Time Sensitivity

The Time Sensitivity classification is a way to categorize content within Echobox. Certain features in Echobox will behave differently depending on the classification.

The three available classifications are Breaking, Normal and Evergreen. When an article is first seen by Echobox (generally from an RSS feed), Echobox will try to automatically classify it by analyzing the content.

Giving your article a classification manually allows Echobox to learn about your content and will improve classifications going forward.

Labeling your content allows Echobox to appropriately schedule your post to reach your audience at the time. When using Optimal timing, any share can be labeled as:

Breaking: Post expected within 15 minutes of scheduling
Normal: Post expected within 12 hours of scheduling
Evergreen: Post expected within 48 hours of scheduling

Time Sensitivity is per article not per post

An article can only have one Time Sensitivity classification. If the Time Sensitivity is changed, it will change it for all future posts of that article, even across different Social Pages. The change will not affect the suggestions which have already been scheduled. If the article is in a scheduled post, that post's time will not change.

Breaking

Articles that are classified as Breaking will be posted by Echobox within 15 minutes when scheduled with the Optimal timing option.

The Autofeed will never reshare an article that has been classified as Breaking.

Normal

Articles that are classified as Normal will be posted by Echobox within 12 hours, when scheduled with the Optimal timing option.

Whilst 12 hours is recommended as the default setting, it is possible to override this value if you would like to decrease or increase it based on your content and editorial approach. Please contact support for more information on configuring this setting.

Normal content can be reshared by the Autofeed as long as the Only reshare content classified as Evergreen setting isn't active.

Evergreen

Articles that are classified as Evergreen will be posted by Echobox within 48 hours when scheduled with the Optimal timing option.

Evergreen content can always be reshared, regardless of the reshare settings for your page.

How it works

The time sensitivity classification uses machine learning to predict articles belonging to each category. Therefore, it cannot be understood with simple rules, such as “articles with topic A will be evergreen”. We create a numerical representation of each article, which reflects the information included in the article as the input data to the machine learning models, using historical user manual labels as what we are trying to predict. What the machine learning model learns is how to map these numerical representations onto each of the categories, as labelled by our users. The machine learning models are language specific, that is to say we have one for each language. We also retrain (update) them regularly so they can learn newer user classifications.

Machine learning of this nature inherently has a margin of error. This is partially because the user labelling is "noisy" and varies from customer to customer due to subjectivity, but also partially as machine learning cannot be 100% accurate, just as a human cannot be.

As we retrain the models regularly, manually labelling/correcting time sensitivity will contribute to there being more of your articles labelled the way you want them to be in the training data. This means that the model will be more likely to learn the way you classify articles. This is why we encourage users to manually update labels where possible.