Customer Experience Analytics Formula Blog

Use Struggle Analytics to Quickly Identify and Understand Customer Struggle to Optimize Customer Experiences

Posted by Steven Perry on Feb 6, 2018 7:00:00 AM


Customers today have come to consistently expect optimal online experiences.  Deliver anything less and you run the risk of not converting a user or potentially losing a loyal customer.   A crucial first step in the process of optimizing your customer experiences is your ability to quickly discover, understand and minimize customer struggle on your site.


Analysis Overview

By utilizing Struggle Analytics, a cognitive feature in Watson Customer Experience Analytics (CXA), you can easily monitor specific processes or promotional pages on your site to quickly identify customer struggle and uncover areas of opportunity to improve and optimize your customer experiences. 


Analysis Benefits:  

  • Easily monitor specific processes or promotional pages for struggle activity with minimal configuration effort
  • Create struggle factors based on pertinent events in your system and relative to your business focus
  • Quickly discover obstacles to superior customer experience on your site by setting alerts to be notified when struggle thresholds are exceeded, allowing you to take swift action to eliminate struggle before it becomes prevalent.
  • Employ session replay technology to view the customer struggle identified and rectify issues to optimize customer experiences on your site.


Analysis Formula

This customer experience analytics (CXA) formula will explain how to use Struggle Analytics in Watson CXA to effectively monitor specific processes, pages and events on your site for struggle activity, allowing you to quickly detect and analyze customer struggles and take corrective action to rectify issues and deliver frictionless customer experiences.


How Struggle Analytics work

Struggle Analytics is a cognitive feature that uses algorithms to learn over time what is normal user behavior on a page and what is not and assigns a struggle score based on struggle factors. The struggle score is calculated from weighted default factors and struggle event factors, and factors are weighed on a scale of 1 to 10.

By default, struggle factors for events in Watson CXA are calculated for a page URL or screen based on the following core attributes:

  1. Repeat pattern – repeated patterns of unproductive user actions
  2. Step count – number of user-initiated actions on a page such as clicks and entering text
  3. Load time for page in seconds
  4. UI element -- detect issues with UI elements, such as a user repeatedly clicking an icon

Note:  There is also the option to add user-defined events as struggle factors, allowing you to closely monitor and weight variables that are specific to your particular business.   Elements to consider could include abandonment events, error pages, no results searches, or events that capture struggle for customers in a particular region, etc.


Consider the pages you want to analyze for customer struggle

Customer struggle can vary depending on your site design or business focus, but some examples of struggles you may want to monitor could include users entering multiple credit card numbers (i.e. potential fraudulent activity) on your check out page, struggling with a form field (i.e. user confusion) on your registration page or having to wait extended periods of time for a page to load (i.e. poor or invalid web design) on your product page or shopping cart page


Configure Struggle Analytics for your application

Step 1 - Map pages and prepare application for analysis

  1. Select Struggle Analytics in Watson CXA and then from the Options menu in the upper right corner, select Map Page.
  2. Select Configure from the left side and enter a recognizable name for your mapped page or you can use the provided pages and icons to name your pages.
  3. Then enter the page URL you want to monitor and select the appropriate conditions you want to include for your page mapping.  The conditions that you specify can be And or OR and are based on the page URL or an Event.
  4. Next, test the set conditions against existing sessions by selecting Test Pages in the lower left.  Then, resolve any issues that are shown in the Test results window.


Page Mapping in Struggle Analytics image

Page mapping for Struggle Analytics displayed in Watson CXA


When all of your page maps have been added and tested, the Struggle Analytics service will begin to analyze customer struggle in sessions for the mapped application.  It takes about 10 minutes for the Struggle Analytics service in Watson CXA to begin posting results.  The generated struggle report will identify customer struggle for your mapped pages using a set of algorithms based on key struggle factors and a struggle score will be assigned.


Modify the default configuration for Struggle Factors

Struggle Analytics provides a predefined configuration for load time, step count, repeat pattern and UI element struggle factors.  However, Watson CXA allows you the option to modify the configuration for struggle factors that are taken into consideration in the algorithm.  You can change the weight (significance/importance) of a factor, add event-based factors, or exclude a struggle factor from being calculated on a particular page.   The ability to modify the configuration for Struggle Factors is helpful if you want to shift your focus to a particular area as your analysis becomes more specific over time.

Step 2 - Modify configuration for struggle factors:

  1. Click Options in the upper right corner
  2. Select Edit Struggle Factor from the menu
  3. The associated Default Factors and Struggle Event Factors are displayed.
    1. Adjust the struggle factor weight by sliding the circle left to right.
    2. Add an UI element factor by clicking on the plus sign “+” next to Default Factors
    3. Add an event-based factor by clicking the plus sign “+” next to Struggle Event Factors
  4. Click Save in the lower right to save your selections.

Edit Struggle Factor_image

 Edit Struggle Factor for Struggle Analytics displayed in Watson CXA


Modify the configuration for Struggle Thresholds

You have the option in Struggle Analytics to edit the configuration for Struggle Thresholds, allowing you to make modifications to struggle thresholds that are most relevant to business on your site.

  • The struggle score threshold is the maximum struggle score that is acceptable, or considered within normal range, for a session and is based on the total struggle score for a session.
  • The Percentage of sessions demonstrating struggleis the percentage of sessions that have a struggle score that is greater that the struggle score threshold.   

For example, in the screenshot below, the struggle score threshold is set to 80 and the Percentage of sessions demonstrating struggle is set to 5%.   If you have enabled the option to receive alerts, an alert will be sent if and when it is detected that 5% of the sessions are showing a struggle score of 80 or more. 


Set threshold and alert image

Edit Struggle Factor for Struggle Analytics displayed in Watson CXA


Step 3 - Modify the configuration for Struggle Threshold 

  1. Click Menu in the upper right corner
  2. Select Set Threshold & Alert from the menu
  3. As shown in the above image, the configuration screen will appear where you can:
    1. Edit the struggle score threshold
    2. Adjust the percentage of sessions demonstrating struggle


Understand and Use the Struggle Analytics data

The reports generated by Struggle Analytics display the sessions that match the page mapping conditions and struggle score and percentage that you set.   Highlighted are the top 10 struggling sessions, the session struggle score, and the session struggle indicators and number of instances of each struggle indicator.  From the details page, you can view the median load time, step count, and UI element struggle for the struggling page or event.   

Utilizing session replay technology in Watson CXA, you can then drill deeper and replay the sessions that included struggle activity and view the actual struggle, allowing you to easily identify the source of struggle and take quick action to rectify issues and optimize the customer experience on your site.


 Have a Question               Subscribe to CXA Blog

Topics: Watson Customer Experience Analytics, Watson CXA

Creating Journey Reports to Understand Customer Journeys and Optimize Conversions

Posted by Steven Perry on Dec 5, 2017 7:00:00 AM


Your customers are engaging with your business across many different journeys – with customer interactions taking place over multiple channels, devices and visits -- and engagements occurring anywhere and anytime.   At the same time, in a competitive environment where your customers have increasing expectations to enjoy optimal experiences at every touchpoint, addressing the challenge of visualizing and understanding your customers’ complete journeys is essential for enriching experiences, building loyal customers, and maximizing conversions. 


Analysis Overview

Creating journey reports in IBM Watson Customer Experience Analytics (CXA) allows you to quantify and visualize customer journeys – individually and in aggregate – across devices, channels and touchpoints, offering you the opportunity to quickly compare path popularity, duration, revenue and customer values across unique segments (i.e., web, mobile, email, etc.) and the ability to resolve potential issues, increase conversions and optimize every journey to create loyal customers.


Analysis Benefits

  • Visualize a consolidated view of customer journeys to recognize the most traveled paths, the shortest paths, the paths that generate the most revenue and the paths that are most successful in leading to conversions.
  • Identify where customers may struggle or succeed in their paths and drill down for a deeper understanding of customer journeys and the source of customer friction or customer conversions.
  • Examine how activity in one customer path may impact performance in other paths, allowing you to optimize all customer journeys and increase conversions in all channels.


