BURIED column shows the statistics for buried products. The CONTROL GROUP column shows the statistics for normal (non-buried) products. The ESTIMATED EFFECT column shows the impact of the bury using the difference between the first two columns.
For example, an Estimated Effect of 67% in displays means that the displays of the buried products are reduced by 67% because of your bury.
Normally the figures in the ESTIMATED EFFECT column are positive, but they can be negative in cases where not enough data has been gathered.
How is the Estimated Effect calculated? (version 4)
The Estimated Effect measures the relative increase or decrease in session activity, sold units, revenue, and profit caused by the bury. It compares numbers for sessions when the bury was active with numbers for sessions in the control group, with no bury active. The effect is estimating the relative change it has caused.
For Sold Units, Revenue, and Profit, we need at least 500 sold units in the control group to calculate the Estimated Effect. This is to guarantee statistical reliability, and when this condition is not met, those rows are highlighted in yellow.
The formula used to calculate Estimated Effect
Estimated Effect = (Activity in total number of sessions / Total number of sessions) ÷ (Activity in control group sessions / Number of control group sessions) − 1
This formula expresses the percentage change in activity between sessions exposed to the bury and those in the control group.
Example
Let's say you have the following data:
- Control group: 5,000 sessions generated 10,000 clicks
- Buried group: 20,000 sessions generated 5,000 clicks
Then:
- Average clicks per control group session: 10,000 ÷ 5,000 = 2
- Average clicks per session: (10,000 + 5,000) ÷ (5,000 + 20,000) = 0.6
- Estimated Effect: 0.6