Trends - BloomReach Experience - Open Source CMS

This article covers a Hippo CMS version 10. There's an updated version available that covers our most recent release.



This feature is only available in Hippo's Enterprise Edition.



Find interesting trends in the stream of visitors to your sites.


The Trends tab in the Audiences perspective consists of two columns. The left columns contains two filters: time period and visitor segment. The right column displays the data in the form of a graph and a "peak performing" list.


Time Period

This filter allows you to select the time period for which to display data. You can select either the past 7 days (1w), the past 30 days (1m) or the past 90 days (3m).

Visitor Segment

This filter allows you to narrow down a visitor segment for which to display data. Data for the selected segment will be displayed in addition to the data for all visits, so you can compare them.

Narrowing down a visitor segment is done by selecting certain facets of the visitors such as "looks at category", "comes from country" and "comes from city". Multiple values can be selected but this does not make sense for all facets. For example a visitor can look at multiple categories, but can only come from one country. Selecting multiple countries simultaneously will shrink the selection to the empty set.



The graph plots the number of visits against the selected time period, as well as the general trend (upwards, downwards).

If no visitor segment has been selected the graph shows all visits:

If a visitor segments has been selected the graph shows the visits for the selected segment in addition to the all visits so you can compare them:

To hide the All visits line altogether you can click on its legend item.

Hovering the mouse pointer over a data point results in a pop-up showing detailed numbers for that data point:

Peak Performing

The Peak Performing list displays items that stand out in some way for the selected time period and visitor segment.

Above the list you can select one of a number of facets, corresponding with the available facets in the visitor segment filter. For example "peak performing category" or "peak performing country".

If no visitor segment is selected, the list will show the number of visits and the percentage of all visits for each item:

If a visitor segment is selected, the list will additionally show the percentage of the selected segment for each item, and a score of 0 to 5 stars:

The score reflects how much the item stands out: it is high when the item represents a large number of visits, or if the item occurs significantly more often in  the selected visitor segment compared to all visitors. See Trends Scores for more details on the way the scores are computed.


In the screenshot below we have selected visitors from the US who have looked at content labeled 'solar' in the last month:

Notice the following:

  • The "comes from country" list only shows the US as visits from the US are not simultaneously also from other origins.
  • The "comes from city" list has adjusted to only show cities that occur in the selection, that is, U.S. cities.
  • The graph shows how visits in the selected visitor segment compare to all visits.
  • In "looks at category", 'solar' is selected but other categories are also shown because visitors that look at 'solar' also look at other categories.
  • The Peak Performing list shows document categories that stand out. For example, 'energy' is looked at in 34% of the selected visits compared to only 7% of all visits. This earns it the maximum number of 5 stars.
  • Category 'food' is looked at in 36% of the selected visits compared to 53% in general. This difference also earns it 5 stars.
  • Category 'solar' itself also occurs in the list, at a spectacular 100% but that's just because we're filtering on 'solar'.


Configuring additional facets

The list of facets available in the Visitor segment...  and Peak Performing panes can be customized per project.




Did you find this page helpful?
How could this documentation serve you better?
On this page
    Did you find this page helpful?
    How could this documentation serve you better?