Analysis Formula

This customer experience analytics (CXA) formula will assist you with creating customer journey reports in Watson CXA to offer you a complete view of the multi-channel paths your customer take, enabling you to identify the paths that are the most relevant to your business, such as paths that lead to conversions or paths that top-tier buyers pursue, allowing you to drill deeper and uncover insight you can use to optimize all channels and customer journeys.   In addition, the analysis of journey reports can also assist you with evaluating the paths that may require additional optimization to increase conversions.   


Components of journey reports

As you can configure journey reports to focus on analysis of particular audiences or specific customer interactions pertinent to your business, it is first helpful to understand the key components of journey reports.

  • Audiences: A journey report can be configured to include all of your customers or just a specific audience, such as top-tier buyers or customers from a specific demographic group.
  • Top paths: Top paths are identified through four metrics, including path duration, most traveled, average revenue generated per customer and how many unique customers traveled the path.   Journey reports can be configured to include the top five paths in the report results or only the top path.
  • Touch points: A touch point is the starting point or the ending point for a path in a journey report.  You must define at least one touch point—a start point or an end point -- or you can define both a start point and an end point.
  • Range of interactions: Touch points are the foundation that defines where the path starts or ends. If you define both a start and end point for a path, the path results include all interactions that occurred between the two touch points.  If you define only a start point as a touch point, you can specify how many days from the start point you want to look forward for interactions. If you define only an end point, you can specify how many days from the end point to look backward for interactions.


Creating customer journey reports

To create a journey report in Watson CXA to visualize the customer journey paths or patterns that you want to analyze for your business, do the following:

  1. In the navigation panel, click Create > Journey report. 
  2. From the Audience menu of the report builder panel, click the Edit icon to select an audience, or accept the default value of All customers.
  3. In the Interactions menu, click the Starting or Ending icon, or click one of the touch points on the page to add an interaction. Then, select an interaction to associate with the touch point.
  4. Accept the default date range to report on or you may click Modify to change it.
  5. While optional, you can add one or more filter conditions to the event by doing the following:
    1. Click Add filter and select an event attribute and operator from the lists.
    2. Enter a value to filter on.   If you want to add another filter criterion, click +.
    3. Select the filtering logic that you want to use from the list at the top of the window and then click Apply:
      • Satisfy all below applies AND logic to the filter processing where events must meet all of the listed filter conditions to be included in the results.
      • Satisfy any below applies OR logic to the filter processing where events that meet any of the listed filter conditions are included in the results.
  6. Then click Apply to add the interaction to the report configuration.  
    • Note:  If you want to define another interaction, repeat the process of selecting an interaction, date range and filter conditions (if applicable.)
  7.  In the Look period menu, you can specify the number of days in the range or accept the default range of 14 days.
    • Note:  If your path includes both a starting interaction and an ending interaction, you do not need to specify a look-back or look-forward range, as the report results will include all interactions that occurred between the starting and ending interactions.
      1. Enter the number of days to look back or forward in the box and then click Done 
  8. From the Metrics menu, select a criteria for ranking the paths in the report results:
    • Top journeys of individual metric. This option returns results for the top five journeys ranked according to the metric you select:  Most traveled, Duration, Average revenue, or Unique customers.
    • Top journey for all metrics. This option returns the top journey for each of the four metrics.
  9. To compile your journey report, click Generate.  You can then accept the default report name or change it or add a description.  Save your report by clicking Save.


Customer Journey image.png


Accessing and analyzing journey reports

Watson CXA processes two report types for your customers’ journeys:  Channel Analysis and Pattern Analysis.    By default, a journey report opens with the journey channel analysis results displayed when you click Manage > Journey reports and select the report you want from the list.  By then clicking on the Report type menu, you can switch to the journey pattern analysis results for your report.

  • Accessing journey channel data: Helps you analyze your journey channel data by providing four path view types: Separated view, Time-lapsed view, Merged view, and Mindset view.
  • Analyzing journey patterns: Helps you discover the sequences of customer behavior that have the greatest impact on the customer journey by using journey pattern analysis.


Using customer journey reports

While there are many use cases for customer journey reports, for this formula example we will apply journey channel analysis to analyze and understand the customer journeys that led to conversions and use this insight to successfully support a new product launch.   

  1. Configure the journey report to understand the paths of top-tier buyers
    1. As we want to better understand what led the top-tier buyers to convert on previous purchases of similar products, we configure the journey report to show only the paths that were taken by these buyers by selecting Top Tier Buyers from the Audience menu.
    2. There may be several successful paths by top-tier buyers, so we will select to view the top five most-traveled paths by selecting this option from the Report options menu. Or, you can select to view just the top path from the options menu, if you wish.
    3. For the selection of the end touch points of the paths, we will select the Cart Purchased event, as we want our report to capture only the paths that resulted in a conversion.
      1. Additional filters can also be applied to this event. For this formula example, we will apply a filter to specify the cart purchase of a particular product, as we want to identify the top paths for customers who converted or purchased a particular product, so that we can apply this insight to the launch of a similar product. 
  2. Analyzing the results of our journey report
    1. The journey channel report results show the top five most traveled paths that led to the purchase of our specified product.
    2. Each path shows the channels that customers used to interact with our brand, and the paths are sorted by most traveled in descending order. The report also provides metrics to compare paths by duration, average revenue, and number of unique customers.
    3. If we want to visually compare the top five paths along the time dimension, we can switch to the Time-lapsed view to quickly gauge the time elapsed between individual channels within a path.
  3. Drill into channel activity for deeper insight
    1. Drilling into the most frequently traveled path in the channel summary, we see an aggregate view of the customer interactions that the top-tier buyers completed in the different channels along their path to purchasing our particular product.
    2. By analyzing the interactions of top-tier buyers, we can identify what influenced their purchase.
      1. Perhaps they first opened a promotional email with a 10% discount offer, clicked the email link to the product summary page on the website, and then launched the promotional product video before making their purchase. Or, maybe they were influenced through a social media promotion to visit the product page, and they then downloaded a product brochure before finalizing their purchase.   By drilling down into the channel activity, we can uncover what interactions led to conversions and replicate those offerings to support conversions for our new product launch.


While our formula example utilizes journey channel analysis to gain insight into customer journeys, journey reports can also be configured to provide journey pattern analysis to help identify high-impact customer behavior patterns in customer journeys.   Armed with deeper insight into your customers’ interactions and journeys using journey reports in Watson CXA, you are better able to replicate successful customer paths, resolve potential issues or customer struggles, and optimize journeys in all channels to increase conversions.  


Watson, Can you help improve digital experience ebook                Subscribe to CXA Blog

Topics: Watson Customer Experience Analytics, IBM Watson Customer Experience Analytics, Watson CXA, customer journeys, journey reports

Creating Business Process Reports to Optimize Process Conversions

Posted by Steven Perry on Nov 28, 2017 7:00:00 AM


Offering an optimized customer experience requires more than just providing your online customers with enhanced web pages or advanced site functions -- it requires you to ensure that your entire online business process is optimized and nurtures conversions.  As customers progress through various process steps to complete transactions on your site, it is important to monitor your critical business processes and identify and alleviate customer friction within individual process steps.  Understanding where your customers may struggle within your business process offers you the opportunity to optimize your process steps, enrich customer experiences and improve conversions.  


Analysis Overview

By using process reports in Watson Customer Experience Analytics (CXA), you can gain visibility into the overall performance of your business processes and identify individual process steps that may require enhancements or improvements.   By then comparing the performance of process steps across a common process flow for different segments of your business, you can recognize segments that may require increased refinements and optimization.


Analysis Benefits:  

  • Measure and score overall business process success by identifying the abandonment and conversion rates across and within each step of a critical business process
  • Uncover customer struggles or usability issues in your business process steps to identify areas requiring improvements and optimization.
  • Compare segments across a common process flow to identify segments that may be underperforming and require fine-tuning to increase the success of your overall business process.


