Crisis Management and Spike Alerts in Listening
Crisis management and spike alerts help brands identify events that could negatively or positively impact the brand’s perception among customers and react to such events in time.
Emplifi Crisis Management for Listening allows you to set up alerts on top of existing Listening queries. These alerts automatically notify you about potential spikes in the number of mentions received during a short period of time. This was you always know something’s happening and you can respond to the situation.
While starting with an existing Listening query, you can target your spike alert rules with keywords and sentiment filters and get notifications only about the most relevant spikes in mentions.
For example, if you’re monitoring your brand name in a Listening query, you can filter down mentions with the keyword “Quality” and negative sentiment.
To properly set the sensitivity of spike identification, you can:
use our AI-driven “Smart Detection” based on analysis of historical data
manually set up the number of mentions you consider significant
or, if needed, get a notification about every mention that meets your definition.
You can get notifications about spikes via email and/or directly within the Emplifi platform. These notifications include links that would take you to Listening Analytics, where you can analyze and understand the data behind the spikes.
Learn more about how to use the functionality up to the maximum below.
Managing Spike Alert Rules
To manage spike alerts, head to Emplifi Settings → Listening → Spike alerts.
The section allows you to create new alerts and edit, monitor and delete existing ones.
It contains basic information about the alerts:
name
number of detected spikes
users to be notified (user notified when a spike is detected)
status of the alert
If you need to filter the alert list, you can do so by query sets, queries and users to be notified.
Alert status
The alert can be in four different states:
Running – The alert is active
Paused – The alert is paused
Invalid – The alert can be invalid when the query it’s supposed to monitor has been deleted.
On hold – The on hold status is affected by the status of the query sets or queries it’s supposed to monitor. It can happen when:
All the original queries have been paused, so they’re not collecting any data.
All the original queries have reached their mentions quota, so they’re not collecting any data.
The Emplifi Account has reached the yearly listening mentions quota. The queries will not collect any new data until the first day of the new billing period.
How to Set Up Spike Alerts
The setup process consists of three steps:
Source & condition
Alert sensitivity
Notifications
Each phase is described separately further below.
Permissions
Spike alerts are, by default, part of the Listening feature. There are no additional permissions, therefore, users with access to Listening settings (specifically, the Manage query sets and queries account permission) also have access to Spike alert settings.
If you need to change these settings, head to Settings → Roles & permissions → Account roles → Listening.
Creating Spike Alerts
To create a new alert, simply press the Create spike alert button to open the alert setup.
Let’s go through each step of spike alert creation.
Step 1 - Source & conditions
In the first step of the setup, you define the alert name, choose data sources it will monitor and set up the monitoring conditions (keywords and sentiment filter).
When choosing a data source, you have two options:
Choosing a query set
By default, all queries within the selected query set are automatically used as a data source.Choosing specific queries
You can select one or multiple queries based on the needs of your workflow.
When setting up the alert conditions:
Apply a keyword filter (optional). You can choose one of the following filters:
Any - searches for content containing at least one of the specified keywords, @mentions, or #hashtags.Technically speaking, it’s an OR logical operation.
All - searches for content containing all specified keywords, @mentions, or #hashtags.Technically speaking, it’s an AND logical operation.
None - excludes from the search all content that includes one or more of the specified keywords, @mentions, or #hashtags.Technically speaking, it’s a NOR logical operation.
Set up the sentiment filter.
All sentiment
Positive only – when you’re interested in monitoring, for example, potentially viral content
Negative only – when you want to monitor, for example, potential PR crises
Step 2 - Alert sensitivity
Choose the alert sensitivity that best fits your workflow. There are three options:
Smart detection
Mentions per hour exceeding…
All mentions
We describe them further below.
Data preview and spike volume estimation
To ensure you set up the spike alert correctly, we’ve implemented a preview tool that helps you predict how many spikes the alert is likely to detect and how many notifications you can expect.
Based on real data from the previous 7 days, we’re previewing how many spikes the current alert setup would detect.
Smart detection
Our AI-powered algorithm calculates the number of mentions per hour that can be considered normal and then compares the actual number to this benchmark based on the queries and query sets you choose.