Analysis Formula

This customer experience analytics (CXA) formula will explain how you can:

  1. Create process reports in Watson CXA to gain visibility into the performance of your business processes by identifying the conversion rates and abandonment rates for all process steps.
  2. Drill deeper with session analysis to understand what causes abandonments in your process steps and implement the necessary improvements to increase conversions.
  3. Compare process reports for different segments across a common process flow to begin to recognize process issues with particular segments of your business, allowing you to improve underperforming segments and processes and overall optimize your customer experiences.


Create process reports for your critical business processes

Online business processes differ, but common examples include placing an order for a product on your ecommerce site or registering a new user account for your business.   Similarly, the process steps and sequence of events can vary and may include the following:   place items in a shopping cart, provide customer information and shipping address, complete payment method, review order and submit order for fulfillment.   The events that follow the placement of a customer order are key in your overall business process as well and can include:  processing payments, verification of orders and shipment of the final products and more.

As such, you will want to ensure you have events built to record the important process steps that make up the sequence of events that define the flow of your critical business processes. 


  1. Create a process report using the Report feature in Watson CXA
    1. Go to Optimize > Analytics and go to your workspace.
    2. Select Add widget and select Create Report > Process report from the Actions menu.  The New Process Report page is displayed.
    3. Select an application profile and click Next.  Select metrics as step page is displayed.
    4. Select the steps that define your business process.
      1. From the Metrics column, select the events that when sequenced together define the process that you want to score or evaluate.
      2. Drag each object one at a time a time to the Steps column. The metrics are linked together in a sequential process flow in the Steps column.
      3. For each step you select, the absolute numbers of visitors are counted, and a conversion rate is calculated.
    5. Enter a step name for each metric you added to the Steps column and click Next
    6. Select any additional segments or use the default segment only. 
    7. Click Create to Preview your Process report. 
    8. Click Save report to create the report and add it to your workspace.
  1. Evaluate your Process Report to understand how the steps in your business process are performing.
    1. Process reports automatically measure and score (against defined goals) the overall success and failure of your online process steps.
    2. Analyzing the scoring for each process step allows you to easily identify process steps that include a percentage of abandonments that fall outside expected or typical percentages for your business.

Ongoing monitoring of your process report is important and can help your identify significant changes in performance across your business processes that might signal a concern, allowing you to quickly rectify issues and keep your business processes and customer experiences optimized.


Drill deeper to understand what causes abandonments

By identifying process steps that reflect a higher than normal percentage of abandonments, you have the ability to analyze sessions to better understand the cause of customer friction or struggle within the process step(s).  


Process Report image_1.png

Example:  Process Report in Watson CXA


For example, if your process step for “Check Out” is showing a high percentage of abandonments, you can drill down into sessions to understand if customers may be entering expired promotion codes or an invalid payment type.  Or, if you have added a new step to your registration process, and abandonments are high for the new process step, you can use session analysis to understand if customers may be confused on the new step or if there may be an issue with your registration form or field.   As a streamlined and frictionless business process is key to creating loyal customers and boosting conversions, it is important to quickly identify and understand what obstacles may prevent customers from successfully progressing through your business process steps with optimized experiences.


Compare segments for a better understanding of your overall business process performance

Comparing your process reports for different segments allows you to look at how your various segments are performing across a common business process flow or how your data from process reports may change over a period of time by comparing different dates.   By comparing process reports, you are able to identify individual segments that encounter challenges in your business process steps and then apply improvements to increase conversions across your entire business process.  

For example, you can compare a process report for all mobile visitors and a process report for all desktop web visitors to determine if a segment encounters a struggle in a particular process step.    If your mobile visitors struggle during a particular process step, while your desktop visitors do not, you can then drill down further to understand what is causing friction for your mobile visitors.  Perhaps one of your process steps is not optimized for mobile?  Or, are your mobile abandonments isolated to a particular operating system?  You can also compare segments to understand if customers struggle in a particular geography or if customers encounter friction while using a particular browser across your business process.


1. Compare process reports for different segments

There are a couple of ways you can compare segments or audiences for process reports in Watson CXA. 

First, you can update the audience view or date view when viewing a process report.    

  1. While viewing a process report from your workspace, click the audience or date boxes in the upper left corner to switch your audience view or date range view
  2. The process report will refresh to reflect the data for the updated audience or date

Or, you can compare process reports for different audiences or dates for a common process flow in a side by side comparison view.

  1. From your workspace Open the Process report for which you want to compare data.
  2. Select the Compare The Compare mode dialog is displayed.
  3. Enter the compare criteria. You can compare by date or by segment.
    1. Enter the date or segment for the current report and enter the dates or segments for the comparison report.
    2. Select Applyto apply your criteria to the report.
    3. The original report and the report with the new date or segment are shown side by side for comparison. You can view the data in chart and grid form.
  4.  Select Save Reportto save the report to your workspace and enter a name for the report.
  5. Select the Full screenicon on the saved report to return to the side-by-side comparison view.
  6. Select the Comparelink to exit Compare mode.


Process Report image_2.png Example:  Comparison of Process Reports for Different Segments in Watson CXA


2. Analyze the comparison data across the process reports

Comparing the performance of process steps for different segments across a common process flow allows you to easily recognize customer struggles for a distinct segment and understand if a particular segment is impacting your business process success.  With this increased insight, you have the ability to make improvements that will increase conversions and the success of your business process across all segments  

By creating process reports, you are able to look at the performance of your overall business process and understand where you may be losing customers in your process, providing you the opportunity to optimize your process steps to improve the customer experience and boost conversions.


Watson, Can you help improve digital experience ebook                Subscribe to CXA Blog



Topics: Watson Customer Experience Analytics, Watson CXA, business process

Employ Real-time Alerting to Assist with Delivering Optimal Experiences

Posted by Steven Perry on Oct 10, 2017 7:00:00 AM


As customers expect an optimal experience every time they engage with your business, understanding where – and when – they may struggle on your site is crucial.  The ability to readily identify customer struggle, and quickly rectify issues for a consistently optimized customer experience, can be further compounded by the number of channels and campaigns that may be involved in driving your business, where you are required to optimize many more pages and potentially new processes.   But, how can you know exactly when customers encounter struggle on your site?  


Analysis Overview

Enabling real-time alerting to be automatically notified when struggle events occur on your site can assist you with quick detection and analysis of customer struggles, allowing you to swiftly take corrective actions to deliver frictionless customer experiences.


Analysis Benefits

  • Discover obstacles to superior customer experience through alerts that notify you at the first sign of customer struggle on your site.
  • Swiftly act on alerts and insights to proactively eliminate areas of struggle before they become prevalent.
  • Quickly optimize customer experiences with real-time struggle detection and alerting.


Analysis Formula

This customer experience analytics (CXA) formula will demonstrate how to create, use and manage alerts for defined struggle events on your site using Watson Customer Experience Analytics (CXA), allowing you quick notification of and increased insight into customer struggles that you can then use to rectify issues and optimize customer experiences. 

Consider the struggle events you want to closely monitor

Events are the foundation of alerts, as alerts are enabled to notify you when specific struggle events exceed a defined threshold on your site.   You can set alerts for:

  1. Struggle events you created using Struggle Analytics in Watson CXA.
  2. Events you have built in Event Manager to monitor struggle on your pages.

Depending on your website design and what activities you would like to monitor, you could consider creating events to evaluate struggle or user confusion on your check out page, product page, registration page or shopping cart page, and more.  

Set up alerts to monitor relevant struggle events

As configuring an alert requires some consideration on the threshold values and threshold intervals that are best to use for accurately monitoring your event activity, it is helpful to first let your events run for a short period of time.  This will allow you to observe what event activity is within expected or typical range on your site, so that you can better configure your alert notifications to trigger when your event activity falls outside of normal range.   However, if you are launching a new page (i.e. campaign or promotional page), a new process (i.e. registration), or perhaps launching into a new market, you may want to set up alerts right away to notify you of any immediate struggles or issues users encounter.

Alerts are created and monitored from Alert Manager in Watson CXA, and alert notifications are based on:

  • Event
  • Alert type
  • Threshold operator, greater than, less than, or in the range of the Threshold value
  • Time between alerts notifications


Select Alert Manager from the left navigation in Watson CXA and click “+ New Alert” in the upper right.   The basic info screen for creating alerts will display.

Alert Manager - Step 1
Enter a prescriptive alert name and ensure active switch is toggled to on
Alert Manager - Step 2
For alert type, there are several options to consider:
  1. Average: Sends an alert notification based on the average of a numeric event value collected during the alert interval based on the threshold value
  2. Count: Sends an alert notification based on the event count as compared to the threshold operator and value.
  3. Maximum: Sends an alert notification if the numeric event value collected exceeds this defined maximum during the alert interval based on the threshold
  4. Minimum: Sends an alert notification if the numeric event value collected exceeds this defined minimum during the alert interval based on the threshold value
  5. Ratio: Sends an alert notification based on the ratio of two events. You choose two events, one as the denominator and the other as the numerator. Alert notification is sent once the threshold has been met. The threshold is set as a percentage.
  6. Sum: Sends an alert notification based on the sum of a numeric event value collected during the alert interval based on the threshold value using the threshold operator.

The most common type of alert is typically count, as there are many events where a high or a low count could signal an issue with your site.  For example, a high number of bounces from your check out page or a low number of counts to your landing page.


Alert_image 1.pngExample:  Basic info screen for real-time alerting in Watson CXA


Additional considerations for configuring alerts

With the ability to select different alert types, as well as add dimensions to the event you would like to closely monitor, you have the flexibility to apply some creativity and uncover different activities that are taking place across your site and move beyond simply tracking page load issues or page not found errors.    

Use case examples include:  

  • Uncovering fraudulent activity on your site
  • Monitoring reception of a new product or campaign
  • Gauge progress in a new market
  • Identifying user struggle during peak or high-traffic times.

For example, if you want to be alerted to possible fraudulent activity taking place on your site like a user entering multiple credit card numbers in the same session, you can configure an alert for events tracking the process on your checkout page.  You would select ratio as the alert type and apply an event recording total credit card entries as the numerator for the ratio and an event recording the total sessions as the denominator.  

For global operations, you can apply dimensions to the events you are monitoring to segment by location and receive alerts for struggles that may occur in particular geographies or regions.   Or, perhaps you need to monitor influencers or referrals to your landing page as part of a new campaign.  Applying dimensions to your monitored event can allow you to be notified, for example, if counts to your campaign landing page drop or fall below expectations for your campaign.

In addition, if you are implementing a new campaign or registration process, you could consider configuring alerts for notification on the potential use of invalid promo codes or login issues on your registration page.   Monitoring abandoned cart rates or percentages is also helpful in alerting you to potential user struggles that could signal potential page or process issues.

Alert Manager - Step 3
After selecting the event and alert type, select the threshold operation (i.e. is greater than, is in the range or is less than) and enter a threshold value for the event activity you are monitoring. Again, you will want to consider what is within normal range for event activity on your site when deciding on these values.
Alert Manager - Step 4
After the threshold value is defined, select the alert interval time (i.e. between 1 and 1440 minutes.)

Click continue once your alert parameters are defined and the Recipient info screen is displayed. 

Step 1
Review the email address, email subject and alert message
  1. While the email address defaults to the email address of the person creating the alert, you can add additional recipients to receive alert notifications.
  2. Helpful tips:
    1. Use Email to SMS Text gateway to send the alert notification email as a text message to your mobile phone or a recipient’s mobile phone.
      1. For example, if you are a Verizon user and your mobile number is 949-555-1212, you can receive the email notification as a text by using [10-digit phone number] or
    2. Review and update the email subject to ensure it is prescriptive and allows you to readily identify the alert notification when it is received. This is especially helpful if you are receiving multiple alert notifications.

Alert_image 3.pngExample:  Recipient info screen for real-time alerting in Watson CXA


Click Continue once your Recipient info is confirmed and the Blackout info screen displays.

Step 2
After an Alert is sent, you can specify an interval of time during which no other alerts of this type are created as part of the Resume alerting. 

Note: You can also specify blackout periods during timeframes in which you may not want to receive alerts like during non-business hours, for example.

It is also important to consider that very few systems will operate with 100% uptime or that your site will experience a struggle-free environment.  So, be sure to check your alert configurations if activity seems unusually quiet. 

Finally, complement real time detection of customer struggles with session replay technology in Watson CXA for increased visibility into your customers’ behaviors and struggles to pinpoint where and why experiences are occurring, allowing you to quickly rectify issues and readily optimize your customers’ experiences.

Topics: Watson Customer Experience Analytics, Watson CXA, alerting, real-time alerting

Optimize Marketing Landing Pages to Drive Marketing Effectiveness

Posted by Steven Perry on Sep 27, 2017 7:00:00 AM


There are a lot of factors to consider when developing your marketing strategy and campaigns.   One very important component that deserves special focus is the marketing landing page.   In many ways, your marketing landing pages are the heart of your campaigns and play a crucial role in a campaign’s success.  A great landing page not only serves to generate leads and create conversions, but tracking performance of your marketing landing pages can provide valuable insight you can use to better align your marketing strategies and maximize the effectiveness of your campaigns.   So, how do you go about ensuring your marketing landing pages are optimized to support your business objectives?


Analysis Overview

In this analysis, we will show how using marketing optimization reports can provide you an end-to-end view of the effectiveness of your marketing landing pages, offering insight from your customer behaviors that can assist you with optimizing your marketing landing pages to help improve engagements, enhance customer experiences and maximize conversions.


Analysis Benefits

  • Optimize marketing landing pages and maximize effectiveness of marketing campaigns utilizing a better understanding of customer behaviors and interactions.
  • Uncover usability and design flaws on your landing pages for increased visibility into areas requiring improvements and fine-tuning.
  • Utilize marketing insights to make better-informed decisions on how to best align your marketing strategy and improve campaign performance.

Analysis Formula

This customer experience analytics (CXA) formula will explain how you can use marketing optimization reports and usability analytics in Watson Customer Experience Analytics (CXA) to track landing page performance and identify areas for improvement on your marketing landing page design and overall campaign marketing strategy.   With this insight, you are better able to optimize your landing pages to improve the customer experiences and conversions while maximizing your marketing efforts and budgets.

Create MMC codes used in generating attribution reports

Step 1
Create the Marketing Management Center tracking codes (MMC codes) used in attribution reports in Watson CXA utilizing the Tracking Code Generator. The attribution reports, such as the Marketing Channels and Marketing Program reports, allow you to track key campaign performance indicators and monitor the performance of your marketing channel to identify trends that may be a concern or require immediate attention.   If you need help or missed our earlier Watson CXA formula, “Getting Started with Marketing Codes and Attribution Reporting,” click here.
  1. Use the Marketing Channels report to track key performance indicators for each channel, such as bounce rate, conversion rate, sales, or percent share of traffic, and monitor for significant changes in performance that might signal a concern.
  2. Use the Marketing Program report that provides a comprehensive view of your online paid marketing activity to drill deeper into your marketing campaigns and help identify potential reasons for changes in performance across your paid marketing elements.

Apply usability analytics for increased insight into customer interactions

Apply usability analytics, like heat maps, to evaluate customer behaviors and understand if your landing page design is optimal or if usability improvements are needed to improve your customer experiences and increase conversions.   Heat maps are helpful in identifying where customers are engaging on your page and recognizing areas of your page that may be causing struggles. 

Step 2
Apply heat maps overlay to your landing page to view and compare the customer interactions on a page. Then, segment your heat map data further to help you understand – and quantify -- if your landing page design is causing struggles that lead to customer abandonment.  If you need assistance, refer to our earlier CXA formula, “Applying Heat Maps to Understand Customer Behavior,” click here,

Generate marketing optimization reports to evaluate landing page performance

A successful marketing campaign – and optimized landing page – encompasses more than one strong component.   So, creating an end-to-end view of campaign activities – before, during and after engagements with your landing pages – allows you to pinpoint what elements are working and identify the segments that may be underperforming.  