You can then select how sensitive your alert should be and whether you want to be notified about:
Slight increase above normal
High increase above normal
Extreme increase above normal
Mentions per hour exceeding…
Get a notification every time the number of mentions per hour exceeds the chosen number of mentions.
All mentions
Get hourly notifications summing up all mentions that match your settings. You’ll get a notification every time there’s at least one mention in an hour window.
Step 3 - Notifications
The last step lets you choose who, how and how often should get notifications about spikes.
By pressing the Add user button, you can add people to the list of users to be notified and choose if they should be notified by email, in-product notification or both.
Notification types
Email notifications are sent to the user’s email connected to Emplifi (the one they’re using to log in).
If you’re not sure what the email address is, you can review it in Settings → Personal Settings → Social Identities.
The email notification includes two buttons that would take you to different parts of the Emplifi platform: Analytics (Review in Analytics) and Content (Open the mentions in Content).
When you’re redirected to the corresponding module, the search bar is automatically prefilled with relevant inputs so that you can immediately focus on the actual spike.
Immediate in-product notifications redirect you to Analytics where you can analyze the spike.
If you miss an immediate in-product notification, you can still find it in the platform’s Notifications Center. Access it by clicking your name in the platform’s navigation panel. If you want to see the spike alert data in Analytics, click the notification you’re interested in. You’ll be automatically redirected to relevant data.
Disabling spike notifications
There are two places in the Emplifi platform where you can set up spike alert notifications: in the platform Settings and in the settings of each alert.
You can disable in-product notifications about all spike alerts in the platform’s Settings → Notifications in the Listening section. However, it is not possible to globally turn off email notifications about spike alerts to help prevent potential issues no notifications could cause.
If you wish to unsubscribe from in-product or email notifications for a specific alert, you can do so in the Spike alerts tab in the Listening settings. Click the alert for which you want to disable the notifications. Go to the Notifications tab and set up your preferences. In this case you can disable both notification types.
How to Access Spike Data and Work With It
There are different ways to access and analyze the spike alert data:
Directly by searching for a specific spike alert in the Listening section in Unified Analytics, legacy Analytics, or Content modules
Directly by filtering up spike data in Community
From an email or in-product notification, you can learn more above
From the spike alert list in Listening settings
Analytics
In Unified Analytics → Listening → Crisis Management and legacy Analytics → Listening → Crisis Management, you can search for spike alerts in the search bar and select them as a data source.
Once selected, you’ll get a table providing basic information about all spikes detected during the selected time range.
For more data about the spikes, review the widgets below the table.
If you wish to get details about one specific spike, you can navigate to its data by clicking the Focus button on the right side of the particular line.
After opening a particular spike, you can see its progress in detail. As the spikes usually do not have a clear beginning and ending, you can drag and drop their “edges” to analyze them in a full context and review what preceded and what followed.
Community
You can use specific spike alerts to filter Listening mentions in Community and react to critical conversations as they happen in real-time.
To use the filter, search for the Spike alerts - Listening filter in the right panel and select the relevant spike alert.
You can select as many spike alerts as you need. Once you do, hit the Apply button to confirm your choice. When your filter settings are ready, click the Filter button in the bottom part of the panel.
Redirection from spike alerts table in Listening settings
Another way of accessing data from a specific alert is through redirect buttons in Listening settings that would take you to Analytics or Content.
You can find these buttons in Emplifi Settings → Listening → Spike alerts. Hover on the spike alert you’re interested in; the buttons will appear next to the alert name. The options are Open in Analytics and Open in Content. Pressing each will take you to its corresponding section in Analytics or Content.
Terminology
Spike alert is an automated alerting rule that monitors selected data sources (Listening data) and detects volume spikes based on defined keywords and sensitivity and notifies users about anomalies in the data.
Spike is an anomaly in the number of Listening mentions.
Spike notification is an in-product or email notification sent to users that provides them with basic information about the detected spike.
Spike alert sensitivity is a basic input needed for the correct settings of a spike alert. It allows users to set up a threshold value for the number of mentions and notify them when the number changes in a way they define as significant.