Step 3
Create a marketing report to track your landing page views by page by day
  1. This is a great first step in helping you quickly understand how all of your landing pages are performing and allows you to zero in on the pages that are readily viewed and those that are not. Is there a landing page or two that really stands out?   Are there pages that peak one day to only drop in views the next?
Marketing Optimization_image 1.png

Example:  Daily landing page views by MMC

Step 4
Create a marketing report to track your daily landing page views by MMC (Marketing Management Center codes) to evaluate the performance of your campaigns.
  1. Understanding how landing pages are performing by campaign allows you to take a further step in identifying what campaigns are driving interest and what campaigns may need fine-tuning or more evaluation. 
Step 5
Create a marketing report to track your daily landing page views by referral source to evaluate how different channels are performing in directing traffic to your landing pages.
  1. This report will help you understand what influencers are driving traffic to your landing pages and what their impact on your campaign is. What marketing channels are driving the most traffic to your pages?   Is there a channel that is underperforming and could be eliminated to conserve budget or perhaps improved to maximize conversions?  
  2. You might then create a report that offers a view of landing page performance by MMC code and referral source allowing you to evaluate if a particular combination of referrals by campaign are working exceptionally well – or perhaps not well at all.
Step 6
Create a marketing report to track call to action (CTA) by landing page to evaluate where users are clicking on once they land on your marketing landing page.
  1. With a better understanding of what channels and what campaigns are driving traffic to your landing pages, you can then begin to take a look at how well your marketing landing pages are performing and driving interest. How well are customers clicking on your CTAs across all your landing pages?   Is there a landing page that received a lot of views, but then CTAs were low?   Is there perhaps a design flaw or poor content placement on your underperforming pages that might be causing confusion or struggle? (Heat maps, as explained above, can help you evaluate if design issues may be an issue.)

Marketing Optimization_image 2.png

Example:  Landing page clicks by landing page


Step 7
Create a marketing report to track conversion rates by landing page to identify what landing pages are the most successful in driving users to convert.
  1. Of course, one of the valuable pieces of information is whether your landing pages are leading users to convert to customers. Are there any pages that perform exceptionally well with conversions?   Is there a conversion event you can identify?  Or was there a progression of success factors that contributed to conversion based on your collective review of the marketing reports?
Step 8
Create a marketing report to track conversions by landing page and by MMC and referral source and CTAs to then get a complete end-to-end view of how the combination of marketing elements are working together.
  1. A cumulative report will allow you to understand what marketing elements are working together to drive your business and what elements may need some further evaluation and fine-tuning. In order to best view this aggregate data, format in a table or a stacked bar or create the report with filters so you can refine the views to get a better understanding on how different elements impact your marketing efforts.


In order to ensure you can create the appropriate marketing reports, you will want to be sure to have the correct events enabled – like events to capture views on your landing pages, clicks on your CTAs and conversions events.   At the same time, the ability to dimensionalize your marketing data by campaign and channel is key, so it is essential to have the correct MMC codes generated for use in the various marketing reports, so that you can further evaluate what is impacting your campaign performance.  With a better understanding of what is driving your campaigns and leading to conversions, you gain valuable insight you can use to develop optimized marketing landing pages – pages designed to create enriched customer experiences and brand loyalty.   So, with the critical role that landing pages play in your campaigns, an optimized landing page ultimately translates into an optimized campaign driving maximized business results.


Watson, Can you help improve digital experience ebook              Subscribe to CXA Blog

Topics: Watson Customer Experience Analytics, Watson CXA, Marketing Optimization, MMC, Marketing Management Center Codes

Getting Started with Marketing Codes and Attribution Reporting

Posted by Steven Perry on Sep 19, 2017 7:00:00 AM


Understanding what influencers drive conversions across the marketing funnel is one of the important data points you can leverage to help drive your marketing campaign success.    Equipped with this valuable insight, you are able to better allocate your marketing resources and best align your marketing strategy to maximize opportunities.   In fact, the marketing attribution reports available in IBM Watson Customer Experience Analytics (CXA) can tell you a lot about what is driving your marketing campaigns – what marketing channels drive the most leads, what paid search terms perform the best, what key words do visitors associate with your brand, what websites refer the most traffic to your site, and more. 


Analysis Overview

In this analysis, we will share how to create the Marketing Management Center (MMC) codes that are used to generate marketing reports that can better inform you on the influencers -- marketing channels, programs, search and referring sites -- that are driving your marketing campaigns, allowing you to better refine your marketing strategy and maximize your marketing spend. 


Analysis Benefits

  • Create marketing analysis offering valuable and actionable insight on influencers that you can use to maximize the effectiveness of your marketing campaigns.
  • Utilize marketing insights to make better-informed decisions on how and where to invest your marketing resources to optimize your return on investment (ROI) and minimize wasted marketing budget.
  • Manage your marketing strategy's performance with increased efficiencies by identifying deficiencies and developing effective marketing processes.

Analysis Formula

This customer experience analytics (CXA) formula will demonstrate how to create the Marketing Management Center tracking codes (MMC codes) used in marketing program reporting in Watson CXA utilizing the Tracking Code Generator.   We will also highlight the marketing reports available that you can use to maximize your marketing efforts and increase your campaign performance.

Get started with marketing campaign analysis

As an entry point into marketing analysis for marketing programs, you can generate, format and append Marketing Management Center (MMC) tracking codes to your destination URLs to track off-site campaign links.  Data from the MMC tracking codes is used in the Marketing Programs report.   To reduce the possibility of error, you will want to create the Marketing Management Center (MMC) codes using the Tracking Code Generator. 

1. Download and install the Tracking Code Generator.

  1. Within IBM Digital Analytics, click Manage > Marketing > Tracking Codes.
    1. Note: Be sure to use the appropriate IBM® Digital Analytics service domain name for your organization.
  2. Launch the downloadable file to install the Tracking Code Generator. Once installation is complete, an icon will appear on your desktop.
  3. You can also access the Tracking Code Generator by clicking: Start > All Programs > IBM Digital Analytics >Tracking Code Generator.

2. Generate Marketing Management Center (MMC) tracking codes used in Marketing Program reports. You can specify marketing program attribute values to attach to the destination URLS for specific marketing programs (for example, cm_mmca1=).

  1. Open the Tracking Code Generator.
  2. Open the Excel file that you want to use, or download the template from the Marketing Program Codes tab of the Tracking Code Generator.
  3. Give your file a unique name and save it to your desktop.
  4. Add the destination URLs and parameter values to your file, then save the file (Refer to Diagram A below).
  5. On the Marketing Program Codes tab of the Tracking Code Generator, follow the on-screen instructions to select options for the marketing programs format, existing codes (if any) in your file, and advanced settings.
  6. Click Browse to select your Excel file.
  7. Click Create Codes, input a new unique file name for the output file, and save to your desktop
  8. If the file is successfully processed, a message displays, indicating the number of rows successfully processed.
  9. Open the Excel output file and you will see the MMC parameter appended to each URL. (Refer to Diagram B below).

                             Diagram A - Example:  Marketing Program codes input file

Marketing Attribution_image 1.png


Diagram B - Example:  Marketing Program codes output file

Marketing Attribution_image 2.png

If any rows in your file contain an error message in column F, correct the error and use the Tracking Code Generator to reprocess the file. If column F is blank, the URL in the row is a valid MMC destination URL.

Note:  The Tracking Code Generator requires proper formatting of Microsoft Excel input files to generate tracking codes correctly.  Be sure to follow these guidelines when using the Tracking Code Generator:

  • Do not skip rows in between values. The Tracking Code Generator will stop processing when it encounters an empty row.
  • Include http://as part of the destination URL.
  • Do not include multiple worksheets in your Excel files. The Tracking Code Generator will not process an Excel file that contains multiple worksheets.
  • Use unique file names for the downloaded Excel templates, and for your input and output files (for example, 1.2013.xlsand MMC_output_9.1.2013.xls).
  • If you use special characters in your parameter values, ensure that the characters you use are supported in Digital Analyticsparameter values.


Recognize what is influencing your campaigns

As mentioned above, generating and appending MMC codes to your destination URLs creates data used in the Market Programs report that offers a comprehensive view of your online paid marketing activity and provides a good start to analyzing the performance of your marketing campaigns and links.   Using the Marketing Programs report, you can analyze marketing programs to improve the performance of content or creative placement, identify creative element within emails that generate the most click-throughs and transactions to enhance future emails, as well as analyze various paid search terms to improve the performance of existing keyword marketing placements, and more. 


Marketing Channels Report_1B.png

Additional acquisition reports available include the Marketing Channels report that provides a high-level overview of the referral sources or vendor channels that direct traffic to your website, allowing you to pinpoint those channels that are the most effective and identify those that may be adding little value to your business.   The Natural Search report identifies website traffic and conversion that results from natural search and helps you identify which keywords visitors associate with your brand, allowing you to refine your paid search marketing strategy and optimize ROI.   Finally, the Referring Sites report recognizes the websites that are referring traffic to your website so you can measure the value of this traffic and identify areas of opportunity or isolate areas for possible improvement.  

Collectively, the insight offered through the marketing attribution reports provides you with actionable data that you can use to better align your marketing strategy and optimize your campaign results.


Watson, Can you help improve digital experience ebook              Subscribe to CXA Blog



Topics: Customer Experience Analytics Formula, Marketing Codes and Attribution Reporting

Search Optimization Through Deeper Insight on Customer Engagement

Posted by Steven Perry on Sep 1, 2017 4:37:27 PM

A great deal of time is often spent in developing content for websites and web pages, and effective content should quickly engage prospective customers and encourage their further interaction on your site. But, do you really understand if your customers find what they are searching for on your site? Do you know if the content you are offering is keeping them actively engaged on your pages?


Analysis Overview

While basic search metrics for click rates, CTA performance, multi-search, etc., can certainly give you a glimpse into customer engagement on your site, taking the metrics a step further with advanced search events and customized reporting can provide you with deeper insight into how engaged your customers are with your content and if they find what they are searching for on your site.


Analysis Benefits

  • Gain a deeper understanding of customer engagement on your web pages to help determine if your page content is effective and engaging for customers.
  • Provide insight that can assist you with optimizing search and improving content on your pages to increase customer experiences.

Search Optimization Graphic.png
Analysis Formula

This customer experience analytics (CXA) formula will outline how to create an advanced event to capture how long customers stay on a product page after a search or click through, and then build a customized report by search term and downstream user engagement to better understand if you are providing the right content to keep your customers actively engaged on your pages.

1) As you want to understand how long a customer stays on a page after they have clicked on an item after a search result or how long they stay on a page after they have clicked through to the page, you will need to create an advanced event. The advanced event should record the time customers spend on a product page after they have performed a search or click through.

For this example, we will illustrate an advanced event that fires ONLY IF the amount of time the customer spend on the product page is LESS THAN 20 seconds, as the purpose of this event is to fire when customers are not staying on the page that they clicked through to.

The advanced event should be structured in the following way to fire IF the noted conditions are met:

  1. //Search Results Page UNLOAD event exists in session
    //we need this to calculate the time spent on the page
    ($F.getEventCount("E_SEARCH_RESULTS_PAGE_LOAD_COPY_1488312064185") > 0
  2. //Search Results Page LOAD event exists in session
    //we need this as well to calculate the time spent on the page
    && $F.getEventCount("E_SEARCH_RESULTS_PAGE_LOAD_1488311781932") > 0
  3. //Last Referrer URL includes /search/
    //checks that this page was referred to by a search page
    && $S["SSV_1111"].toUpperCase().indexOf("/SEARCH/") >= 0 
  4. //Last Referrer URL includes ?terms=
    //checks that the referring search page contains a search term
    && $S["SSV_1111"].toUpperCase().indexOf("?TERMS=") >= 0
  5. //checks if the number of seconds from page LOAD to page UNLOAD is less than 20
    && ($F.getLastEventNumericValue("E_SEARCH_RESULTS_PAGE_LOAD_COPY_1488312064185") - $F.getLastEventNumericValue("E_SEARCH_RESULTS_PAGE_LOAD_1488311781932")) < 20)

IF all of the above conditions are met, the event will fire and record the time spent (i.e., less than 20 seconds) on the product page.

2) Next, build a customized report that will allow you to better understand the relation of search term and the amount of time (i.e. time under 20 seconds, in this example) that customers spend on your product page after they have searched or after they have clicked through.

In order to appropriately build and segment the report, you should do the following:

  1. As the underlying event that captures the search term is “Event Value Terms,” create a dimension out of this event called   “QueryString Value.”
  2. Then, place the dimension “QueryString Value” in the dimension group “QuerySearch.” By adding this dimension group to the advanced event, you are then able to segment the report by search term

 Watson, Can you help improve digital experience ebook              Subscribe to CXA Blog


Topics: Customer Experience Analytics Formula, Search Optimization

Applying Geospatial Analysis for Geographical Insight into your Data

Posted by Steven Perry on Jun 27, 2017 7:00:00 AM


Developing a complete view of your customers is critical in today’s competitive environment where it is essential to understand how customers are interacting with your business and why particular customer behaviors may be recognized on your site.   By expanding your reach with geospatial analytics available in Watson Customer Experience Analytics (CXA), you can now also understand where geographically customers are engaging or struggling with your organization, allowing you additional insight to make more informed and targeted decisions about your business on a global level.


Analysis Overview

By applying geo analytics to your reports, you can understand how your business is performing across geographical areas, allowing you to identify particular areas that may require improvement or adjustments and overall enabling you to maximize wide-scale business opportunities.


Analysis Benefits  

Geospatial analysis can be applied to many use cases to help identify geographical areas for improvement and assist with targeting key areas to maximize potential opportunities.   Some business examples could include:   

  • Examine online abandonment rates by state to remarket to customers with a targeted campaign or promotional offer for additional savings at a local storefront.
  • Evaluate error conditions or struggle indicators like broken pages by country for global operations to deploy necessary site improvements.
  • Assess mobile traffic by city, state or zip code to inform your mobile strategy, optimize mobile site design and/or develop improved mobile applications.
  • Review marketing landing page traffic by country or region to understand campaign effectiveness by geography and identify potential areas for campaign improvement.


Analysis Formula

This customer experience analytics (CXA) formula will assist you with applying geo analytics to event reporting available in Watson CXA to understand how your business is performing across states, regions or the globe.   At the same time, the formula will illustrate how to set a threshold to measure geo analytics data by positive or negative performance, as well as drill down into particular sessions, as appropriate, to visualize customer experiences for increased optimization.

This analysis is flexible and you should apply geo analytics to the event reports that are most important to your business goals or most relevant to your key performance indicators (KPIs).

Create a Geo report to begin to evaluate your data from a geographical view

Just as events populate reports, the Geo reports populate event data based on the location of a customer’s device or machine.

Step 1
To create a Geo report, navigate to the Event manager to either build an event or modify an existing event for geo analytics.
  1. The event you build or the event you select to modify will depend, of course, on what data you want to evaluate on a geographical level. Any event that you would like to evaluate on a geographical level can be created or modified for geo analysis.
  2. Looking at our example scenarios, you could build or modify an existing event tracking hits to particular website pages or to marketing landing pages you may have developed to support a campaign. Or, you could create or select an existing event tracking mobile interactions on your site or build an event to record online abandonments, as examples.         
Step 2
Once you have created an event or selected an event to modify, you can enable the event to collect geo analytic data by simply scrolling to the bottom of the event’s configuration, and then selecting the ON switch for Geo Analytics.
  1. When the geo analytic-enabled event is saved, you can use the event in geo analytic reports OR in standard reports using geo analytics as breakout values.
  2. Similarly, you can also enable dimensions for geo-analytics and integrate the dimensions as a breakout in a report.
  3. Using the Geo out of the box dimension groups, you can segment report data by continent, country, state or city.                                                                                                                            
Step 3
From an open Workspace, select “Add Widget” and select “Create Report.” From there, select “Geo report.”
  1. From the Geo report screen, select an application’s profile in the upper right. You could be monitoring both a mobile application and a desktop website for your business.
  2. Next, select a timeframe to populate report data You can select a fixed data range OR select to keep your report updated with rolling date ranges. 
  3. Finally, click “metrics” and then “add metrics.” An “add metrics” screen is displayed.  As Geo reports can be populated by multiple metrics, you can select the metric or metrics to be used in your report.
    1. From the metrics selector, you can select the metric by filtering by the tag that was assigned to the metric OR select the metric by using Search. For example, in the first use case on examining online abandonment rates by state, you could search by the word “abandonment” and then select the applicable event that was created to capture abandonments in reporting.
  4. When the correct event or events are highlighted and then applied from the metric selector, the Geo report is then displayed.

Apply thresholds to identify negative and positive event performance in Geo reports


Geo reports can track both positive and negative performances of events.  To identify positive and negative events in the Geo report, you can set a threshold to analyze geo analytics data.   The threshold value determines at what value we establish the geo report as positive (or healthy) or establish the geo report as negative (or critical).   

Step 1
From the displayed Geo report, select the “summary” tab from the upper right. A dashboard appears showing the “average” value of the mapped data, as well as the highest value and lowest value of the mapped data.   
  1. Looking again at our first scenario on examining online abandonment rates by state, for example, the dashboard will show the state with the highest frequency of abandonments and the total number of abandonments for that state, as well as display the data for the state with the lowest rate of abandonments and identify the average number of abandonments across the Geo report data.
Step 2
After reviewing the dashboard data, select the “threshold” indicator in the lower left. From the threshold indicator box, you can select to measure a positive performance or select to measure a negative performance.   Again using our example of abandonment rates by state, we would select to measure a negative performance, as we are populating our Geo report to track abandonments.                                  
Step 3
Finally, insert the threshold value to be used in your Geo reporting.
  1. While you can select any value for your threshold, it is helpful to consider the data in the summary dashboard when selecting your threshold value.  For example, if we again refer to our first use case, if the state with the highest rate of online abandonments includes 660 abandonments, we could use half of that number, or 330, to start out as our threshold indicator.
  2. You should consider KPI’s and performance goals assigned to your business when selecting the appropriate threshold value to include in your Geo reports.    

                              Geospatial image_test cut.jpg

Example:  Geo report with threshold values applied


Step 4
After applying the threshold value, the Geo report is updated to reflect positive or healthy performance in green. When numbers exceed the defined threshold, the numbers are displayed in red for negative or critical performance.
  1. The ability to view data in a Geo report allows you to quickly understand how business goals are performing across geographical areas and to easily identify geographies, regions, states or even cities where your business may require improvements, adjustments or enhancements.
    1. Did your campaign landing pages show critical performance in one or two countries compared to healthy performance reflected in 5 or 6 additional countries? What contributed to the poor performance in particular countries?
  2. Using replay capabilities available in Watson CXA, you can drill down into sessions from areas where a critical performance is identified in the Geo report and visualize what struggles or obstacles may be contributing to the negative results.
    1. Were campaign landing pages slow to load in critical performing countries? Did a design flaw on the landing page cause customers to struggle?    Geospatial analysis complemented with replay capabilities in Watson CXA can help you identify poor performance in select geographical areas and assist with visibility into the necessary improvements.

Geospatial analytics offers additional insight into your data and helps you better understand how your business is performing across countries, regions, states and even cities.   The observations from Geo reports can help you realize how promotional campaigns, render times, abandonments and more measure up across the globe, allowing you to reach beyond transactional data into locational information that can identify geographic trends and assist with making better business decisions to maximize global opportunities.  


Topics: Customer Experience Analytics Formula

Applying Heat Maps to Understand Customer Behavior

Posted by Steven Perry on May 30, 2017 7:00:00 AM


Customer expectations to have optimal experiences on today’s websites and mobile applications are soaring, and more and more customers are quick to abandon a site if they do not immediately find it engaging or cannot quickly locate what they are looking for.   But, how do you know if your site is optimized to instantly engage customers and designed to readily meet their needs?   How does click behavior differ on your site for those customers who abandon versus those who purchase?  While web analytics can identify what pages your customers click on, applying usability analytics such as heat maps allows you to view where customers are interacting with your pages and understand customer behaviors for different segments.   Armed with this increased insight, you are better equipped to implement design and layout changes to optimize your site, improve engagements and conversions and enhance customer experiences.


Analysis Overview

A heat map overlay identifies regions of a page where customers either click or hover or hover to click. Heat maps are very useful in identifying high and low interest areas on a page and helping you recognize usability issues in the page design.  By implementing heat map analysis from Tealeaf cxOverstat available with IBM Watson CXA, and applying dimensions to filter the report data by different segments of customers, you can understand where customers engage the most on your page and uncover design flaws that may be causing customer struggles – and abandonments --on your site.   


Analysis Benefits

  • Identify content areas on your pages that are of the most interest to customers and areas that may be causing confusion or struggles on your pages.
  • Uncover usability and design flaws on your site for increased visibility into areas requiring improvements and fine-tuning.
  • Facilitates optimal design elements that can improve engagements, maximize conversions and optimize customer experiences.


Analysis Formula

This customer experience analytics (CXA) formula will assist you with using heat maps analytics available in Watson CXA to view and compare customer interaction rates on areas of a page to identify the content areas that are the most engaging to customers and the areas that may be causing struggles.   The formula will then illustrate how applying dimensions to filter the data by different segments allows you to further evaluate customer behaviors and understand if your site design is optimal or if usability improvements are needed to enhance customer experiences and increase conversions. 


View and compare customer interaction rates on a page

For this analysis, you will apply the heat maps overlay to a page snapshot.  So, you will first want to:

  • Select a page snapshot to evaluate -- your home page, product page or even a campaign page would be good options – and select it from the Snapshot Gallery.
  • Select Heat map report from the list of available overlays.
  • Select the magnifying glass icon from the top tool bar to adjust the zoom and resize the page to ensure it is viewable on your screen, if necessary.

Heat Map Image_1.png                                                  Example of Heat map overlay applied to a page


In viewing the heat maps report, the red indicates areas of high customer interactions and the blue indicates areas of low customer interactions, allowing you to quickly view where customers are interacting with your page.    Do you see a high number of customer interactions on areas of the page that you wouldn’t expect?    Are most customers clicking where you intended with your page design?

You will then want to further analyze the report data and quantify the number of customers who interacted with the particular areas on your page.  To see the data summary for a specific area of the page:

  1. Select the “Sub-select” tool in the Snapshot Galley to open a sub-select box
  2. Drag the sub-select box over the area of the page you would like to investigate The sub-select tool will then display the data summary showing number of customers who clicked on the area you have selected.
  3. Drag the sub-select box over additional areas of the page to compare data.

Comparing interactions with the number of customers who clicked on different areas of your page can help you identify if customers are clicking where you expected or intended on your page.   For example, do you observe a lower click rate on your linked buttons or text?   But, then discover a higher click rate on unlinked text?   If so, perhaps there is a design flaw with your page that is causing customer struggle and confusion.  By visualizing where customers click the most on your pages, you can optimize content and link placement to increase engagements and maximize conversions.

Heat Map Image_2.png

        Heat map overlay with data summary displayed for interactions on center text


Use dimensions to filter and segment the data

To better understand how your page design is impacting customer behavior for different segments, you can apply dimensions to the heat map overlay data to filter the report by a specific dimension value.   For example, applying dimensions to segment mobile customers and desktop web customers to better understand how each user is engaging with your pages can offer you the opportunity to optimize your site for all customers.    Do you see higher interactions on your linked text for desktop web customers compared to mobile customers?   Could some design changes make it easier for your mobile customers to find important information on your page?   A deeper understanding of customer behavior relative to different segments can help you identify where particular customers struggle and others succeed.

Applying dimensions to your heat map data can also help you understand – and quantify -- if your page design is causing struggles that lead to abandonment.  To do so, filter the report data by those customers who interacted with a page and then abandoned the session without purchasing:

  1. Viewing the heat map overlay data report, select the filter icon (on upper tool bar) to see a list of dimensions for the data.
  2. Select a dimension group. For this analysis, select Overstat-goals – heat map and then purchase success.
  3. Select a dimension value. For this analysis, select abandoned as the dimension value, as we want to segment the customers who abandoned without purchasing.  
  4. Click filter and the report snapshot updates to display only the interactions of customers who clicked on the page and then abandoned the session without making a purchase.
  5. To quantify the data, select the sub-select tool and drag the sub-select box over the areas of interaction you want to evaluate. The data summary will show the number of customers who interacted on the page and then abandoned without converting.  

Heat Map_Image 3.png

Updated heat map report showing interactions for those customers who interacted on the page and then abandoned without purchasing.  Data summary display quantifies users who clicked on center of page and abandoned without purchasing.


Applying usability analysis like heat maps allows you to easily see areas of high – and low – interest on your pages by identifying the intensity of customer interactions.   Understanding where customers tend to focus on your pages can allow you to implement important design changes – like repositioning links on a page, redesigning a link button or modifying content placement -- that can directly improve brand engagement and increase conversions.   Recognizing what content your customers are engaging with the most also offers you the opportunity to customize and make your pages more personal for your customers.   Finally, on-going monitoring of customer interactions on your pages allows you to fine-tune and maintain optimal design elements and continually enhance customer experiences.

 Watson, Can you help improve digital experience ebook               Subscribe to CXA Blog

Topics: Customer Experience Analytics Formula

Evaluating Your Customers’ Mobile Experiences with Watson CXA

Posted by Steven Perry on May 23, 2017 7:00:00 AM


You understand the importance of offering exceptional customer experiences across all the journeys that customers take with your brand.   Of growing importance across these journeys is the mobile experience, as mobile is fast becoming the preferred customer tool for interacting with brands and it serves to integrate customer experiences across channels.   But, how can you begin to assess the mobile experience your site is offering to your customers?  Do your mobile customers struggle?  And if they do, where and why do they encounter issues? 


Analysis Overview

Using IBM Watson CXA, you can segment two major categories of customer engagement – desktop web and mobile web  – and compare conversion and abandonment rates for each to begin to understand how mobile is performing.  By further segmenting mobile by operating system, evaluating mobile gestures across different sessions and employing additional usability analytics, you can gain increased insight into your mobile customers’ intentions, experiences and struggles and better implement site design changes to ensure your mobile opportunities are maximized.


Analysis Benefits:  

  • Assess your customers’ mobile experiences to identify potential areas of struggle and better understand how your mobile web business is performing relative to desktop web.
  • Gain insight to address opportunities for increased optimization in your mobile channel
  • Offers opportunity to retarget customers who may have abandoned due to poor mobile experiences and the potential to build brand loyalty and maximize mobile sales.

Analysis Formula

This customer experience analytics (CXA) formula will assist you with using Watson CXA to further segment data you use to evaluate standard business processes to generate reports that provide views into your mobile business and mobile customers’ experiences, and then advise on how you can apply additional usability analysis like heat maps to achieve even more valuable insight into your customers’ mobile interactions with your site.

This analysis is flexible and you should select the standard business process metric reports that are most applicable to your site and business key performance indicators (KPIs.

Create Session lists segmented with mobile-specific data

Many of the reports you create to regularly examine standard business practices on your site – like reports on abandoned carts or abandoned sessions and reports on order confirmation -- can easily be constructed and segmented to display mobile-specific data.   

  1. To include mobile-specific data in your reporting, you will want to create events to record relevant information on mobile customers such as platform, operating system (OS), OS version, device, device model, device vendor, mobile carrier, as well as orientation changes, gestures, unresponsive gestures, and resize gestures.

As a starting point, compare your desktop web business to your mobile web business to get a baseline view of how mobile is performing.    For example, you could build your “abandoned carts” or “abandoned sessions” reports to include platform so you can take a look at both desktop web and mobile web customers.   Or, you could compare desktop web and mobile web customers in your “order confirmation” reports.    How many mobile customers complete orders compared with desktop web customers?    Do more mobile web customers abandon their carts compared to desktop web customers?   How many desktop web sessions take place in a day, week or month compared to the number of mobile web sessions?  

After you have a basic view of your mobile business, you will want to create session lists to further evaluate your mobile customers and better understand why and where they may struggle.   To create mobile session lists, you can select from the reports you built to evaluate your mobile business, as described above (i.e. order confirmation reports, abandoned carts or abandoned session reports), and use session search to further filter and segment mobile customers by mobile-specific attributes including orientation changes, gestures and mobile OS, for example.  

Mobile experience_image.png

Using session search, you can select from user-defined or out of the box objects to filter your search results by specific criteria or particular user behavior.    With session search, you can:

  1. You can select to “include” or “exclude” particular conditions in your session search.
  2. From the “Field” option, you can select from events, dimensions, session attributes and more to be included or excluded for your session search results.
  3. You can also select “add another condition” to include or exclude an additional condition to your session search.
  4. After you configure the desired search, a list of sessions matching your selected criteria will be displayed. Each row represents an individual customer session.
  5. You can replay any of the sessions by clicking the replay arrow as displayed to the far left of each row.
  6. You can also select from different “views” such as overview, customer, performance, gestures or environment.
  7. After the mobile data is properly configured, you can “export” the reported data and pin to your workspace for on-going analysis of your mobile performance.

As a start, you may want to filter your session search by mobile OS as a session attribute to begin to take a look at mobile performance (i.e. abandonments and conversions) relevant to specific mobile OS.     Was a particular mobile OS linked to a high number of abandonments?   To drill down further, you can filter the data by OS version or device model to better understand if a particular OS version or model of device is causing a struggle for the mobile customers on your site.  

Insight into mobile gestures like resizing and orientation changes are also important events and session attributes to consider when evaluating your mobile data.   Do you observe a high number or orientation changes for customers using iOS versus Android?   Do most customers using Windows Mobile perform resizing gestures?   To help answer these questions – and most importantly, to understand why these particular actions may be happening -- you can access replay for specific sessions from the mobile lists you create to see the exact customer interaction and experience on your site.  Are important function buttons readily displayed on the mobile screen?   Do particular mobile customers need to resize their screen or change orientation to properly see key information on your site?  The ability to replay sessions and observe the customer experience can help you answer these questions and work to better optimize your site. 


Apply usability analytics for increased insight into mobile experiences

Finally, applying additional usability analytics like heat maps can be very beneficial in identifying usability flaws on your mobile site that may cause customer confusion and struggles.   By identifying particular sessions and pages where your mobile customers struggled – and applying heat map analytics to the particular page – you can isolate what caused struggles for your mobile customers and better enable an optimal mobile site design.  Heat mapping can be a powerful tool in helping you to maximize opportunities with our mobile customers, and we will cover this in more detail in a future CXA formula.

Ensuring that you are offering exceptional experiences to your mobile customers is becoming a paramount component of an overall successful CX strategy.   Using Watson CXA, you can effectively segment standard business processes using mobile-specific attributes and begin to understand how your mobile channel is performing.   Further segmentation and analysis of the mobile data can help you drill down and understand where – and why – your mobile customers may struggle, allowing you the ability to optimize your mobile experience and maximize opportunities with mobile.  


Topics: Customer Experience Analytics Formula

About this blog

The CXA Formula Blog is designed to provide formulas or recipes in using IBM Watson Customer Experience Analytics and Tealeaf CX on Cloud. This includes formulas in digital analysis, customer experience analytics, journey analytics, and Universal Behavior Exchange. These formulas are designed to provide you insight and best practices in customer analytics.

Subscribe to the CXA Formula Blog

Watson, Can you help improve digital experience ebook
Read the Shubert TIcketing CXA Case Study

Recent Posts